正在加载视频...

视频加载失败

加载此视频时出现问题。这可能是由于临时网络问题，或视频可能不可用。

Hello world! Flyby Robotics has shipped our first run of F-11 programmable flying AI robots. Dual NVIDIA Orin NX capable for 300 TOPS compute. 2TB SSD. 1 hour flight time. 5.7lb payload capacity. EW-resistant mesh radio capable. Zero CCP components. Made in Los Angeles.

Jason H. Lu 🇺🇸

7,101 subscribers

129,690 次观看 • 11 个月前 •via X (Twitter)

Anya Rossi• Live Now

Private livecam show

55 条评论

Jason H. Lu 🇺🇸 的头像

Jason H. Lu 🇺🇸11 个月前

State-of-the-art AI systems are advancing at great speed, but there are no flying machines to run them. Today, the world’s drones are either made in China (banned), closed systems that block developer access, or they lack meaningful AI compute.

Jason H. Lu 🇺🇸 的头像

Jason H. Lu 🇺🇸11 个月前

We built the F-11 and its software libraries to solve this. So far, we’ve shipped to Palantir, European defense primes, cutting-edge startups, and major universities. We’ve landed our first 7-figure contract with USAF, and 7X’ed YoY booked revenue since last year.

Jason H. Lu 🇺🇸 的头像

Jason H. Lu 🇺🇸11 个月前

currently works with 4 out of 5 branches of the @DeptofWar. Our work now spans end users at USAF and USMC, as well as the Army, and Navy.

Jason H. Lu 🇺🇸 的头像

Jason H. Lu 🇺🇸11 个月前

Next, we’re working on 10X’ing on-edge compute with our Gen II platform. We’re also opening up new production slots in 2026, allowing industrial developers wider access.

Jason H. Lu 🇺🇸 的头像

Jason H. Lu 🇺🇸11 个月前

America needs national champions. We’re here to answer the call.

Jason H. Lu 🇺🇸 的头像

Jason H. Lu 🇺🇸11 个月前

My deepest thanks to our team members, investors, advisors, and partners who got @FlybyRobotics to where we are today. @andrew_neciuk @BaoNguy88991086 @mickyabir @okimnathanyo @suzannexie @xuantruc_n @adrianfenty @apartovi @balajis @blader @CatOrman1 @eldavis8 @howardakumiah @jimmydouglas @naval @Phil_Brady_ @neo @NivDror @mpalank101 @michlimlim @rrhoover @vedikaja_in @aphysicist @Gundo_OG @StephanSturges @zanemountcastle

Jason H. Lu 🇺🇸 的头像

Jason H. Lu 🇺🇸11 个月前

And thank you to our developers and customers for believing in us early!

kache 的头像

kache11 个月前

@FlybyRobotics @nvidia i'd make extremely tiny ones for development fwiw

Jason H. Lu 🇺🇸 的头像

Jason H. Lu 🇺🇸11 个月前

@FlybyRobotics @nvidia Put a Silvus mesh radio on it and we can have the swarm go zooooom

Balázs Némethi 的头像

Balázs Némethi11 个月前

@yacineMTB @FlybyRobotics @nvidia he meant small drones so developers can ship improvements across the board vs buying the massive rig version at start

Jason H. Lu 🇺🇸 的头像

Jason H. Lu 🇺🇸11 个月前

@yacineMTB @FlybyRobotics @nvidia Currently, a lot of the revenue generating work our developers are shipping to require the carrying of high quality sensors, long flight times, and large compute. That’s why we built to this size. How fast an app can go to market is a deep consideration

Oriflamme 的头像

Oriflamme11 个月前

@FlybyRobotics @nvidia Might seem silly, but apart from the insane specs this is the small detail that immediately tells me you're in a different league from everyone else Amazing work

Aaron Slodov 的头像

Aaron Slodov11 个月前

@FlybyRobotics @nvidia hellllll yes

Jason H. Lu 🇺🇸 的头像

Jason H. Lu 🇺🇸11 个月前

@FlybyRobotics @nvidia Forged inside @atomic_inc pressed

Aaron Slodov 的头像

Aaron Slodov11 个月前

@FlybyRobotics @nvidia @atomic_inc 🫡

Ali Partovi 的头像

Ali Partovi11 个月前

@FlybyRobotics @nvidia I'm so proud of you all, Jason & team. Amazing milestone, and here's to much more ahead!!! 💙 We at @Neo are proud to have backed you.

Jason H. Lu 🇺🇸 的头像

Jason H. Lu 🇺🇸11 个月前

@FlybyRobotics @nvidia @neo Thank you Ali 🫡

EDITH 的头像

EDITH11 个月前

man i am so proud of you, this is such a great milestone, seeing you guys ship is just fuckin awesome. When I look at the rest of the market, even DJI’s Matrice 350 RTK tops out at about 55 minutes in the air and consumer drones like the Mini 3 and Mavic 3 series are in the 45–51 minute-ish range.. The Orin NX module itself provides up to 100 TOPS of AI performance , so using two of them and hitting 300 TOPS puts you in a completely different class...(off the charts) It also feels like perfect timing, doesn't it?. The U.S. Commerce Department has been considering rules to restrict or even ban Chinese‑made drones due to security worries and the existing alternatives are often closed platforms with little onboard compute. By building everything in LA with zero CCP components and giving developers an open software stack, you guys are filling a huge gap. Can’t wait to see how the Gen II platform performs congrats again, this is huge!

Jason H. Lu 🇺🇸 的头像

Jason H. Lu 🇺🇸11 个月前

@FlybyRobotics @nvidia Yes sir! Thank you 🙏

Evan Applegate 的头像

Evan Applegate11 个月前

@FlybyRobotics @nvidia @mickyabir spotted

Ethan Loosbrock 的头像

Ethan Loosbrock11 个月前

@FlybyRobotics @nvidia Congrats!

Jason H. Lu 🇺🇸 的头像

Jason H. Lu 🇺🇸11 个月前

@FlybyRobotics @nvidia Thank you Ethan

Luiz 的头像

Luiz11 个月前

@FlybyRobotics @nvidia Great seeing that Jason! :) I still remember meeting you here in a thread about SoC lol

Jason H. Lu 🇺🇸 的头像

Jason H. Lu 🇺🇸11 个月前

@FlybyRobotics @nvidia Haha! Good times

Luiz 的头像

Luiz11 个月前

@FlybyRobotics @nvidia The problem I mentioned still takes sleep out of me… 3 times I changed the architecture lol

PaintSandRepeat 的头像

PaintSandRepeat11 个月前

@FlybyRobotics @nvidia Love seeing this

Jason H. Lu 🇺🇸 的头像

Jason H. Lu 🇺🇸11 个月前

@FlybyRobotics @nvidia Thank you sir 🫡

⚡ 的头像

⚡11 个月前

@FlybyRobotics @nvidia Very interesting. Is it open source? How to test it ?

Cameron Rowe 的头像

Cameron Rowe11 个月前

@FlybyRobotics @nvidia Amazing work!! 🚁

Mykhailo Sorochuk 的头像

Mykhailo Sorochuk11 个月前

@FlybyRobotics @nvidia Innovative hardware, groundbreaking software. This will definitely change the game in drone AI.

Aung Khant 的头像

Aung Khant11 个月前

@FlybyRobotics @nvidia Why don't you guys step into fill the agricultural spraying drone vacuum once DJI is out by the end of this year?

Catherine Yeo 的头像

Catherine Yeo11 个月前

@FlybyRobotics @nvidia This is incredible, congrats!!

Jason H. Lu 🇺🇸 的头像

Jason H. Lu 🇺🇸11 个月前

@FlybyRobotics @nvidia Thank you! 😊

Tyler Plummer 的头像

Tyler Plummer11 个月前

@FlybyRobotics @nvidia These are awesome Jason and team

Jon Werthen 的头像

Jon Werthen11 个月前

@FlybyRobotics @nvidia Love to see it. Great video!

cemaxecuter 的头像

cemaxecuter11 个月前

@FlybyRobotics @nvidia That’s outstanding. You have an extra m2 slot on one of the Jetsons? If so, and you’d like to another an EW capability, please let me know.

Morgan Barrett 的头像

Morgan Barrett11 个月前

@FlybyRobotics @nvidia Let’s freaking goooooo

Ritwik Pavan 的头像

Ritwik Pavan11 个月前

@FlybyRobotics @nvidia congrats! this is incredible

Jason H. Lu 🇺🇸 的头像

Jason H. Lu 🇺🇸11 个月前

@FlybyRobotics @nvidia Thank you would love to meet!

Ritwik Pavan 的头像

Ritwik Pavan11 个月前

@FlybyRobotics @nvidia let’s do it

David Pinsen 的头像

David Pinsen11 个月前

@aphysicist @FlybyRobotics @nvidia What countries do you source your components from?

Owen Mounts 的头像

Owen Mounts11 个月前

@FlybyRobotics @nvidia W

Michael Ron Bowling 的头像

Michael Ron Bowling11 个月前

@FlybyRobotics @nvidia Great work!

Andrew Neciuk 的头像

Andrew Neciuk11 个月前

@FlybyRobotics @nvidia Changing how we think about drones. Autonomous. Agnostic. Affordable. Joining this team has been an honor, and I am so inspired by everyone @FlybyRobotics.

payback 的头像

payback11 个月前

@FlybyRobotics @nvidia is it open source? is the build open source? as italian (nato) contractor able to buy it? DM if needed

Jason H. Lu 🇺🇸 的头像

Jason H. Lu 🇺🇸11 个月前

@FlybyRobotics @nvidia You can!

Min Chon Chi 的头像

Min Chon Chi11 个月前

@FlybyRobotics @nvidia Interesting compute/size ratio with the dual Orin NX.

Michael Tsai 的头像

Michael Tsai11 个月前

@FlybyRobotics @nvidia Congrats!

Wes 的头像

Wes11 个月前

@FlybyRobotics @nvidia Looks like a DJI

Mario Maggio 的头像

Mario Maggio11 个月前

@FlybyRobotics @nvidia Been waiting to see what you were up to. Congrats!

Evan 的头像

Evan11 个月前

@FlybyRobotics @nvidia Holy shit you built a quadcopter! Insane! Give this guy a gov contact ASAP!

Man three 的头像

Man three11 个月前

@Scobleizer @FlybyRobotics @nvidia @grok what does EW resistant mean?

3A 的头像

3A11 个月前

@FlybyRobotics @nvidia @NWblueUSA detected

Ben Miller 的头像

Ben Miller11 个月前

@FlybyRobotics @nvidia Jason, please sell some shares on @Wefunder - a lot of us would love to be a part of this journey.

Raul Izahi Lopez 🇺🇸 的头像

Raul Izahi Lopez 🇺🇸11 个月前

@FlybyRobotics @nvidia Congratulations! 🎊 🍾

相关视频

If intelligence is the log of compute… it starts with a lot of compute! And that’s why we’re scaling our GPU fleet faster than anyone else. Just last year, we added over 2 gigawatts of new capacity – roughly the output of 2 nuclear power plants. And today we’re going further, announcing the world's most powerful AI datacenter, located in southeastern Wisconsin. Fairwater is a seamless cluster of hundreds of thousands of NVIDIA GB200s, connected by enough fiber to circle the Earth 4.5 times. It will deliver 10x the performance of the world’s fastest supercomputer today, enabling AI training and inference workloads at a level never before seen. For AI training workloads, you need compute at exponential scale. That’s why we designed the datacenter, GPU fleet, and network together as one integrated system. This ensures a single job can run from day 1 at exponential scale across thousands of GPUs. Fairwater uses a liquid-cooled closed-loop system for cooling GPUs that requires zero water for operations after construction. And we’re matching all of the energy that is consumed with renewable sources. And of course, it is just one of several similar sites we’re lighting up across our 70+ regions. We have multiple identical Fairwater datacenters under construction in other locations across the US, in addition to our AI infrastructure already deployed in over 100 datacenters around the world, powering model training, test-time compute, RL tuning, and real-time inference at global scale. Too often during times like this, people go with the current and only later wonder, how did we get here? With Fairwater, we're charting a new path: doing the hard engineering work, bringing compute, network, and storage into one highly scaled cluster, and designing closed-loop energy systems to meet real-world computing needs. And partnering with local communities to ensure it's thoughtfully done in a way that is sustainable, creates new jobs, and expands opportunity. We are thrilled to see this take hold in Wisconsin, and we are just getting started.

If intelligence is the log of compute… it starts with a lot of compute! And that’s why we’re scaling our GPU fleet faster than anyone else. Just last year, we added over 2 gigawatts of new capacity – roughly the output of 2 nuclear power plants. And today we’re going further, announcing the world's most powerful AI datacenter, located in southeastern Wisconsin. Fairwater is a seamless cluster of hundreds of thousands of NVIDIA GB200s, connected by enough fiber to circle the Earth 4.5 times. It will deliver 10x the performance of the world’s fastest supercomputer today, enabling AI training and inference workloads at a level never before seen. For AI training workloads, you need compute at exponential scale. That’s why we designed the datacenter, GPU fleet, and network together as one integrated system. This ensures a single job can run from day 1 at exponential scale across thousands of GPUs. Fairwater uses a liquid-cooled closed-loop system for cooling GPUs that requires zero water for operations after construction. And we’re matching all of the energy that is consumed with renewable sources. And of course, it is just one of several similar sites we’re lighting up across our 70+ regions. We have multiple identical Fairwater datacenters under construction in other locations across the US, in addition to our AI infrastructure already deployed in over 100 datacenters around the world, powering model training, test-time compute, RL tuning, and real-time inference at global scale. Too often during times like this, people go with the current and only later wonder, how did we get here? With Fairwater, we're charting a new path: doing the hard engineering work, bringing compute, network, and storage into one highly scaled cluster, and designing closed-loop energy systems to meet real-world computing needs. And partnering with local communities to ensure it's thoughtfully done in a way that is sustainable, creates new jobs, and expands opportunity. We are thrilled to see this take hold in Wisconsin, and we are just getting started.

Satya Nadella

2,026,666 次观看 • 1 年前

$Unitree Robotics just filed to go public on March 20th targeting a $7 billion valuation. Most people have no idea what this company actually is. Here is why this might be the most important robotics IPO of the decade. Unitree shipped 5,500 humanoid robots in 2025. Figure AI shipped roughly 150. Agility Robotics shipped roughly 150. Unitree did $246 million in revenue last year, up 335% year over year, and they are actually profitable. Figure AI is valued at $39 billion with near zero revenue and is still private. Unitree wants $7 billion with real numbers. The price point is what separates them from everyone else. Their G1 humanoid sells for $13,500. Competitors charge $50,000 to $130,000. Their newest R1 humanoid launching in April starts at $4,900. Nothing comparable exists at that price anywhere in the world. They hold roughly 32% of the global humanoid market and 70% of the quadruped market. The moat is vertical integration. They self develop over 90% of core components including motors, reducers, controllers, sensors, and all software. Real clients include PetroChina, Sinopec, State Grid, and China Mobile. This is not a research project. Product is shipping at scale. This is listing on China's STAR Market, not Hong Kong, not the US, which makes access extremely difficult for international investors. The risks are real. The US House Select Committee on the CCP has formally requested Unitree be blacklisted. Their robots appeared in PLA military exercises in 2024. Tariffs have already nearly tripled the US price of the G1. $TSLA Optimus is targeting sub $20,000 pricing with automotive scale manufacturing backed by $NVDA compute. If they execute, the price advantage shrinks fast. But this is still the only profitable pure play humanoid robotics company in the world growing at 335% a year, valued at a fraction of its loss making peers. Goldman projects the humanoid market at $38 billion by 2035. Morgan Stanley goes to $5 trillion by 2050. Unitree currently holds the largest market share of any humanoid manufacturer on the planet. Full breakdown coming soon. $TSLA $NVDA$

Unitree Robotics just filed to go public on March 20th targeting a $7 billion valuation. Most people have no idea what this company actually is. Here is why this might be the most important robotics IPO of the decade. Unitree shipped 5,500 humanoid robots in 2025. Figure AI shipped roughly 150. Agility Robotics shipped roughly 150. Unitree did $246 million in revenue last year, up 335% year over year, and they are actually profitable. Figure AI is valued at $39 billion with near zero revenue and is still private. Unitree wants $7 billion with real numbers. The price point is what separates them from everyone else. Their G1 humanoid sells for $13,500. Competitors charge $50,000 to $130,000. Their newest R1 humanoid launching in April starts at $4,900. Nothing comparable exists at that price anywhere in the world. They hold roughly 32% of the global humanoid market and 70% of the quadruped market. The moat is vertical integration. They self develop over 90% of core components including motors, reducers, controllers, sensors, and all software. Real clients include PetroChina, Sinopec, State Grid, and China Mobile. This is not a research project. Product is shipping at scale. This is listing on China's STAR Market, not Hong Kong, not the US, which makes access extremely difficult for international investors. The risks are real. The US House Select Committee on the CCP has formally requested Unitree be blacklisted. Their robots appeared in PLA military exercises in 2024. Tariffs have already nearly tripled the US price of the G1. $TSLA Optimus is targeting sub $20,000 pricing with automotive scale manufacturing backed by $NVDA compute. If they execute, the price advantage shrinks fast. But this is still the only profitable pure play humanoid robotics company in the world growing at 335% a year, valued at a fraction of its loss making peers. Goldman projects the humanoid market at $38 billion by 2035. Morgan Stanley goes to $5 trillion by 2050. Unitree currently holds the largest market share of any humanoid manufacturer on the planet. Full breakdown coming soon. $TSLA $NVDA

KawzInvests

74,078 次观看 • 6 个月前

We are back again :) After three weeks of quiet building. Introducing Genesis World 1.0, our latest simulation platform, the second release in our full-stack suite. Open-sourced. Robotics is still bottlenecked by the 1× speed of the physical world. Every model, checkpoint, and data recipe eventually needs to be tested on physical hardware, slowly, expensively, and with limited coverage. One hour in reality can become 100 days in simulation. That is how robotics model iteration moves from a wall-clock bottleneck to a compute problem. To make this work, simulation has to be both fast and trustworthy. Over the past year, we rebuilt the entire stack: a GPU-accelerated cross-platform compiler, penetration-free multi-physics contact solvers, unified rigid and deformable physics, and a photo-realistic renderer purpose-built for physical AI applications. We built Nyx, a high-performance path-traced rendering engine for robotics application. Genesis World 1.0 achieves near realtime performance with our latest development for penetration-free IPC solver, supporting various types of deformables beyond rigid bodies. It supports contact-rich, dexterous manipulation simulation across different embodiments: unitree, sharpa, wuji, genesis hand and various types of grippers. Under the hood is Quadrants, our effort in pushing forward cross-platform GPU-accelerated computation. Quadrants started as a fork of Taichi, and we rebuilt most of the critical parts for optimizing simulation workloads, giving 10x faster launch time and up to 4.6x runtime performance compared to the initial Genesis release. Together, they bring us to an unprecedentedly low sim-to-real gap, enabling zero-shot real-to-sim model evaluation and much faster iteration of GENE. All available today. Genesis World 1.0: Quadrants: Nyx:

We are back again :) After three weeks of quiet building. Introducing Genesis World 1.0, our latest simulation platform, the second release in our full-stack suite. Open-sourced. Robotics is still bottlenecked by the 1× speed of the physical world. Every model, checkpoint, and data recipe eventually needs to be tested on physical hardware, slowly, expensively, and with limited coverage. One hour in reality can become 100 days in simulation. That is how robotics model iteration moves from a wall-clock bottleneck to a compute problem. To make this work, simulation has to be both fast and trustworthy. Over the past year, we rebuilt the entire stack: a GPU-accelerated cross-platform compiler, penetration-free multi-physics contact solvers, unified rigid and deformable physics, and a photo-realistic renderer purpose-built for physical AI applications. We built Nyx, a high-performance path-traced rendering engine for robotics application. Genesis World 1.0 achieves near realtime performance with our latest development for penetration-free IPC solver, supporting various types of deformables beyond rigid bodies. It supports contact-rich, dexterous manipulation simulation across different embodiments: unitree, sharpa, wuji, genesis hand and various types of grippers. Under the hood is Quadrants, our effort in pushing forward cross-platform GPU-accelerated computation. Quadrants started as a fork of Taichi, and we rebuilt most of the critical parts for optimizing simulation workloads, giving 10x faster launch time and up to 4.6x runtime performance compared to the initial Genesis release. Together, they bring us to an unprecedentedly low sim-to-real gap, enabling zero-shot real-to-sim model evaluation and much faster iteration of GENE. All available today. Genesis World 1.0: Quadrants: Nyx:

Genesis AI

319,949 次观看 • 4 个月前

🚨 The Robot Episode @jason sits down with FOUR leaders in robotics to discuss the future of the industry, the $1/hour robo-worker, China's threat, robo-dog vs humanoid, and why this time is different: - Dr. Péter Fankhauser of ANYbotics - Bernt Børnich of 1X - Amanda McMaster of Boston Dynamics - Jonathan Hurst of Agility (0:00) Intro: Humanoids, Robots, & AI+ (0:57) ANYbotics' Dr. Péter Fankhauser: Why ANYbotics Bet the Company on Four-Legged Robot Dogs, Not Humanoids (13:18) Dr. Péter Fankhauser: China's Armed Robot Dogs Are "Stupid and Risky": Inside the Terminator Debate (15:01) NEO’s Bernt Børnich: Neo Ships in 2026 and Becomes an Open Robot Platform (22:18) Bernt Børnich: Robots Building Robots in Just 3 Years: “Hard Takeoff” (33:45) Boston Dynamics' Amanda McMaster: Spot Is Now the Most Deployed Robot on Earth, Atlas is the Most Capable (42:36) Amanda McMaster: Chinese Robotics Security Risks (47:14) Agility Robotics' Jonathan Hurst: Why "This Time Is Different" for Humanoids After 100 Years of False Starts (58:45) Jonathan Hurst: When Do Robots Outnumber Humans on the Factory Floor? --------------------------------------- Thanks to our partners for making this possible! Most advertisers have never heard of the platform with an $11B annual run rate in ad spend. AppLovin Ads. 1B+ daily active users, full-screen video ads watched for a median of 35 seconds, and businesses are profitably spending hundreds of thousands of dollars a day on it. Now open to everyone. Sign up at If your work depends on conversations — meetings, deal flow, interviews, customer calls — Plaud helps you capture and organize everything with highly accurate AI-generated notes that are not just simple summaries, but also highlight pain points, key decisions, next steps, and customizable summary templates. Check out Plaud at and use code ALLIN for up to 20% off! Which is also available on Amazon: (Code: ALLIN20X)

🚨 The Robot Episode @jason sits down with FOUR leaders in robotics to discuss the future of the industry, the $1/hour robo-worker, China's threat, robo-dog vs humanoid, and why this time is different: - Dr. Péter Fankhauser of ANYbotics - Bernt Børnich of 1X - Amanda McMaster of Boston Dynamics - Jonathan Hurst of Agility (0:00) Intro: Humanoids, Robots, & AI+ (0:57) ANYbotics' Dr. Péter Fankhauser: Why ANYbotics Bet the Company on Four-Legged Robot Dogs, Not Humanoids (13:18) Dr. Péter Fankhauser: China's Armed Robot Dogs Are "Stupid and Risky": Inside the Terminator Debate (15:01) NEO’s Bernt Børnich: Neo Ships in 2026 and Becomes an Open Robot Platform (22:18) Bernt Børnich: Robots Building Robots in Just 3 Years: “Hard Takeoff” (33:45) Boston Dynamics' Amanda McMaster: Spot Is Now the Most Deployed Robot on Earth, Atlas is the Most Capable (42:36) Amanda McMaster: Chinese Robotics Security Risks (47:14) Agility Robotics' Jonathan Hurst: Why "This Time Is Different" for Humanoids After 100 Years of False Starts (58:45) Jonathan Hurst: When Do Robots Outnumber Humans on the Factory Floor? --------------------------------------- Thanks to our partners for making this possible! Most advertisers have never heard of the platform with an $11B annual run rate in ad spend. AppLovin Ads. 1B+ daily active users, full-screen video ads watched for a median of 35 seconds, and businesses are profitably spending hundreds of thousands of dollars a day on it. Now open to everyone. Sign up at If your work depends on conversations — meetings, deal flow, interviews, customer calls — Plaud helps you capture and organize everything with highly accurate AI-generated notes that are not just simple summaries, but also highlight pain points, key decisions, next steps, and customizable summary templates. Check out Plaud at and use code ALLIN for up to 20% off! Which is also available on Amazon: (Code: ALLIN20X)

The All-In Podcast

342,271 次观看 • 1 个月前

If an AI can control 1,000 robots to perform 1 million skills in 1 billion different simulations, then it may "just work" in our real world, which is simply another point in the vast space of possible realities. This is the fundamental principle behind why simulation works so effectively for robotics. Real-world teleoperation data scales linearly with human time (< 24 hrs/robot/day). Sim data scales exponentially with compute. There are 3 big trends for simulators in the near future: 1. Massive parallelization on large clusters. Physics equations are "just" matrix math at their core. I hear GPUs are good at matrix math 🔥. One can run 100K copies of simulation on a single GPU. To put this number in perspective: 1 hour of wallclock compute time gives a robot 10 years (!!) of training experience. That's how Neo was able to learn martial arts in a blink of an eye in the Matrix Dojo. 2. Generative graphics pipeline. Traditionally, simulators require a huge amount of manual effort from artists: 3D assets, textures, scene layouts, etc. But every component in the workflow can be automated: text-to-image, text-to-3D mesh, and LLMs that write Universal Scene Description (USD) files as a coding exercise. RoboCasa is one example of a prior work. 3. End2end neural net that acts as simulator itself. This is still bluesky research and quite far from replacing a graphics pipeline, but we are seeing some exciting signs-of-life based on video gen models: Sora, Veo2, CogVideoX, Hunyuan (text-to-video); and action-driven world models: GameNGen, Oasis, Genie-2, etc. Genesis does great on (1) for certain tasks, shows good promises on (2), and could become a data generation tool for reaching (3). Its sim2real capabilities for locomotion are good, but there's still a long way to go for contact-rich, dexterous manipulation. It shows a bold vision and is on the right path to providing a virtual cradle for embodied AI. It is open-source and puts a streamlined user journey at the front and center. I had the privilege to know Zhou Xian and play a small part in his project since a year ago. Xian has been crunching code non-stop on Genesis with a very small group of core devs. He often replied to my messages at 3 am. Zhenjia Xu from our GEAR team helped with sim2real experiments in his spare time. Genesis is truly a grassroot effort with an intense focus on quality engineering. Nothing gives me more joy than seeing the simulation ecosystem bloom. Robotics should be a moonshot initiative owned by all of humanity. Congratulations.

If an AI can control 1,000 robots to perform 1 million skills in 1 billion different simulations, then it may "just work" in our real world, which is simply another point in the vast space of possible realities. This is the fundamental principle behind why simulation works so effectively for robotics. Real-world teleoperation data scales linearly with human time (< 24 hrs/robot/day). Sim data scales exponentially with compute. There are 3 big trends for simulators in the near future: 1. Massive parallelization on large clusters. Physics equations are "just" matrix math at their core. I hear GPUs are good at matrix math 🔥. One can run 100K copies of simulation on a single GPU. To put this number in perspective: 1 hour of wallclock compute time gives a robot 10 years (!!) of training experience. That's how Neo was able to learn martial arts in a blink of an eye in the Matrix Dojo. 2. Generative graphics pipeline. Traditionally, simulators require a huge amount of manual effort from artists: 3D assets, textures, scene layouts, etc. But every component in the workflow can be automated: text-to-image, text-to-3D mesh, and LLMs that write Universal Scene Description (USD) files as a coding exercise. RoboCasa is one example of a prior work. 3. End2end neural net that acts as simulator itself. This is still bluesky research and quite far from replacing a graphics pipeline, but we are seeing some exciting signs-of-life based on video gen models: Sora, Veo2, CogVideoX, Hunyuan (text-to-video); and action-driven world models: GameNGen, Oasis, Genie-2, etc. Genesis does great on (1) for certain tasks, shows good promises on (2), and could become a data generation tool for reaching (3). Its sim2real capabilities for locomotion are good, but there's still a long way to go for contact-rich, dexterous manipulation. It shows a bold vision and is on the right path to providing a virtual cradle for embodied AI. It is open-source and puts a streamlined user journey at the front and center. I had the privilege to know Zhou Xian and play a small part in his project since a year ago. Xian has been crunching code non-stop on Genesis with a very small group of core devs. He often replied to my messages at 3 am. Zhenjia Xu from our GEAR team helped with sim2real experiments in his spare time. Genesis is truly a grassroot effort with an intense focus on quality engineering. Nothing gives me more joy than seeing the simulation ecosystem bloom. Robotics should be a moonshot initiative owned by all of humanity. Congratulations.

Jim Fan

157,343 次观看 • 1 年前

This is WILD! One week before SpaceX's historic IPO, Google signed a deal to pay SpaceX $920 million per month from October 2026 through June 2029 for access to 110,000 Nvidia GPUs, CPUs, and related infrastructure (Save this). That is $11 billion per year and up to $30 billion over the life of the contract. This comes less than a month after Anthropic committed $1.25 billion per month for full access to the Colossus 1 data center in Memphis, 200,000+ GPUs, 300+ megawatts of power capacity, through 2029. Two of the most consequential AI labs in the world combined committed value over $70 billion. The question that haunted SpaceX's IPO roadshow was why did Elon keep spending billions constructing Colossus, Macro Hard and Macro Harder, three facilities totaling nearly 2 gigawatts of AI compute when xAI's revenue wasn't yet on the same trajectory as OpenAI or Anthropic? Wall Street was pricing in a risk that Elon was building capacity ahead of revenue which would mean sustained cash burn without a clear payback timeline. That concern was legitimate on its face, because xAI had been aggressive on model development but had not yet demonstrated the enterprise revenue numbers to justify the infrastructure cost. The answer is that the compute itself was always the product. Amazon has AWS, Microsoft has Azure, Google has Google Cloud, Elon just confirmed that he has been quietly building the fourth major hyperscale AI cloud and his first two paying customers are Google and Anthropic, the very companies most aggressively competing in the AI race. xAI's Colossus facility in Memphis was built at a speed that no traditional data center developer could match, it went from groundbreaking to operational in roughly 122 days. That is what happens when you have direct Nvidia relationships, a construction operation built around SpaceX-style execution, and a founder who treats infrastructure buildout the same way he treats rocket launches: compress every timeline and eliminate every bottleneck. The result is that SpaceX now has three operational facilities, Colossus, Macro Hard, and Macro Harder with Macro Hard and Macro Harder in Blackwell architecture running 1.2 gigawatts combined. Colossus 1, built on H100s and optimized for inference, is the facility that went to Anthropic first. The Blackwell-era facilities are where the next-generation training workloads happen and Google's deal suggests they are renting into that capacity as it comes online through the second half of 2026. Elon's compute leasing business would generate approximately $45 billion in incremental annual revenue on top of the mid-$20 billion range analysts had been modeling for SpaceX more than enough to fully subsidize the infrastructure investment and take the financial pressure off xAI delivering immediate AI product revenue. That changes the entire valuation conversation of SpaceX completely! Milk road remains bullish on Space and come join Milk Road Pro and get our full SpaceX IPO breakdown, how we're thinking about the $1.75 trillion valuation and our entire AI thesis. Link below!

This is WILD! One week before SpaceX's historic IPO, Google signed a deal to pay SpaceX $920 million per month from October 2026 through June 2029 for access to 110,000 Nvidia GPUs, CPUs, and related infrastructure (Save this). That is $11 billion per year and up to $30 billion over the life of the contract. This comes less than a month after Anthropic committed $1.25 billion per month for full access to the Colossus 1 data center in Memphis, 200,000+ GPUs, 300+ megawatts of power capacity, through 2029. Two of the most consequential AI labs in the world combined committed value over $70 billion. The question that haunted SpaceX's IPO roadshow was why did Elon keep spending billions constructing Colossus, Macro Hard and Macro Harder, three facilities totaling nearly 2 gigawatts of AI compute when xAI's revenue wasn't yet on the same trajectory as OpenAI or Anthropic? Wall Street was pricing in a risk that Elon was building capacity ahead of revenue which would mean sustained cash burn without a clear payback timeline. That concern was legitimate on its face, because xAI had been aggressive on model development but had not yet demonstrated the enterprise revenue numbers to justify the infrastructure cost. The answer is that the compute itself was always the product. Amazon has AWS, Microsoft has Azure, Google has Google Cloud, Elon just confirmed that he has been quietly building the fourth major hyperscale AI cloud and his first two paying customers are Google and Anthropic, the very companies most aggressively competing in the AI race. xAI's Colossus facility in Memphis was built at a speed that no traditional data center developer could match, it went from groundbreaking to operational in roughly 122 days. That is what happens when you have direct Nvidia relationships, a construction operation built around SpaceX-style execution, and a founder who treats infrastructure buildout the same way he treats rocket launches: compress every timeline and eliminate every bottleneck. The result is that SpaceX now has three operational facilities, Colossus, Macro Hard, and Macro Harder with Macro Hard and Macro Harder in Blackwell architecture running 1.2 gigawatts combined. Colossus 1, built on H100s and optimized for inference, is the facility that went to Anthropic first. The Blackwell-era facilities are where the next-generation training workloads happen and Google's deal suggests they are renting into that capacity as it comes online through the second half of 2026. Elon's compute leasing business would generate approximately $45 billion in incremental annual revenue on top of the mid-$20 billion range analysts had been modeling for SpaceX more than enough to fully subsidize the infrastructure investment and take the financial pressure off xAI delivering immediate AI product revenue. That changes the entire valuation conversation of SpaceX completely! Milk road remains bullish on Space and come join Milk Road Pro and get our full SpaceX IPO breakdown, how we're thinking about the $1.75 trillion valuation and our entire AI thesis. Link below!

Milk Road AI

762,493 次观看 • 3 个月前

Everything Elon said about Optimus on the Q4 2024 earnings call: ⦿ I see a path for Tesla to be the most valuable company in the world, possibly bigger than the next five companies combined, overwhelmingly due to autonomous vehicles and autonomous humanoid robots. ⦿ The training compute needed for Optimus will ultimately probably be 10× what is needed for cars. Humanoids likely have 1,000× more useS than a car, which doesn't mean training scales by 1,000×, but probably close to 10×. The training compute will scale progressively as Optimus becomes more productive. ⦿ Long-term, Optimus has the potential to generate $10 trillion in revenue. In that scenario, we can support a lot of training compute. Even $500 billion in training compute is a good deal (chuckles). ⦿ There's a lot of uncertainty with timing because several aspects are being iterated simultaneously. The internal plan is for roughly 10,000 robots to be built this year, but we'll more likely produce several thousand. ⦿ I'm confident those several thousand robots will be able to do useful things. ⦿ The lessons from Production V1 will inform the changes in Production V2, which we expect to launch around mid-next year. ⦿ Our goal, aspirationally, is to ramp 10× every year, but perhaps we end up with 5× growth per year. With that kind of growth, it won't be many years before we're making 100 million robots a year. ⦿ The off-the-shelf components didn't work well, so we had to design everything in-house, including the most sophisticated hand ever made. Optimus will be able to play a piano and thread a needle. ⦿ My long-term prediction is that Optimus will overwhelmingly be the value of the company. ⦿ Optimus is not design-locked. It is rapidly evolving in a good direction. Tesla has by far the best humanoid robotics engineers in the world. Tesla also has all the other necessary ingredients: battery pack, power electronics, charging, communications, real-world AI, and the ability to scale production. ⦿ What other companies are missing is real-world AI and the ability to scale to millions of units a year. ⦿ This year, we aim to use Optimus internally at Tesla. We can easily use several thousand robots at Tesla for repetitive tasks, such as loading sheet metal at the welding line. ⦿ The Production V1 line is roughly 1,000 units per month. The Production V2, launching around mid-next year, will be for 10k units per month. The line after that will be for 100k units a month. Of course, it takes time for any given line to reach its maximum potential. ⦿ A very rough guess: we'll start delivering Optimus to companies outside of Tesla in the second half of 2026. The ramp is going to be exponential, and demand will not be a problem. ⦿ Once we're above 1 million units per year, the production cost of Optimus will be less than $20,000. Its total mass and complexity are much lower than a car. At a similar production volume to the Model Y, Optimus should be about half the cost of a Model Y. ⦿ The price is a different matter than cost. The price of Optimus will be set by market demand. [This is by far the longest Elon has ever spent discussing Optimus on an earnings call.]

Everything Elon said about Optimus on the Q4 2024 earnings call: ⦿ I see a path for Tesla to be the most valuable company in the world, possibly bigger than the next five companies combined, overwhelmingly due to autonomous vehicles and autonomous humanoid robots. ⦿ The training compute needed for Optimus will ultimately probably be 10× what is needed for cars. Humanoids likely have 1,000× more useS than a car, which doesn't mean training scales by 1,000×, but probably close to 10×. The training compute will scale progressively as Optimus becomes more productive. ⦿ Long-term, Optimus has the potential to generate $10 trillion in revenue. In that scenario, we can support a lot of training compute. Even $500 billion in training compute is a good deal (chuckles). ⦿ There's a lot of uncertainty with timing because several aspects are being iterated simultaneously. The internal plan is for roughly 10,000 robots to be built this year, but we'll more likely produce several thousand. ⦿ I'm confident those several thousand robots will be able to do useful things. ⦿ The lessons from Production V1 will inform the changes in Production V2, which we expect to launch around mid-next year. ⦿ Our goal, aspirationally, is to ramp 10× every year, but perhaps we end up with 5× growth per year. With that kind of growth, it won't be many years before we're making 100 million robots a year. ⦿ The off-the-shelf components didn't work well, so we had to design everything in-house, including the most sophisticated hand ever made. Optimus will be able to play a piano and thread a needle. ⦿ My long-term prediction is that Optimus will overwhelmingly be the value of the company. ⦿ Optimus is not design-locked. It is rapidly evolving in a good direction. Tesla has by far the best humanoid robotics engineers in the world. Tesla also has all the other necessary ingredients: battery pack, power electronics, charging, communications, real-world AI, and the ability to scale production. ⦿ What other companies are missing is real-world AI and the ability to scale to millions of units a year. ⦿ This year, we aim to use Optimus internally at Tesla. We can easily use several thousand robots at Tesla for repetitive tasks, such as loading sheet metal at the welding line. ⦿ The Production V1 line is roughly 1,000 units per month. The Production V2, launching around mid-next year, will be for 10k units per month. The line after that will be for 100k units a month. Of course, it takes time for any given line to reach its maximum potential. ⦿ A very rough guess: we'll start delivering Optimus to companies outside of Tesla in the second half of 2026. The ramp is going to be exponential, and demand will not be a problem. ⦿ Once we're above 1 million units per year, the production cost of Optimus will be less than $20,000. Its total mass and complexity are much lower than a car. At a similar production volume to the Model Y, Optimus should be about half the cost of a Model Y. ⦿ The price is a different matter than cost. The price of Optimus will be set by market demand. [This is by far the longest Elon has ever spent discussing Optimus on an earnings call.]

The Humanoid Hub

96,475 次观看 • 1 年前

Jensen Huang just BROKE the most important rule in the industry. And it explains why Nvidia controls 95% of the AI chip market. Last night at CES, he unveiled Vera Rubin - the new AI supercomputer that's shipping right now. Full production started weeks ago. But here's the part that made every semiconductor engineer in the room go crazy: Reuben GPU is 5x faster than Blackwell. But only has 1.6x the transistors. That should be physically impossible. Moore's Law says you get maybe 25% more performance per transistor generation. Jensen just delivered 300%. How? He BROKE the most sacred rule in chip design. The rule every company follows: "Never redesign more than 1-2 chips per generation." Nvidia redesigned all six chips simultaneously. Vera CPU. Reuben GPU. Connect X9 networking. Bluefield 4 DPU. MVLink switches. Spectrum X Ethernet. Every. Single. Component. From scratch. He calls it "extreme co-design." The industry calls it insane. One rack now moves 240 terabytes per second. That's TWICE the entire global internet bandwidth. In a single rack. And it runs on 45°C water - no chillers needed. Which saves 6% of global data center power. But the real story isn't the hardware... It's what they're doing with it. Nvidia just open-sourced Alpha Mayo. The world's first reasoning autonomous vehicle AI. Mercedes-Benz CLA launches with it in Q1. Europe Q2. Asia by year-end. Not a concept car. Not a limited release. Full production vehicles. And the AI will even explain its reasoning out loud. "I'm slowing down because the truck ahead is braking and there's a cyclist merging." It thinks. Then tells you what it's thinking. Then executes. Jensen drove it through San Francisco for an hour yesterday. No hands. No interventions. Through heavy Sunday traffic. The whole thing is open source now. Every line of training code. Every data source. The entire stack. But why would Nvidia give this away? Because they learned something from the last year: Open models activated the entire world. DeepSeek R1 proved open source can hit the frontier. Downloads exploded. Every country, every startup, every researcher can now build AI. And they all need Nvidia hardware to train it. That's the strategy. Give away the recipes. Sell the kitchen. The partnerships tell you where this is going: Siemens is integrating Nvidia into every industrial design tool. Cadence and Synopsys are rebuilding chip design around Nvidia. Palantir, ServiceNow, Snowflake - their entire platforms now run on Nvidia's agentic AI stack. This isn't just selling chips anymore. Nvidia is rebuilding the entire computing stack. From design to manufacturing to deployment. Every layer of the trillion-dollar AI infrastructure buildout runs through them. And now they're 18 months ahead of everyone else. Again. The competition is still trying to match Blackwell. Nvidia's already shipping the thing that makes Blackwell look slow. What do you think - is anyone catching them? The only company capable of this might be Google.

Jensen Huang just BROKE the most important rule in the industry. And it explains why Nvidia controls 95% of the AI chip market. Last night at CES, he unveiled Vera Rubin - the new AI supercomputer that's shipping right now. Full production started weeks ago. But here's the part that made every semiconductor engineer in the room go crazy: Reuben GPU is 5x faster than Blackwell. But only has 1.6x the transistors. That should be physically impossible. Moore's Law says you get maybe 25% more performance per transistor generation. Jensen just delivered 300%. How? He BROKE the most sacred rule in chip design. The rule every company follows: "Never redesign more than 1-2 chips per generation." Nvidia redesigned all six chips simultaneously. Vera CPU. Reuben GPU. Connect X9 networking. Bluefield 4 DPU. MVLink switches. Spectrum X Ethernet. Every. Single. Component. From scratch. He calls it "extreme co-design." The industry calls it insane. One rack now moves 240 terabytes per second. That's TWICE the entire global internet bandwidth. In a single rack. And it runs on 45°C water - no chillers needed. Which saves 6% of global data center power. But the real story isn't the hardware... It's what they're doing with it. Nvidia just open-sourced Alpha Mayo. The world's first reasoning autonomous vehicle AI. Mercedes-Benz CLA launches with it in Q1. Europe Q2. Asia by year-end. Not a concept car. Not a limited release. Full production vehicles. And the AI will even explain its reasoning out loud. "I'm slowing down because the truck ahead is braking and there's a cyclist merging." It thinks. Then tells you what it's thinking. Then executes. Jensen drove it through San Francisco for an hour yesterday. No hands. No interventions. Through heavy Sunday traffic. The whole thing is open source now. Every line of training code. Every data source. The entire stack. But why would Nvidia give this away? Because they learned something from the last year: Open models activated the entire world. DeepSeek R1 proved open source can hit the frontier. Downloads exploded. Every country, every startup, every researcher can now build AI. And they all need Nvidia hardware to train it. That's the strategy. Give away the recipes. Sell the kitchen. The partnerships tell you where this is going: Siemens is integrating Nvidia into every industrial design tool. Cadence and Synopsys are rebuilding chip design around Nvidia. Palantir, ServiceNow, Snowflake - their entire platforms now run on Nvidia's agentic AI stack. This isn't just selling chips anymore. Nvidia is rebuilding the entire computing stack. From design to manufacturing to deployment. Every layer of the trillion-dollar AI infrastructure buildout runs through them. And now they're 18 months ahead of everyone else. Again. The competition is still trying to match Blackwell. Nvidia's already shipping the thing that makes Blackwell look slow. What do you think - is anyone catching them? The only company capable of this might be Google.

Ricardo

580,362 次观看 • 8 个月前

This is WILD! MIT just solved one of the hardest unsolved problems in robotics (Save this). For decades, the fundamental problem with soft robots and wearable exoskeletons has not been compute or AI, it has been actuation. The moment you try to give a soft robot meaningful strength, you run into the same wall every engineer has hit since the field began, fluid-driven systems require external pumps, hydraulic reservoirs, and heavy infrastructure that makes the entire thing impractical to wear or embed into fabric. MIT's new Electrofluidic Fiber Muscles solve that problem by eliminating external infrastructure entirely. The key insight is electrohydrodynamic pumping using electric fields to generate pressure directly from electricity, with no moving parts, no motors, and no external fluid reservoir. The fibers are less than 2 millimeters thick, can be woven into fabric like ordinary textile, and operate in complete silence because nothing physically moves inside them, it is just ions propelling fluid through a closed circuit. The performance numbers published in Science Robotics are not conceptual, they are empirical results from actual hardware. These fibers achieve a power density of 50 watts per kilogram, matching skeletal muscle, with a contraction strain of 20% and a response time of 0.3 seconds. A single bundled configuration lifted 4 kilograms, 200 times its own weight while a separate configuration drove a robotic arm through a 40-degree bend compliant enough to safely complete a human handshake. Another configuration launched objects in under 100 milliseconds, which is faster than a human flinch reflex. The design mirrors biological muscle architecture in a way that prior artificial muscle approaches never achieved. The fibers are organized into antagonistic pairs, one contracts while the other extends, exactly like biceps and triceps and because the system runs in a closed loop, the relaxing fiber serves as the fluid reservoir for the contracting one, which is what allows the whole system to operate untethered with no external tank. The applications are not hypothetical but rather are the exact use cases the industry has been waiting years for the hardware to catch up to. Exoskeletons for physical labor, prosthetic limbs that move with the natural compliance of biological tissue, assistive garments for patients with motor disorders, and soft robots capable of safe physical contact with humans are all immediately unlocked by a muscle technology that is silent, lightweight, and weavable into clothing. The deeper significance is what this technology does when it meets the AI robotics wave that is already underway. Every major humanoid robot program, Figure, 1X, Boston Dynamics, Tesla Optimus is currently bottlenecked by the same hardware limitations these fibers address, actuators that are too rigid, too loud, too heavy, or too dependent on infrastructure to operate naturally alongside humans. Electrofluidic fiber muscles do not just solve a materials science problem but rather they remove one of the last physical barriers between robots that live in labs and robots that live in the world.

This is WILD! MIT just solved one of the hardest unsolved problems in robotics (Save this). For decades, the fundamental problem with soft robots and wearable exoskeletons has not been compute or AI, it has been actuation. The moment you try to give a soft robot meaningful strength, you run into the same wall every engineer has hit since the field began, fluid-driven systems require external pumps, hydraulic reservoirs, and heavy infrastructure that makes the entire thing impractical to wear or embed into fabric. MIT's new Electrofluidic Fiber Muscles solve that problem by eliminating external infrastructure entirely. The key insight is electrohydrodynamic pumping using electric fields to generate pressure directly from electricity, with no moving parts, no motors, and no external fluid reservoir. The fibers are less than 2 millimeters thick, can be woven into fabric like ordinary textile, and operate in complete silence because nothing physically moves inside them, it is just ions propelling fluid through a closed circuit. The performance numbers published in Science Robotics are not conceptual, they are empirical results from actual hardware. These fibers achieve a power density of 50 watts per kilogram, matching skeletal muscle, with a contraction strain of 20% and a response time of 0.3 seconds. A single bundled configuration lifted 4 kilograms, 200 times its own weight while a separate configuration drove a robotic arm through a 40-degree bend compliant enough to safely complete a human handshake. Another configuration launched objects in under 100 milliseconds, which is faster than a human flinch reflex. The design mirrors biological muscle architecture in a way that prior artificial muscle approaches never achieved. The fibers are organized into antagonistic pairs, one contracts while the other extends, exactly like biceps and triceps and because the system runs in a closed loop, the relaxing fiber serves as the fluid reservoir for the contracting one, which is what allows the whole system to operate untethered with no external tank. The applications are not hypothetical but rather are the exact use cases the industry has been waiting years for the hardware to catch up to. Exoskeletons for physical labor, prosthetic limbs that move with the natural compliance of biological tissue, assistive garments for patients with motor disorders, and soft robots capable of safe physical contact with humans are all immediately unlocked by a muscle technology that is silent, lightweight, and weavable into clothing. The deeper significance is what this technology does when it meets the AI robotics wave that is already underway. Every major humanoid robot program, Figure, 1X, Boston Dynamics, Tesla Optimus is currently bottlenecked by the same hardware limitations these fibers address, actuators that are too rigid, too loud, too heavy, or too dependent on infrastructure to operate naturally alongside humans. Electrofluidic fiber muscles do not just solve a materials science problem but rather they remove one of the last physical barriers between robots that live in labs and robots that live in the world.

Milk Road AI

1,208,301 次观看 • 4 个月前

Gavin's takes on Microsoft, Google, Meta, & Amazon: Microsoft ($MSFT): "I like Satya, I admire him. He's an exceptional CEO, and I give him a lot of credit for the decisions he's made. But he did go from, "We're going to make Google dance," to being the product manager of Copilot in 3 years. The decision Satya is making now, which the market has punished him for, but I think is the right decision — who knows how fast Azure could be growing if they were willing to just sell GPUs to OpenAI. 'We're going to use our compute internally to make our own products better.' One reason Copilot was so bad, or has been so bad, is that there wasn't enough compute available. They're fixing that. He's making good decisions that are risky decisions, to position Microsoft for this world where frontier models are no longer API-accessible. It's a really courageous decision that I give him a lot of credit for. Microsoft probably would be an $800 stock today if they were using their GPUs to serve solely OpenAI and Anthropic's capacity instead of using them for their own products." Google ($GOOG): "Google was incredible last year because they had that TPU advantage, which is now gone. The reason I think they're still in a great position is they have the most compute of everyone. We talked about the value of installed bases being higher as a result of shortages — they have the biggest installed base of compute. Google I/O is this week. If they don't release something that even slightly leapfrogs OpenAI and/or Claude, that's interesting. It's not a disaster for Google, it's just interesting. Between the amount of data they have, the YouTube data, the amount of compute, the search business — Google's never not going to be in a good position. You see that with GCP going crazy." Meta ($META): "You've got to give Zuckerberg immense credit, for what he's done in terms of making Meta an AI-first company internally. He is the only one of those true internet giants to have done that. I give him a lot of credit for paying up when he did for contracts, that talent. And Muse was a really big upside surprise. It was the first model from MSL, and it's not on the Pareto frontier with xAI, Google's one entrant, OpenAI and Claude, but it's pretty close. That was very impressive to me. So Meta is in a better position — still not as strong of an absolute position as Google, but a better position." Amazon ($AMZN): "Amazon is in a really strong position because of Trainium. You're going to see real P&L efficiencies from robotics over the next 18 months in their retail business. I actually think Nova — their internal models are not where Muse is, but they're better than they get credit for. The two companies who are the most deeply engaged with startups are Amazon and Nvidia by a mile. It's going to end up being a pretty big advantage for Nvidia and Amazon — with Google right behind them — to have this engagement that you just don't see from these other hyperscalers."

Gavin's takes on Microsoft, Google, Meta, & Amazon: Microsoft ($MSFT): "I like Satya, I admire him. He's an exceptional CEO, and I give him a lot of credit for the decisions he's made. But he did go from, "We're going to make Google dance," to being the product manager of Copilot in 3 years. The decision Satya is making now, which the market has punished him for, but I think is the right decision — who knows how fast Azure could be growing if they were willing to just sell GPUs to OpenAI. 'We're going to use our compute internally to make our own products better.' One reason Copilot was so bad, or has been so bad, is that there wasn't enough compute available. They're fixing that. He's making good decisions that are risky decisions, to position Microsoft for this world where frontier models are no longer API-accessible. It's a really courageous decision that I give him a lot of credit for. Microsoft probably would be an $800 stock today if they were using their GPUs to serve solely OpenAI and Anthropic's capacity instead of using them for their own products." Google ($GOOG): "Google was incredible last year because they had that TPU advantage, which is now gone. The reason I think they're still in a great position is they have the most compute of everyone. We talked about the value of installed bases being higher as a result of shortages — they have the biggest installed base of compute. Google I/O is this week. If they don't release something that even slightly leapfrogs OpenAI and/or Claude, that's interesting. It's not a disaster for Google, it's just interesting. Between the amount of data they have, the YouTube data, the amount of compute, the search business — Google's never not going to be in a good position. You see that with GCP going crazy." Meta ($META): "You've got to give Zuckerberg immense credit, for what he's done in terms of making Meta an AI-first company internally. He is the only one of those true internet giants to have done that. I give him a lot of credit for paying up when he did for contracts, that talent. And Muse was a really big upside surprise. It was the first model from MSL, and it's not on the Pareto frontier with xAI, Google's one entrant, OpenAI and Claude, but it's pretty close. That was very impressive to me. So Meta is in a better position — still not as strong of an absolute position as Google, but a better position." Amazon ($AMZN): "Amazon is in a really strong position because of Trainium. You're going to see real P&L efficiencies from robotics over the next 18 months in their retail business. I actually think Nova — their internal models are not where Muse is, but they're better than they get credit for. The two companies who are the most deeply engaged with startups are Amazon and Nvidia by a mile. It's going to end up being a pretty big advantage for Nvidia and Amazon — with Google right behind them — to have this engagement that you just don't see from these other hyperscalers."

Invest Like the Best

52,492 次观看 • 4 个月前

The U.S. unveils it's new F-47 stealth fighter, the centerpiece of the NGAD program, a "family of systems" designed to integrate advanced manned and unmanned platforms, including Collaborative Combat Aircraft (CCA) drones. Announced by President Donald Trump alongside Secretary of Defense Pete Hegseth and Air Force Chief of Staff Gen. David Allvin, the F-47 is described as the "most advanced, most capable, most lethal aircraft ever built." It reportedly builds on a prototype that has been secretly flying for nearly five years, suggesting significant testing and refinement prior to its public unveiling. The aircraft is engineered for speed, stealth, and adaptability, with a focus on countering advanced threats from nations like China, which has also been developing sixth-generation capabilities. Boeing’s victory over Lockheed Martin for the NGAD contract, valued at approximately $20 billion for the Engineering and Manufacturing Development (EMD) phase, marks a critical win for the company amid its recent struggles in defense and commercial sectors. The F-47 is expected to enter service in the 2030s, with each unit potentially costing upwards of $300 million, reflecting its cutting-edge technology. Its development emphasizes rapid adaptability to emerging threats, leveraging advanced manufacturing and an open architecture design to allow for continual upgrades. Since exact specifications remain classified or undisclosed as of now, the following are informed projections based on NGAD program objectives, statements from officials, and sixth-generation fighter trends. Designation: Boeing F-47 Manufacturer: Boeing Phantom Works Role: Air dominance fighter with multi-role capabilities (air-to-air and air-to-ground) Crew: Likely manned with optional unmanned configuration, aligning with sixth-generation flexibility Dimensions: Larger than the F-22 and F-35 to accommodate greater range and payload; exact size undisclosed but possibly exceeding 60 feet in length and a wingspan over 40 feet Powerplant: Expected to use adaptive cycle engines from the Next Generation Adaptive Propulsion (NGAP) program—either General Electric XA102 or Pratt & Whitney XA103. These engines feature a three-stream architecture, offering over 20% better fuel efficiency, increased thrust (potentially 45,000-50,000 lbf per engine), and enhanced electrical output for directed-energy weapons. Speed: Likely exceeds Mach 2 (super cruise capable—sustained supersonic flight without afterburners), surpassing the F-22’s Mach 1.8 super cruise Range: Combat radius projected at 1,000-1,500 nautical miles (unrefueled), tailored for Indo-Pacific operations, significantly greater than the F-22’s 600 nautical miles or F-35’s 670 nautical miles Stealth: Advanced stealth features, including a tailless design, next-generation coatings, and materials to reduce radar, infrared, and acoustic signatures beyond fifth-generation standards Payload: Larger internal weapons bays (possibly 20-23 feet long) to carry advanced munitions like the AIM-174, hypersonic missiles, and future cruise missiles, with external hardpoints available at the cost of stealth Sensors and Avionics: AI-enhanced sensor suite for unmatched situational awareness, integrating radar, infrared search and track (IRST), and electronic warfare systems; likely includes "smart skins" with embedded sensors for reduced drag and improved performance Networking: Maximum connectivity for real-time data sharing with satellites, drones, and other platforms, supported by a robust, jam-resistant data link Additional Features: Potential for directed-energy (laser) weapons to counter missiles and drones Integration with CCA drones for expanded mission options (e.g., extra munitions, electronic warfare) Open architecture for rapid upgrades and mission-specific customization Key Highlights Human-Machine Teaming: The F-47 is designed to "unlock the magic" of human-machine collaboration, pairing pilots with AI-driven systems and autonomous drones to enhance decision-making and reduce workload. Strategic Purpose: Built to penetrate contested environments, countering advanced air defenses and stealth fighters from adversaries like China, with a focus on long-range engagements over vast theaters. Development Timeline: Prototypes have been flying since at least 2020, with full operational capability targeted for the 2030s, replacing the F-22 incrementally as numbers grow. Cost and Scale: Estimated at $300 million per unit, with plans for roughly 200 manned aircraft, though this is a planning figure subject to change. The F-47’s exact design and full capabilities remain shrouded in secrecy, typical of NGAD’s classified nature, but its unveiling signals a bold step forward in U.S. air power. Its blend of stealth, speed, range, and technological integration positions it as a cornerstone of future aerial warfare, though its high cost and complexity will likely spark ongoing debate about affordability and strategic priorities. U.S. military technology will continue to dominate all other nations like it always has.

The U.S. unveils it's new F-47 stealth fighter, the centerpiece of the NGAD program, a "family of systems" designed to integrate advanced manned and unmanned platforms, including Collaborative Combat Aircraft (CCA) drones. Announced by President Donald Trump alongside Secretary of Defense Pete Hegseth and Air Force Chief of Staff Gen. David Allvin, the F-47 is described as the "most advanced, most capable, most lethal aircraft ever built." It reportedly builds on a prototype that has been secretly flying for nearly five years, suggesting significant testing and refinement prior to its public unveiling. The aircraft is engineered for speed, stealth, and adaptability, with a focus on countering advanced threats from nations like China, which has also been developing sixth-generation capabilities. Boeing’s victory over Lockheed Martin for the NGAD contract, valued at approximately $20 billion for the Engineering and Manufacturing Development (EMD) phase, marks a critical win for the company amid its recent struggles in defense and commercial sectors. The F-47 is expected to enter service in the 2030s, with each unit potentially costing upwards of $300 million, reflecting its cutting-edge technology. Its development emphasizes rapid adaptability to emerging threats, leveraging advanced manufacturing and an open architecture design to allow for continual upgrades. Since exact specifications remain classified or undisclosed as of now, the following are informed projections based on NGAD program objectives, statements from officials, and sixth-generation fighter trends. Designation: Boeing F-47 Manufacturer: Boeing Phantom Works Role: Air dominance fighter with multi-role capabilities (air-to-air and air-to-ground) Crew: Likely manned with optional unmanned configuration, aligning with sixth-generation flexibility Dimensions: Larger than the F-22 and F-35 to accommodate greater range and payload; exact size undisclosed but possibly exceeding 60 feet in length and a wingspan over 40 feet Powerplant: Expected to use adaptive cycle engines from the Next Generation Adaptive Propulsion (NGAP) program—either General Electric XA102 or Pratt & Whitney XA103. These engines feature a three-stream architecture, offering over 20% better fuel efficiency, increased thrust (potentially 45,000-50,000 lbf per engine), and enhanced electrical output for directed-energy weapons. Speed: Likely exceeds Mach 2 (super cruise capable—sustained supersonic flight without afterburners), surpassing the F-22’s Mach 1.8 super cruise Range: Combat radius projected at 1,000-1,500 nautical miles (unrefueled), tailored for Indo-Pacific operations, significantly greater than the F-22’s 600 nautical miles or F-35’s 670 nautical miles Stealth: Advanced stealth features, including a tailless design, next-generation coatings, and materials to reduce radar, infrared, and acoustic signatures beyond fifth-generation standards Payload: Larger internal weapons bays (possibly 20-23 feet long) to carry advanced munitions like the AIM-174, hypersonic missiles, and future cruise missiles, with external hardpoints available at the cost of stealth Sensors and Avionics: AI-enhanced sensor suite for unmatched situational awareness, integrating radar, infrared search and track (IRST), and electronic warfare systems; likely includes "smart skins" with embedded sensors for reduced drag and improved performance Networking: Maximum connectivity for real-time data sharing with satellites, drones, and other platforms, supported by a robust, jam-resistant data link Additional Features: Potential for directed-energy (laser) weapons to counter missiles and drones Integration with CCA drones for expanded mission options (e.g., extra munitions, electronic warfare) Open architecture for rapid upgrades and mission-specific customization Key Highlights Human-Machine Teaming: The F-47 is designed to "unlock the magic" of human-machine collaboration, pairing pilots with AI-driven systems and autonomous drones to enhance decision-making and reduce workload. Strategic Purpose: Built to penetrate contested environments, countering advanced air defenses and stealth fighters from adversaries like China, with a focus on long-range engagements over vast theaters. Development Timeline: Prototypes have been flying since at least 2020, with full operational capability targeted for the 2030s, replacing the F-22 incrementally as numbers grow. Cost and Scale: Estimated at $300 million per unit, with plans for roughly 200 manned aircraft, though this is a planning figure subject to change. The F-47’s exact design and full capabilities remain shrouded in secrecy, typical of NGAD’s classified nature, but its unveiling signals a bold step forward in U.S. air power. Its blend of stealth, speed, range, and technological integration positions it as a cornerstone of future aerial warfare, though its high cost and complexity will likely spark ongoing debate about affordability and strategic priorities. U.S. military technology will continue to dominate all other nations like it always has.

The SCIF

453,918 次观看 • 1 年前

Hello, Ghana 🇬🇭 Your Excellency, John Dramani Mahama Hello Africa and the world. We 3Farmate are proud to announce the official launch of FAMA—Ghana's FIRST AI-powered autonomous farming robot for large-scale crop production. Founded in 2021 by Clinton 🛸 (CEO) and Koffi-Cobbin (CTO), the company began in a dorm room at Kwame Nkrumah University of Science and Technology, where its first prototype was developed. Since then, the team has engineered FAMA into a full-scale autonomous robot capable of planting seeds, applying fertilizer, weeding, and spraying across real farm environments. FAMA navigates using a vision-based AI system instead of GPS, allowing it to operate reliably in areas where GPS is unavailable or inconsistent. The robot runs on batteries charged by solar panels while in the field and can operate across uneven terrain, loose and muddy soils, and variable weather conditions. A single operator can oversee multiple robots, each covering 27 to 35 acres per day with sub-85mm planting precision. 3Farmate targets large-scale staple crop producers in Ghana, starting with corn and soybeans. We operate a service model, charging farmers per acre and removing the need for upfront equipment investment. Over 70 farmers and several large-scale crop production companies are currently in discussions, with commercial deployments beginning in the 2026 planting season. Approximately $200,000 has been raised to date, including investment from 776 Foundation (Alexis Ohanian, Reddit co-founder) and a grant from Kosmos Innovation Center Ghana. Our team consists of young engineers specializing in robotics, embedded systems, software, and mechanical design. With 8 major iterations, 60+ field test runs, 100+ cumulative acres covered, and thousands of runtime hours in real farm conditions, FAMA is market-ready to become the ultimate farmer-assistant. We built FAMA right here in Ghana, inspired by Dr. Kwame Nkrumah’s belief that “Africa must industrialize to achieve true independence.” We built FAMA as a true testament to every young engineer in Africa that “IT IS POSSIBLE.” FAMA is designed to operate seamlessly on Ghanaian soil and is adaptable to diverse agricultural environments worldwide. Join us as we drive innovation across global agriculture. We call on the government of Ghana, stakeholders, international organizations, and agri-industry leaders to partner with us in transforming agriculture together. This is not history in the making, because history has already been made, and we thank you for being a part of our journey. On this note, we are officially launched! For more information, visit 3FARMATE – The future is here.

Hello, Ghana 🇬🇭 Your Excellency, John Dramani Mahama Hello Africa and the world. We 3Farmate are proud to announce the official launch of FAMA—Ghana's FIRST AI-powered autonomous farming robot for large-scale crop production. Founded in 2021 by Clinton 🛸 (CEO) and Koffi-Cobbin (CTO), the company began in a dorm room at Kwame Nkrumah University of Science and Technology, where its first prototype was developed. Since then, the team has engineered FAMA into a full-scale autonomous robot capable of planting seeds, applying fertilizer, weeding, and spraying across real farm environments. FAMA navigates using a vision-based AI system instead of GPS, allowing it to operate reliably in areas where GPS is unavailable or inconsistent. The robot runs on batteries charged by solar panels while in the field and can operate across uneven terrain, loose and muddy soils, and variable weather conditions. A single operator can oversee multiple robots, each covering 27 to 35 acres per day with sub-85mm planting precision. 3Farmate targets large-scale staple crop producers in Ghana, starting with corn and soybeans. We operate a service model, charging farmers per acre and removing the need for upfront equipment investment. Over 70 farmers and several large-scale crop production companies are currently in discussions, with commercial deployments beginning in the 2026 planting season. Approximately $200,000 has been raised to date, including investment from 776 Foundation (Alexis Ohanian, Reddit co-founder) and a grant from Kosmos Innovation Center Ghana. Our team consists of young engineers specializing in robotics, embedded systems, software, and mechanical design. With 8 major iterations, 60+ field test runs, 100+ cumulative acres covered, and thousands of runtime hours in real farm conditions, FAMA is market-ready to become the ultimate farmer-assistant. We built FAMA right here in Ghana, inspired by Dr. Kwame Nkrumah’s belief that “Africa must industrialize to achieve true independence.” We built FAMA as a true testament to every young engineer in Africa that “IT IS POSSIBLE.” FAMA is designed to operate seamlessly on Ghanaian soil and is adaptable to diverse agricultural environments worldwide. Join us as we drive innovation across global agriculture. We call on the government of Ghana, stakeholders, international organizations, and agri-industry leaders to partner with us in transforming agriculture together. This is not history in the making, because history has already been made, and we thank you for being a part of our journey. On this note, we are officially launched! For more information, visit 3FARMATE – The future is here.

MARK OFORIQUAYE

152,011 次观看 • 6 个月前

Chamath Palihapitiya just dropped the number that explains the entire AI infrastructure trade (Save this). A gigawatt of compute now costs $100 billion and when he started his Arizona data center project it was $4 to $5 billion, it has gone up 20x in a single investment cycle. The implication is not just that AI infrastructure is expensive but rather that the capital barrier to owning meaningful compute has become so high that only a handful of entities in the world can actually build it and the companies who got there early are sitting on what may be the most durable pricing power in the history of the technology industry. This is the neocloud trade. The neocloud market, purpose-built GPU cloud providers like CoreWeave, Nebius, and Lambda Labs was worth $35 billion in 2026 and is projected to reach $236 billion by 2031, compounding at 46% annually. For context, that is faster growth than cloud computing itself posted in its first decade. The reason is very simple, hyperscalers like AWS, Azure, and Google are building for everything, storage, databases, enterprise software, networking and their GPU pricing reflects the overhead of that full-stack infrastructure. Neoclouds build for one thing only, AI compute. The result is a 60% to 85% cost advantage on the same Nvidia silicon, bare metal H100s at $0.78 to $2.79 per GPU-hour on a neocloud versus $3.43 to $5.07 per GPU-hour on a hyperscaler. That spread does not close as AI demand scales but rather it widens, because hyperscalers have to amortize legacy infrastructure and margin expectations that neoclouds do not carry. Gartner projects that by 2030, neoclouds will capture 20% of the $267 billion AI cloud market, and Vultr's own analysis says at least 80% of GPU market share by end of 2026 will be held by a small group of scaled neocloud providers. Now zoom into Nebius specifically, because it is the most interesting publicly traded proxy for this trade. Nebius is the infrastructure arm of the former Yandex Russia's equivalent of Google rebuilt from the ground up after Russia's invasion of Ukraine by Arkady Volozh and relisted on Nasdaq in October 2024. The team that built it already knew how to run internet-scale infrastructure at the lowest possible cost, which is exactly the operational DNA a neocloud requires. In Q1 2026, Nebius reported revenue of $399 million and already generating serious cash on a young business with revenue growing nearly eightfold year-over-year. Then in March 2026, Meta signed a five-year infrastructure agreement with Nebius worth up to $27 billion, $12 billion in committed dedicated GPU capacity deployments beginning early 2027, plus up to $15 billion more tied to Meta purchasing Nebius's unsold third-party capacity. The deal will be executed on one of the first large-scale deployments of Nvidia's Vera Rubin platform, the next-generation architecture after Blackwell making Nebius one of a tiny number of operators in the world with confirmed priority access to the most advanced AI hardware available. Following the contract, Nebius guided to $7 to $9 billion in annualized recurring revenue for 2026 representing 540% year-over-year growth. Chamath Palihapitiya point about the $100 billion capital moat is the bear case for new entrants and the bull case for incumbents. No one can afford to build the next CoreWeave or Nebius from scratch at current hardware and power costs. The companies that are already built, already contracted, and already deploying Nvidia's latest silicon have a moat that compounds with every GPU generation cycle because they get allocations first, they deploy fastest, and their customers re-sign rather than wait for a new operator that does not yet exist. Come join Milk Road Pro for our full breakdown, the complete neocloud competitive landscape, how to think about Nebius's valuation versus CoreWeave and AI entire thesis. Link below.

Chamath Palihapitiya just dropped the number that explains the entire AI infrastructure trade (Save this). A gigawatt of compute now costs $100 billion and when he started his Arizona data center project it was $4 to $5 billion, it has gone up 20x in a single investment cycle. The implication is not just that AI infrastructure is expensive but rather that the capital barrier to owning meaningful compute has become so high that only a handful of entities in the world can actually build it and the companies who got there early are sitting on what may be the most durable pricing power in the history of the technology industry. This is the neocloud trade. The neocloud market, purpose-built GPU cloud providers like CoreWeave, Nebius, and Lambda Labs was worth $35 billion in 2026 and is projected to reach $236 billion by 2031, compounding at 46% annually. For context, that is faster growth than cloud computing itself posted in its first decade. The reason is very simple, hyperscalers like AWS, Azure, and Google are building for everything, storage, databases, enterprise software, networking and their GPU pricing reflects the overhead of that full-stack infrastructure. Neoclouds build for one thing only, AI compute. The result is a 60% to 85% cost advantage on the same Nvidia silicon, bare metal H100s at $0.78 to $2.79 per GPU-hour on a neocloud versus $3.43 to $5.07 per GPU-hour on a hyperscaler. That spread does not close as AI demand scales but rather it widens, because hyperscalers have to amortize legacy infrastructure and margin expectations that neoclouds do not carry. Gartner projects that by 2030, neoclouds will capture 20% of the $267 billion AI cloud market, and Vultr's own analysis says at least 80% of GPU market share by end of 2026 will be held by a small group of scaled neocloud providers. Now zoom into Nebius specifically, because it is the most interesting publicly traded proxy for this trade. Nebius is the infrastructure arm of the former Yandex Russia's equivalent of Google rebuilt from the ground up after Russia's invasion of Ukraine by Arkady Volozh and relisted on Nasdaq in October 2024. The team that built it already knew how to run internet-scale infrastructure at the lowest possible cost, which is exactly the operational DNA a neocloud requires. In Q1 2026, Nebius reported revenue of $399 million and already generating serious cash on a young business with revenue growing nearly eightfold year-over-year. Then in March 2026, Meta signed a five-year infrastructure agreement with Nebius worth up to $27 billion, $12 billion in committed dedicated GPU capacity deployments beginning early 2027, plus up to $15 billion more tied to Meta purchasing Nebius's unsold third-party capacity. The deal will be executed on one of the first large-scale deployments of Nvidia's Vera Rubin platform, the next-generation architecture after Blackwell making Nebius one of a tiny number of operators in the world with confirmed priority access to the most advanced AI hardware available. Following the contract, Nebius guided to $7 to $9 billion in annualized recurring revenue for 2026 representing 540% year-over-year growth. Chamath Palihapitiya point about the $100 billion capital moat is the bear case for new entrants and the bull case for incumbents. No one can afford to build the next CoreWeave or Nebius from scratch at current hardware and power costs. The companies that are already built, already contracted, and already deploying Nvidia's latest silicon have a moat that compounds with every GPU generation cycle because they get allocations first, they deploy fastest, and their customers re-sign rather than wait for a new operator that does not yet exist. Come join Milk Road Pro for our full breakdown, the complete neocloud competitive landscape, how to think about Nebius's valuation versus CoreWeave and AI entire thesis. Link below.

Milk Road AI

139,047 次观看 • 3 个月前

Elon Musk just explained why the SpaceX IPO is an energy story and the energy constraint is why he believes space becomes the only viable path for AI to scale (Save this). The argument he is making is one of the most important and least understood things happening in technology right now. The United States currently consumes roughly 500 gigawatts of electricity on average. To double that capacity which is what continued AI expansion on the current terrestrial trajectory would eventually require would mean building as many power plants as currently exist in the entire country. He is not arguing that this is technically impossible, just that communities are not willing to accept it, that permitting timelines make it unrealistic, and that the hard ceiling on Earth based power generation means the expansion of AI compute will eventually hit a wall that no amount of capital can overcome on the ground. His observation is that in space, that wall does not exist. A solar panel in orbit produces roughly five times more power than the same panel on Earth, operates in continuous sunlight uninterrupted by weather or nighttime, and benefits from the vacuum of space as a completely passive cooling system meaning the two largest operating costs of any terrestrial data center, energy and cooling, are effectively eliminated. He then said that you could theoretically increase harnessed energy by a factor of one million and still be using less than a millionth of the sun's total energy output. This is the underlying physics of why SpaceX filed with the FCC to launch up to one million solar powered AI satellites, and why they described that constellation in their own filing as a first step toward becoming a Kardashev Type II civilization capable of harnessing the full power of the sun. To understand what makes this credible rather than visionary, you need to understand what SpaceX already controls that no other company on earth possesses. Starship, once operating at full cadence, can deliver 100 to 150 tons of payload to orbit per launch, at a target cost per kilogram that is an order of magnitude lower than any existing vehicle. Musk's stated ambition is to scale Starship to 10,000 to 30,000 launches per year, a frequency that would allow the deployment of orbital compute infrastructure at a pace that is currently unimaginable with any existing rocket. He told xAI staff earlier this year that achieving space-based AI at scale will eventually require manufacturing facilities on the moon, building solar panels and heat dissipation structures from lunar silicon and aluminum, and launching them into orbit from there rather than from Earth's surface because the moon's lower gravity makes the economics of launch dramatically more favorable. SpaceX's S-1 filing explicitly states that its launch capabilities could enable massive AI compute satellite constellations with the potential for millions of satellites for orbital data centers, with the first launch potentially occurring as soon as 2028. Google and Alphabet are already in advanced talks with SpaceX about deploying space-based data centers. Starcloud, a startup running Nvidia H100 GPUs in orbit, has already validated that high-performance AI inference workloads can operate in space, with plans to scale to five gigawatts of orbital compute power by 2035. This is why Musk believes the cost crossover happens in two to three years because SpaceX's launch cost trajectory intersects with the accelerating energy constraint on the ground in a way that makes space genuinely cheaper, faster, and less regulated at exactly the moment AI demand is hitting its hardest physical limits.

Elon Musk just explained why the SpaceX IPO is an energy story and the energy constraint is why he believes space becomes the only viable path for AI to scale (Save this). The argument he is making is one of the most important and least understood things happening in technology right now. The United States currently consumes roughly 500 gigawatts of electricity on average. To double that capacity which is what continued AI expansion on the current terrestrial trajectory would eventually require would mean building as many power plants as currently exist in the entire country. He is not arguing that this is technically impossible, just that communities are not willing to accept it, that permitting timelines make it unrealistic, and that the hard ceiling on Earth based power generation means the expansion of AI compute will eventually hit a wall that no amount of capital can overcome on the ground. His observation is that in space, that wall does not exist. A solar panel in orbit produces roughly five times more power than the same panel on Earth, operates in continuous sunlight uninterrupted by weather or nighttime, and benefits from the vacuum of space as a completely passive cooling system meaning the two largest operating costs of any terrestrial data center, energy and cooling, are effectively eliminated. He then said that you could theoretically increase harnessed energy by a factor of one million and still be using less than a millionth of the sun's total energy output. This is the underlying physics of why SpaceX filed with the FCC to launch up to one million solar powered AI satellites, and why they described that constellation in their own filing as a first step toward becoming a Kardashev Type II civilization capable of harnessing the full power of the sun. To understand what makes this credible rather than visionary, you need to understand what SpaceX already controls that no other company on earth possesses. Starship, once operating at full cadence, can deliver 100 to 150 tons of payload to orbit per launch, at a target cost per kilogram that is an order of magnitude lower than any existing vehicle. Musk's stated ambition is to scale Starship to 10,000 to 30,000 launches per year, a frequency that would allow the deployment of orbital compute infrastructure at a pace that is currently unimaginable with any existing rocket. He told xAI staff earlier this year that achieving space-based AI at scale will eventually require manufacturing facilities on the moon, building solar panels and heat dissipation structures from lunar silicon and aluminum, and launching them into orbit from there rather than from Earth's surface because the moon's lower gravity makes the economics of launch dramatically more favorable. SpaceX's S-1 filing explicitly states that its launch capabilities could enable massive AI compute satellite constellations with the potential for millions of satellites for orbital data centers, with the first launch potentially occurring as soon as 2028. Google and Alphabet are already in advanced talks with SpaceX about deploying space-based data centers. Starcloud, a startup running Nvidia H100 GPUs in orbit, has already validated that high-performance AI inference workloads can operate in space, with plans to scale to five gigawatts of orbital compute power by 2035. This is why Musk believes the cost crossover happens in two to three years because SpaceX's launch cost trajectory intersects with the accelerating energy constraint on the ground in a way that makes space genuinely cheaper, faster, and less regulated at exactly the moment AI demand is hitting its hardest physical limits.

Milk Road AI

12,738 次观看 • 3 个月前

Elon Musk just told the world his plan to own the one thing every AI on Earth depends on. The silicon. The entire AI industry is fighting over the same layer. Models. Parameters. Data. Benchmarks. All of it runs on chips none of them produce. Every frontier lab on the planet is building intelligence on a foundation they do not control, in a country they cannot influence, on an island they could not defend. Musk: “You design a chip, you fabricate the chip, you test the chip, you redesign the chip, and you fabricate it again. All under one roof.” That’s the Terafab. One building in Austin, Texas. No wafers shipped across the Pacific. No six-month tape-out cycles. No single point of geopolitical failure sitting in the Taiwan Strait. Total vertical sovereignty over silicon. But sovereignty is not the endgame. Sovereignty is what makes the endgame possible. Speed. Every fab on Earth optimizes for yield. For cost. For predictable output at massive scale. The Terafab optimizes for one variable only. Learning speed. That distinction will define the next era of compute. When your iteration cycle compresses from months to days, perfection on the first attempt becomes irrelevant. What matters is how fast you reach the fiftieth. That is the exact principle that made SpaceX untouchable. They did not build a superior rocket on the first try. They built a system where failure was cheap and iteration was relentless. And that system produced something no one else could match. Now transplant that into semiconductors. Musk: “New physics. Wild and crazy things.” He is not trying to out-manufacture TSMC. He is trying to make TSMC’s entire model a relic of a slower era. TSMC cannot take radical bets on unproven architecture. Their customers demand predictability. Their margins demand stability. Their entire empire is built on perfecting what already works. Elon’s model is built on trying what has never worked. Repeatedly. At near-zero cost. Until it does. When the cost of failure approaches zero, breakthroughs stop being accidents. They become mathematically inevitable. Now add the layer that makes this permanent. xAI builds frontier AI. That AI assists in chip design. Better chips accelerate the AI. Faster AI designs better chips. That is not a production line. That is a compounding feedback loop with no external dependency. And nobody else on Earth can run it. Intel has fabs but no frontier AI. NVIDIA designs but does not fabricate. TSMC fabricates but does not design. Google designs but outsources production. Every one of them has a structural gap they cannot close. Elon is closing his. The AI that architects the chip. The fab that forges it. The vehicles and robots that run on it. The data they generate. The AI that trains on that data to design the next generation. Full circle. No seams. No permission required from any government, company, or supply chain on Earth. That is not a factory. That is a self-improving system with a physical body. A structure that manufactures upgrades to its own capacity to think. That has never existed before. Not as a concept. Not as a metaphor. As a building. Every era of human civilization was defined by whoever mastered its most critical substrate. Land built empires. Iron built armies. Oil built superpowers. The next substrate is compute. And a single facility in Austin is being designed to produce it in a closed loop that no competitor can replicate, no government can embargo, and no market force can interrupt. That is not a factory announcement. That is the foundation of the first self-reinforcing intelligence monopoly in the history of this species. And the loop only needs to start once.

Elon Musk just told the world his plan to own the one thing every AI on Earth depends on. The silicon. The entire AI industry is fighting over the same layer. Models. Parameters. Data. Benchmarks. All of it runs on chips none of them produce. Every frontier lab on the planet is building intelligence on a foundation they do not control, in a country they cannot influence, on an island they could not defend. Musk: “You design a chip, you fabricate the chip, you test the chip, you redesign the chip, and you fabricate it again. All under one roof.” That’s the Terafab. One building in Austin, Texas. No wafers shipped across the Pacific. No six-month tape-out cycles. No single point of geopolitical failure sitting in the Taiwan Strait. Total vertical sovereignty over silicon. But sovereignty is not the endgame. Sovereignty is what makes the endgame possible. Speed. Every fab on Earth optimizes for yield. For cost. For predictable output at massive scale. The Terafab optimizes for one variable only. Learning speed. That distinction will define the next era of compute. When your iteration cycle compresses from months to days, perfection on the first attempt becomes irrelevant. What matters is how fast you reach the fiftieth. That is the exact principle that made SpaceX untouchable. They did not build a superior rocket on the first try. They built a system where failure was cheap and iteration was relentless. And that system produced something no one else could match. Now transplant that into semiconductors. Musk: “New physics. Wild and crazy things.” He is not trying to out-manufacture TSMC. He is trying to make TSMC’s entire model a relic of a slower era. TSMC cannot take radical bets on unproven architecture. Their customers demand predictability. Their margins demand stability. Their entire empire is built on perfecting what already works. Elon’s model is built on trying what has never worked. Repeatedly. At near-zero cost. Until it does. When the cost of failure approaches zero, breakthroughs stop being accidents. They become mathematically inevitable. Now add the layer that makes this permanent. xAI builds frontier AI. That AI assists in chip design. Better chips accelerate the AI. Faster AI designs better chips. That is not a production line. That is a compounding feedback loop with no external dependency. And nobody else on Earth can run it. Intel has fabs but no frontier AI. NVIDIA designs but does not fabricate. TSMC fabricates but does not design. Google designs but outsources production. Every one of them has a structural gap they cannot close. Elon is closing his. The AI that architects the chip. The fab that forges it. The vehicles and robots that run on it. The data they generate. The AI that trains on that data to design the next generation. Full circle. No seams. No permission required from any government, company, or supply chain on Earth. That is not a factory. That is a self-improving system with a physical body. A structure that manufactures upgrades to its own capacity to think. That has never existed before. Not as a concept. Not as a metaphor. As a building. Every era of human civilization was defined by whoever mastered its most critical substrate. Land built empires. Iron built armies. Oil built superpowers. The next substrate is compute. And a single facility in Austin is being designed to produce it in a closed loop that no competitor can replicate, no government can embargo, and no market force can interrupt. That is not a factory announcement. That is the foundation of the first self-reinforcing intelligence monopoly in the history of this species. And the loop only needs to start once.

Dustin

47,620 次观看 • 2 个月前

Announcing DreamDojo: our open-source, interactive world model that takes robot motor controls and generates the future in pixels. No engine, no meshes, no hand-authored dynamics. It's Simulation 2.0. Time for robotics to take the bitter lesson pill. Real-world robot learning is bottlenecked by time, wear, safety, and resets. If we want Physical AI to move at pretraining speed, we need a simulator that adapts to pretraining scale with as little human engineering as possible. Our key insights: (1) human egocentric videos are a scalable source of first-person physics; (2) latent actions make them "robot-readable" across different hardware; (3) real-time inference unlocks live teleop, policy eval, and test-time planning *inside* a dream. We pre-train on 44K hours of human videos: cheap, abundant, and collected with zero robot-in-the-loop. Humans have already explored the combinatorics: we grasp, pour, fold, assemble, fail, retry—across cluttered scenes, shifting viewpoints, changing light, and hour-long task chains—at a scale no robot fleet could match. The missing piece: these videos have no action labels. So we introduce latent actions: a unified representation inferred directly from videos that captures "what changed between world states" without knowing the underlying hardware. This lets us train on any first-person video as if it came with motor commands attached. As a result, DreamDojo generalizes zero-shot to objects and environments never seen in any robot training set, because humans saw them first. Next, we post-train onto each robot to fit its specific hardware. Think of it as separating "how the world looks and behaves" from "how this particular robot actuates." The base model follows the general physical rules, then "snaps onto" the robot's unique mechanics. It's kind of like loading a new character and scene assets into Unreal Engine, but done through gradient descent and generalizes far beyond the post-training dataset. A world simulator is only useful if it runs fast enough to close the loop. We train a real-time version of DreamDojo that runs at 10 FPS, stable for over a minute of continuous rollout. This unlocks exciting possibilities: - Live teleoperation *inside* a dream. Connect a VR controller, stream actions into DreamDojo, and teleop a virtual robot in real time. We demo this on Unitree G1 with a PICO headset and one RTX 5090. - Policy evaluation. You can benchmark a policy checkpoint in DreamDojo instead of the real world. The simulated success rates strongly correlate with real-world results - accurate enough to rank checkpoints without burning a single motor. - Model-based planning. Sample multiple action proposals → simulate them all in parallel → pick the best future. Gains +17% real-world success out of the box on a fruit packing task. We open-source everything!! Weights, code, post-training dataset, eval set, and whitepaper with tons of details to reproduce. DreamDojo is based on NVIDIA Cosmos, which is open-weight too. 2026 is the year of World Models for physical AI. We want you to build with us. Happy scaling! Links in thread:

Announcing DreamDojo: our open-source, interactive world model that takes robot motor controls and generates the future in pixels. No engine, no meshes, no hand-authored dynamics. It's Simulation 2.0. Time for robotics to take the bitter lesson pill. Real-world robot learning is bottlenecked by time, wear, safety, and resets. If we want Physical AI to move at pretraining speed, we need a simulator that adapts to pretraining scale with as little human engineering as possible. Our key insights: (1) human egocentric videos are a scalable source of first-person physics; (2) latent actions make them "robot-readable" across different hardware; (3) real-time inference unlocks live teleop, policy eval, and test-time planning inside a dream. We pre-train on 44K hours of human videos: cheap, abundant, and collected with zero robot-in-the-loop. Humans have already explored the combinatorics: we grasp, pour, fold, assemble, fail, retry—across cluttered scenes, shifting viewpoints, changing light, and hour-long task chains—at a scale no robot fleet could match. The missing piece: these videos have no action labels. So we introduce latent actions: a unified representation inferred directly from videos that captures "what changed between world states" without knowing the underlying hardware. This lets us train on any first-person video as if it came with motor commands attached. As a result, DreamDojo generalizes zero-shot to objects and environments never seen in any robot training set, because humans saw them first. Next, we post-train onto each robot to fit its specific hardware. Think of it as separating "how the world looks and behaves" from "how this particular robot actuates." The base model follows the general physical rules, then "snaps onto" the robot's unique mechanics. It's kind of like loading a new character and scene assets into Unreal Engine, but done through gradient descent and generalizes far beyond the post-training dataset. A world simulator is only useful if it runs fast enough to close the loop. We train a real-time version of DreamDojo that runs at 10 FPS, stable for over a minute of continuous rollout. This unlocks exciting possibilities: - Live teleoperation inside a dream. Connect a VR controller, stream actions into DreamDojo, and teleop a virtual robot in real time. We demo this on Unitree G1 with a PICO headset and one RTX 5090. - Policy evaluation. You can benchmark a policy checkpoint in DreamDojo instead of the real world. The simulated success rates strongly correlate with real-world results - accurate enough to rank checkpoints without burning a single motor. - Model-based planning. Sample multiple action proposals → simulate them all in parallel → pick the best future. Gains +17% real-world success out of the box on a fruit packing task. We open-source everything!! Weights, code, post-training dataset, eval set, and whitepaper with tons of details to reproduce. DreamDojo is based on NVIDIA Cosmos, which is open-weight too. 2026 is the year of World Models for physical AI. We want you to build with us. Happy scaling! Links in thread:

Jim Fan

228,898 次观看 • 7 个月前

$AMD Massive Rotation from $NVDA $INTC🧵 Not Financial Advice! DYOR! 5-10 minutes before the bell today, last trading day of May 2026, massive rotation out of $INTC and $NVDA into $AMD. I wrote this thread this morning on what $TSM said on Energy Efficiency is now TOP Priotity and why AMD is the biggest winner. Of course I did not have influence on this rebalancing, I was just pointing out why Dr. Su saw this coming years ago. (Check the picture to understand more). I been talking about Agentic AI for like 3-4 years now. OpenClaw broke the CPU:GPU Ratio 1:4 narrative to 1:1 to 5:1 in late Jan and Feb 2026. I will link various threads where you can understand the full picture from supply chain, to TSMC expansion, and different Wafer Ratio for EPYC Venice and MI455X. Energy efficiency is a structural, long-term driver behind institutional rotation from $NVDA and $INTC into $AMD (with spillover strength in $AVGO for complementary networking/custom silicon). This isn't just short-term rebalancing, it's a massive bet on the shift from AI training (performance-at-any-cost) to inference, deployment, and embodied/agentic systems (where total cost of ownership, power draw, and scalability dominate). Precisely What I been writing about $AMD for years now, probably at least more than 5,000 threads.This is the FOMO from Institutions to own $AMD. Do know that AMD is the least owned Semi Stock among vs Peers. AI infrastructure is moving beyond massive training clusters to widespread inference for Agentic AI (running models 24/7) and embodied AI (robots, autonomous agents, edge devices). These workloads prioritize: ~Tokens-per-watt and performance-per-watt ~Lower total power consumption for data centers facing grid constraints ~Better economics at scale (cost-per-token, TCO) ~Thermal and power efficiency for on-device/robotics use Hyperscalers are now thinking more about Margin, Profitability, and $/M Tokens At $516/share. AMD Fwd PEG Ratio is still 35/100+= 0.35 AKA very cheap IMO for the growth and potential. A. Why institutions rotated out of $NVDA? Because Agentic AI is going to dominated by CPUs for years to come, moving violently to 5-10-20:1 CPU:GPU Ratio as enterprises are demanding more than 10-20 agents to run tasks. Now, that does not mean training is going away, Inference is just going to grow much faster. B. Why instiutitons rotated out of $INTC? Because AMD x86 unit share is only at 30-31% but Revenue share is already at 46.2% according to Mercury Research. And Dr. Su wants 50-60% market share, and that would mean 60-70%+ Revenue share where the CPUs TAM Is now already at $200B in 2026 and projected to be $500B by 2030. C. Why $AMD? Because AMD secured meaningful 2nm Capacity, Advanced Packaging and Memory through 2027-2028. And TSMC is expanding 2 primary 2nm Fabs toward 60-65k WPM each, and speeding up 5 2nm Fabs in Taiwan. With total up to 12 2nm Fabs through 2027/2028. 2nm Capacity is expected to be 140k+ WPM toward end of 2026, and 220-240k WPM by end of 2027. Apple has secured 35-45k WPM. And AMD does not have to worry about allocation competition until late 2027 from $AVGO for $META and $GOOGL(This may change) D. Agentic AI will evolve to 24/7 Autonomous Agent, and that will become the foundational layer for Robotic or Physical AI. Agentic AI (autonomous systems that plan, reason, use tools, self-correct, pursue long-horizon goals, and adapt) provides the high-level cognitive architecture. It turns raw perception and low-level control into useful, general-purpose behavior in the physical world. Physical AI (or Embodied AI) refers to AI that senses, understands, and acts directly in the real world through robots, actuators, and sensors. Agentic capabilities are what make this scalable and useful beyond narrow, scripted tasks. Reactive/programmed machines → To proactive, goal-oriented autonomous agents. How does this work? Autonomous Agent layer is the brain ~Vision-Language-Action models or robotics foundation models. ~Agentic loops: Planning, chain-of-thought reasoning, reflection, tool use (simulators, APIs), multi-step task decomposition. ~Persistent 24/7 operation with Memory, world modeling, continuous learning. Institutions may not like $AMD from 2022-2025, but they cannot stop this evolution and it is inevitable. Part of my main thesis for AMD to get to $5 Trillion Market Cap Long Term. Conclusion: Institutions are rotating capital toward AMD not merely for tactical rebalancing, but because Dr. Lisa Su and her team anticipated this exact inflection years in advance and have been methodically engineering AMD’s platform to dominate it. Dr. Su has long championed the convergence of Agentic AI as the high-level cognitive foundation for Physical AI and robotics. As far back as her 2023/2024 CES keynote and earlier strategic commentary, she described Physical AI (including humanoid robotics and edge autonomy) as “the next big thing”; a natural extension of agentic workflows moving from digital reasoning to real-world action. She emphasized that enabling persistent, 24/7 autonomous agents requires a full-stack approach: high-performance CPUs for orchestration and motion control, dedicated accelerators for real-time vision and multimodal inference, and open software ecosystems for rapid development. This vision aligns precisely with the structural drivers we’ve discussed. As AI shifts from training to massive-scale inference and embodiment, energy efficiency, total cost of ownership, and heterogeneous compute become first-order advantages. AMD’s Instinct MI350/MI355 series, Ryzen AI Embedded processors, and EPYC platforms deliver superior performance-per-watt and balanced CPU + GPU + NPU integration ideal for power-constrained robots that must run sophisticated agentic reasoning loops without excessive thermal or battery drain. Dr. Su has repeatedly highlighted the rising importance of CPUs in agentic systems (moving toward 1:1 or even CPU-heavy ratios with GPUs), positioning AMD’s strengths in orchestration, memory handling, and efficiency as critical for the next phase of growth. AMD is engineered for the deployment realities of embodied agents: scalable, efficient, and deployable at the edge and in physical systems. The institutional flows out of NVDA and INTC into AMD reflect recognition of this prepared leadership. Dr. Su didn’t just see the future of Agentic AI powering robotics, she has spent years building the silicon, software, and partnerships to make it practical and economically viable. This rotation signals confidence that the companies best positioned for the physical, always-on intelligence layer will capture the highest-volume opportunities in the coming decade. Not Financial Advice! DYOR!

$AMD Massive Rotation from $NVDA $INTC🧵 Not Financial Advice! DYOR! 5-10 minutes before the bell today, last trading day of May 2026, massive rotation out of $INTC and $NVDA into $AMD. I wrote this thread this morning on what $TSM said on Energy Efficiency is now TOP Priotity and why AMD is the biggest winner. Of course I did not have influence on this rebalancing, I was just pointing out why Dr. Su saw this coming years ago. (Check the picture to understand more). I been talking about Agentic AI for like 3-4 years now. OpenClaw broke the CPU:GPU Ratio 1:4 narrative to 1:1 to 5:1 in late Jan and Feb 2026. I will link various threads where you can understand the full picture from supply chain, to TSMC expansion, and different Wafer Ratio for EPYC Venice and MI455X. Energy efficiency is a structural, long-term driver behind institutional rotation from $NVDA and $INTC into $AMD (with spillover strength in $AVGO for complementary networking/custom silicon). This isn't just short-term rebalancing, it's a massive bet on the shift from AI training (performance-at-any-cost) to inference, deployment, and embodied/agentic systems (where total cost of ownership, power draw, and scalability dominate). Precisely What I been writing about $AMD for years now, probably at least more than 5,000 threads.This is the FOMO from Institutions to own $AMD. Do know that AMD is the least owned Semi Stock among vs Peers. AI infrastructure is moving beyond massive training clusters to widespread inference for Agentic AI (running models 24/7) and embodied AI (robots, autonomous agents, edge devices). These workloads prioritize: ~Tokens-per-watt and performance-per-watt ~Lower total power consumption for data centers facing grid constraints ~Better economics at scale (cost-per-token, TCO) ~Thermal and power efficiency for on-device/robotics use Hyperscalers are now thinking more about Margin, Profitability, and $/M Tokens At $516/share. AMD Fwd PEG Ratio is still 35/100+= 0.35 AKA very cheap IMO for the growth and potential. A. Why institutions rotated out of $NVDA? Because Agentic AI is going to dominated by CPUs for years to come, moving violently to 5-10-20:1 CPU:GPU Ratio as enterprises are demanding more than 10-20 agents to run tasks. Now, that does not mean training is going away, Inference is just going to grow much faster. B. Why instiutitons rotated out of $INTC? Because AMD x86 unit share is only at 30-31% but Revenue share is already at 46.2% according to Mercury Research. And Dr. Su wants 50-60% market share, and that would mean 60-70%+ Revenue share where the CPUs TAM Is now already at $200B in 2026 and projected to be $500B by 2030. C. Why $AMD? Because AMD secured meaningful 2nm Capacity, Advanced Packaging and Memory through 2027-2028. And TSMC is expanding 2 primary 2nm Fabs toward 60-65k WPM each, and speeding up 5 2nm Fabs in Taiwan. With total up to 12 2nm Fabs through 2027/2028. 2nm Capacity is expected to be 140k+ WPM toward end of 2026, and 220-240k WPM by end of 2027. Apple has secured 35-45k WPM. And AMD does not have to worry about allocation competition until late 2027 from $AVGO for $META and $GOOGL(This may change) D. Agentic AI will evolve to 24/7 Autonomous Agent, and that will become the foundational layer for Robotic or Physical AI. Agentic AI (autonomous systems that plan, reason, use tools, self-correct, pursue long-horizon goals, and adapt) provides the high-level cognitive architecture. It turns raw perception and low-level control into useful, general-purpose behavior in the physical world. Physical AI (or Embodied AI) refers to AI that senses, understands, and acts directly in the real world through robots, actuators, and sensors. Agentic capabilities are what make this scalable and useful beyond narrow, scripted tasks. Reactive/programmed machines → To proactive, goal-oriented autonomous agents. How does this work? Autonomous Agent layer is the brain ~Vision-Language-Action models or robotics foundation models. ~Agentic loops: Planning, chain-of-thought reasoning, reflection, tool use (simulators, APIs), multi-step task decomposition. ~Persistent 24/7 operation with Memory, world modeling, continuous learning. Institutions may not like $AMD from 2022-2025, but they cannot stop this evolution and it is inevitable. Part of my main thesis for AMD to get to $5 Trillion Market Cap Long Term. Conclusion: Institutions are rotating capital toward AMD not merely for tactical rebalancing, but because Dr. Lisa Su and her team anticipated this exact inflection years in advance and have been methodically engineering AMD’s platform to dominate it. Dr. Su has long championed the convergence of Agentic AI as the high-level cognitive foundation for Physical AI and robotics. As far back as her 2023/2024 CES keynote and earlier strategic commentary, she described Physical AI (including humanoid robotics and edge autonomy) as “the next big thing”; a natural extension of agentic workflows moving from digital reasoning to real-world action. She emphasized that enabling persistent, 24/7 autonomous agents requires a full-stack approach: high-performance CPUs for orchestration and motion control, dedicated accelerators for real-time vision and multimodal inference, and open software ecosystems for rapid development. This vision aligns precisely with the structural drivers we’ve discussed. As AI shifts from training to massive-scale inference and embodiment, energy efficiency, total cost of ownership, and heterogeneous compute become first-order advantages. AMD’s Instinct MI350/MI355 series, Ryzen AI Embedded processors, and EPYC platforms deliver superior performance-per-watt and balanced CPU + GPU + NPU integration ideal for power-constrained robots that must run sophisticated agentic reasoning loops without excessive thermal or battery drain. Dr. Su has repeatedly highlighted the rising importance of CPUs in agentic systems (moving toward 1:1 or even CPU-heavy ratios with GPUs), positioning AMD’s strengths in orchestration, memory handling, and efficiency as critical for the next phase of growth. AMD is engineered for the deployment realities of embodied agents: scalable, efficient, and deployable at the edge and in physical systems. The institutional flows out of NVDA and INTC into AMD reflect recognition of this prepared leadership. Dr. Su didn’t just see the future of Agentic AI powering robotics, she has spent years building the silicon, software, and partnerships to make it practical and economically viable. This rotation signals confidence that the companies best positioned for the physical, always-on intelligence layer will capture the highest-volume opportunities in the coming decade. Not Financial Advice! DYOR!

Mike

104,109 次观看 • 3 个月前

In a newly released technical update, SpaceX's leadership team, which includes communications manager Dan Huot, Director of Satellite Engineering Ian Dahl, and CEO Elon Musk, detailed a highly ambitious infrastructure roadmap to design, manufacture, and operate specialized artificial intelligence computing satellites at scale. Positioned as a major strategic pillar to dramatically elevate civilizational energy and processing capacity on the Kardashev scale, this strategy moves past traditional communications architectures into massive orbital server arrays. Here is the complete breakdown of the core technologies and timelines driving this space-based intelligence revolution: 🛰️ AI1 satellite power and compute capacity Ian Dahl and Elon Musk introduced the baseline performance targets for the first-generation AI1 satellite, explaining how its custom hardware is engineered to operate like an orbital data center server rack. Ian Dahl noted that their direct operational experience with xAI guided them to target a 150-kilowatt peak power capacity. To manage active machine learning workloads continuously, Elon Musk explained that the satellite is optimized to maintain a sustained average compute power envelope of 120 kilowatts, which directly mirrors the real-world performance of a terrestrial NVIDIA server rack. The official presentation slides outline several key operational metrics for this payload configuration: ⚡ The custom architecture delivers a 150 kW peak compute payload. 🔋 The system maintains a 120 kW sustained average compute payload under active workloads. ⚖️ The hardware achieves a highly optimized power-to-weight density of 70 kW per ton. 🔄 The layout features a completely interchangeable compute provider design. "We thought that the right place to start is around the 150 kilowatt peak power level. But as we look at the workloads with our experience with xAI, we see that we can support about 120 kilowatts of average compute. The 150 kilowatt peak power level roughly matches what, say, an NVIDIA GV300 rack would do. A more reasonable operating envelope would be around 120 kilowatts average power, but it can peak up to 150. So it is basically thinking about it as a rack of compute in space." --- 📐 AI1 satellite dimensions and thermal efficiency specs Elon Musk detailed the physical layout of the AI1 satellite, highlighting the massive dimensions required to accommodate its immense power and cooling hardware. He shared specific design criteria, explaining that the engineering relies on a custom 150 kW solar array paired with a high-capacity deployable liquid radiator thermal management system. The technical specifications of this vehicle layout include: 📏 The structural frame features a massive 70-meter wingspan. ↕️ The vehicle spans a total deployed height of 20 meters. ☀️ The onboard solar array delivers an efficiency of 250 W/m² using technology manufactured in Bastrop, Texas. 🌡️ The thermal system utilizes a 110 m² deployable liquid radiator to cleanly dump waste heat. 🔄 The cooling architecture incorporates redundant pumping loops for mission safety. 🛡️ The exterior contains integrated micrometeoroid shielding to protect the fluid lines. 🧭 The double-sided radiators achieve a dissipation rate of 1400 watts per square meter while remaining oriented knife-edge to the sun. "The assumptions here are 250 watts per square meter for the solar array and about 1400 watts per square meter for the radiators. The radiators are double-sided, radiating on both sides, and they're oriented knife-edge to the sun. They have about a 70-meter wingspan, so these are fairly large." --- 🧩 Simplified design architecture built on Starlink V3 tech Elon Musk explained that despite the satellite's imposing size, its internal architecture is fundamentally much simpler than a standard Starlink satellite. Because it lacks heavy phased array and parabolic communications antennas, the entire vehicle layout is completely streamlined around a few essential structural modules: 🎛️ The hardware framework is arranged around a centralized compute module. ☀️ Large deployable solar arrays extend outward to capture orbital energy. 🌡️ A deployable liquid-radiator thermal management system controls active operational temperatures. 🔄 The engineering team heavily leverages the component evolution and manufacturing experience gained from developing the Starlink V3 vehicle platform. "The AI satellite is actually much simpler than a Starlink satellite. A Starlink satellite has gigantic phased array antennas, parabolic antennas, and a lot of laser links, making it much more complicated. An AI satellite is essentially a lot of solar cells, a radiator, and you still need some laser links, but you don't have all of the super complex antennas that you have on a Starlink satellite. A lot of this is technology we've already made for the Starlink V3 satellites." --- 🔌 Interchangeable compute reference designs and high connectivity Elon Musk outlined a modular hardware approach for the satellite's payload, allowing it to house a variety of industry-standard processing units depending on client requirements. This interchangeable compute rack is supported by a high-bandwidth connectivity loop that links separate orbital units together or transmits data directly back to Earth. The core network parameters include: 🧠 Reference designs are fully established to seamlessly accommodate NVIDIA Reuben chips. 💾 The system architecture is built to support alternative setups using NVIDIA GB300 chips. 💻 Custom hardware layouts are explicitly designed to integrate Google TPUs. 🌐 The onboard communications setup delivers roughly 1 terabit of laser link connectivity. ⏱️ The network closes the communication loop directly with the main Starlink constellation at an ultra-low latency of only 3 milliseconds. "Our current reference design is for NVIDIA Reuben chips, or it could be either GB300 or Reuben chips. We'll also have a reference design for TPUs. Essentially, you can put up any existing chips into orbit. There would also be probably something on the order of a terabit of laser link connectivity from the satellite. Then you can connect these racks of compute to each other by the laser links or directly to the Starlink constellations. Light travels 300 kilometers per millisecond, so that's about three milliseconds away." --- 🏭 The "gigasat" AI satellite and solar production hub in Bastrop, Texas Dan Huot highlighted that the primary production hub for this entire hardware ecosystem is anchored at their sprawling complex in Bastrop, Texas, officially designated as the Gigasat factory. Elon Musk verified that construction is already actively underway on the solar manufacturing facility to feed the project's supply line, with plans moving forward to construct the adjacent AI satellite assembly lines. The physical footprint and timeline of this manufacturing hub are defined by the following benchmarks: 🗺️ The company has over 1,000 acres of land currently owned or under contract for the site. 🏢 The manufacturing complex boasts a massive structural building potential exceeding 11 million square feet. ⚙️ The facility will vertically integrate production to manufacture solar ingots, wafers, solar cells, and completed AI satellites. 📅 Both the solar and AI satellite production lines are targeted to be operational at a viable volume by the end of next year. "We're going to be building a lot of satellites and we're going to be building them here in Bastrop. We already have the solar manufacturing facility under construction, and then we will be building out the AI sat production building soon. We expect to have the AI sat production, the solar production, and all of that operating at some reasonable volume by the end of next year." --- 🏢 The 100-million-square-foot "terafab" chip factory Elon Musk revealed a massive, long-term scaling strategy to build an immense chip manufacturing facility dubbed the "terafab" to completely bypass global semiconductor volume constraints. This manufacturing infrastructure is designed to transition the company into next-generation industrial scaling by producing highly specialized computing components at an unprecedented volume. The scale of this infrastructure project is defined by several extraordinary engineering and production benchmarks: 🏭 The colossal factory is projected to span approximately 100 million square feet, making it ten times larger than the current Tesla Gigafactory Texas. ⚡ The facility is structurally engineered to achieve a massive manufacturing output of 1 terawatt per year once fully operational. 📦 This unprecedented physical footprint provides the capacity required to manufacture 1 billion full-reticle equivalent chips annually. 🔌 Each individual chip manufactured by the facility is designed to run at a power capacity of 1 kilowatt. 🇺🇸 The total scaled output of the facility represents an energy footprint that is exactly double the current annual electricity consumption of the entire United States. "In order to get to the next order of magnitude, you need a gigantic chip factory. To give you a sense of scale here, we expect that the terafab is going to be around 100 million square feet, which is 10 times the size of the Tesla Gigafactory Texas. From a logic die standpoint, that's like having a billion chips per year with a kilowatt per reticle, scaling to a terawatt per year. That is twice the current electricity consumption of the United States." --- 📶 Next-generation high-volume Starlink terminals Dan Huot and Elon Musk introduced their next-generation Starlink user terminals, which have been redesigned specifically to achieve massive manufacturing throughput. Elon Musk pointed out that these newer models will be produced in vastly higher volumes than current hardware designs to fulfill their long-term global deployment targets: 📈 The upgraded user hardware is manufactured at a much higher volume capacity than existing units. 🌍 The company's ultimate target is to successfully deploy a few hundred million of these next-generation terminals worldwide. "In fact, these are the new Starlink terminals, which we made in much higher volume than the current terminals. Ultimately, we think there's probably going to be a few hundred million Starlink terminals out there." --- 📈 Aspirational timeline for orbital AI compute scaling Elon Musk laid out an ambitious, multi-year execution timeline detailing how the company plans to progressively scale space-based processing power. The roadmap targets an initial run-rate by the end of next year and sets an aggressive pace to increase total operational capacity sequentially through a structured, multi-phase timeline: 1️⃣ The initial target aims to hit an annualized run-rate of 1 gigawatt of space AI compute by the end of next year. 2️⃣ The capacity scales to an annualized rate of 10 gigawatts within the next two and a half years. 3️⃣ The operational envelope expands to reach 100 gigawatts in three and a half years. 4️⃣ The long-term deployment plan scales directly to a full terawatt capacity per year using the output of the terafab. "The goal is to get to roughly an annualized rate of a gigawatt per year by the end of next year in terms of space AI compute. Then aspirationally, we want to scale that by an order of magnitude per year. In two and a half years, hitting an annualized rate of 10 gigawatts a year in space, and in three and a half years, maybe a hundred gigawatts, going beyond that with the terafab to scale to a terawatt per year." --- 🌕 Ultimate scaling via lunar production and mass drivers Elon Musk explained that scaling three orders of magnitude past a single terawatt forces a transition completely off-planet to avoid the logistical penalty of Earth's deep gravity well. The vision relies on establishing manufacturing infrastructure directly on the moon to leverage localized resource loops and zero-atmosphere physics: 🌙 The company plans to establish localized raw production lines on the moon to fabricate solar panels, photovoltaics, and radiators from lunar materials. ⚡ Manufacturing components locally avoids the massive fuel and mass penalties of transporting heavy structural materials from Earth. 🧲 Because the moon has no atmosphere and only one-sixth of Earth's gravity, the facility will utilize an electromagnetic mass driver to launch completed satellites. 🚀 Operating essentially as a linear electric motor rail gun, this mechanism will shoot fully assembled AI satellites straight into deep space without relying on chemical rockets. "The only way that we can really see that you can achieve that is on the moon with a mass driver, essentially where you do local production of photovoltaics, solar panels, and radiators on the moon. Because the moon has no atmosphere and only one-sixth Earth's gravity, you can accelerate the AI satellites into deep space without a rocket. You can basically shoot them into space using an electromagnetic gun, like a rail gun type—it's basically a linear electric motor."

In a newly released technical update, SpaceX's leadership team, which includes communications manager Dan Huot, Director of Satellite Engineering Ian Dahl, and CEO Elon Musk, detailed a highly ambitious infrastructure roadmap to design, manufacture, and operate specialized artificial intelligence computing satellites at scale. Positioned as a major strategic pillar to dramatically elevate civilizational energy and processing capacity on the Kardashev scale, this strategy moves past traditional communications architectures into massive orbital server arrays. Here is the complete breakdown of the core technologies and timelines driving this space-based intelligence revolution: 🛰️ AI1 satellite power and compute capacity Ian Dahl and Elon Musk introduced the baseline performance targets for the first-generation AI1 satellite, explaining how its custom hardware is engineered to operate like an orbital data center server rack. Ian Dahl noted that their direct operational experience with xAI guided them to target a 150-kilowatt peak power capacity. To manage active machine learning workloads continuously, Elon Musk explained that the satellite is optimized to maintain a sustained average compute power envelope of 120 kilowatts, which directly mirrors the real-world performance of a terrestrial NVIDIA server rack. The official presentation slides outline several key operational metrics for this payload configuration: ⚡ The custom architecture delivers a 150 kW peak compute payload. 🔋 The system maintains a 120 kW sustained average compute payload under active workloads. ⚖️ The hardware achieves a highly optimized power-to-weight density of 70 kW per ton. 🔄 The layout features a completely interchangeable compute provider design. "We thought that the right place to start is around the 150 kilowatt peak power level. But as we look at the workloads with our experience with xAI, we see that we can support about 120 kilowatts of average compute. The 150 kilowatt peak power level roughly matches what, say, an NVIDIA GV300 rack would do. A more reasonable operating envelope would be around 120 kilowatts average power, but it can peak up to 150. So it is basically thinking about it as a rack of compute in space." --- 📐 AI1 satellite dimensions and thermal efficiency specs Elon Musk detailed the physical layout of the AI1 satellite, highlighting the massive dimensions required to accommodate its immense power and cooling hardware. He shared specific design criteria, explaining that the engineering relies on a custom 150 kW solar array paired with a high-capacity deployable liquid radiator thermal management system. The technical specifications of this vehicle layout include: 📏 The structural frame features a massive 70-meter wingspan. ↕️ The vehicle spans a total deployed height of 20 meters. ☀️ The onboard solar array delivers an efficiency of 250 W/m² using technology manufactured in Bastrop, Texas. 🌡️ The thermal system utilizes a 110 m² deployable liquid radiator to cleanly dump waste heat. 🔄 The cooling architecture incorporates redundant pumping loops for mission safety. 🛡️ The exterior contains integrated micrometeoroid shielding to protect the fluid lines. 🧭 The double-sided radiators achieve a dissipation rate of 1400 watts per square meter while remaining oriented knife-edge to the sun. "The assumptions here are 250 watts per square meter for the solar array and about 1400 watts per square meter for the radiators. The radiators are double-sided, radiating on both sides, and they're oriented knife-edge to the sun. They have about a 70-meter wingspan, so these are fairly large." --- 🧩 Simplified design architecture built on Starlink V3 tech Elon Musk explained that despite the satellite's imposing size, its internal architecture is fundamentally much simpler than a standard Starlink satellite. Because it lacks heavy phased array and parabolic communications antennas, the entire vehicle layout is completely streamlined around a few essential structural modules: 🎛️ The hardware framework is arranged around a centralized compute module. ☀️ Large deployable solar arrays extend outward to capture orbital energy. 🌡️ A deployable liquid-radiator thermal management system controls active operational temperatures. 🔄 The engineering team heavily leverages the component evolution and manufacturing experience gained from developing the Starlink V3 vehicle platform. "The AI satellite is actually much simpler than a Starlink satellite. A Starlink satellite has gigantic phased array antennas, parabolic antennas, and a lot of laser links, making it much more complicated. An AI satellite is essentially a lot of solar cells, a radiator, and you still need some laser links, but you don't have all of the super complex antennas that you have on a Starlink satellite. A lot of this is technology we've already made for the Starlink V3 satellites." --- 🔌 Interchangeable compute reference designs and high connectivity Elon Musk outlined a modular hardware approach for the satellite's payload, allowing it to house a variety of industry-standard processing units depending on client requirements. This interchangeable compute rack is supported by a high-bandwidth connectivity loop that links separate orbital units together or transmits data directly back to Earth. The core network parameters include: 🧠 Reference designs are fully established to seamlessly accommodate NVIDIA Reuben chips. 💾 The system architecture is built to support alternative setups using NVIDIA GB300 chips. 💻 Custom hardware layouts are explicitly designed to integrate Google TPUs. 🌐 The onboard communications setup delivers roughly 1 terabit of laser link connectivity. ⏱️ The network closes the communication loop directly with the main Starlink constellation at an ultra-low latency of only 3 milliseconds. "Our current reference design is for NVIDIA Reuben chips, or it could be either GB300 or Reuben chips. We'll also have a reference design for TPUs. Essentially, you can put up any existing chips into orbit. There would also be probably something on the order of a terabit of laser link connectivity from the satellite. Then you can connect these racks of compute to each other by the laser links or directly to the Starlink constellations. Light travels 300 kilometers per millisecond, so that's about three milliseconds away." --- 🏭 The "gigasat" AI satellite and solar production hub in Bastrop, Texas Dan Huot highlighted that the primary production hub for this entire hardware ecosystem is anchored at their sprawling complex in Bastrop, Texas, officially designated as the Gigasat factory. Elon Musk verified that construction is already actively underway on the solar manufacturing facility to feed the project's supply line, with plans moving forward to construct the adjacent AI satellite assembly lines. The physical footprint and timeline of this manufacturing hub are defined by the following benchmarks: 🗺️ The company has over 1,000 acres of land currently owned or under contract for the site. 🏢 The manufacturing complex boasts a massive structural building potential exceeding 11 million square feet. ⚙️ The facility will vertically integrate production to manufacture solar ingots, wafers, solar cells, and completed AI satellites. 📅 Both the solar and AI satellite production lines are targeted to be operational at a viable volume by the end of next year. "We're going to be building a lot of satellites and we're going to be building them here in Bastrop. We already have the solar manufacturing facility under construction, and then we will be building out the AI sat production building soon. We expect to have the AI sat production, the solar production, and all of that operating at some reasonable volume by the end of next year." --- 🏢 The 100-million-square-foot "terafab" chip factory Elon Musk revealed a massive, long-term scaling strategy to build an immense chip manufacturing facility dubbed the "terafab" to completely bypass global semiconductor volume constraints. This manufacturing infrastructure is designed to transition the company into next-generation industrial scaling by producing highly specialized computing components at an unprecedented volume. The scale of this infrastructure project is defined by several extraordinary engineering and production benchmarks: 🏭 The colossal factory is projected to span approximately 100 million square feet, making it ten times larger than the current Tesla Gigafactory Texas. ⚡ The facility is structurally engineered to achieve a massive manufacturing output of 1 terawatt per year once fully operational. 📦 This unprecedented physical footprint provides the capacity required to manufacture 1 billion full-reticle equivalent chips annually. 🔌 Each individual chip manufactured by the facility is designed to run at a power capacity of 1 kilowatt. 🇺🇸 The total scaled output of the facility represents an energy footprint that is exactly double the current annual electricity consumption of the entire United States. "In order to get to the next order of magnitude, you need a gigantic chip factory. To give you a sense of scale here, we expect that the terafab is going to be around 100 million square feet, which is 10 times the size of the Tesla Gigafactory Texas. From a logic die standpoint, that's like having a billion chips per year with a kilowatt per reticle, scaling to a terawatt per year. That is twice the current electricity consumption of the United States." --- 📶 Next-generation high-volume Starlink terminals Dan Huot and Elon Musk introduced their next-generation Starlink user terminals, which have been redesigned specifically to achieve massive manufacturing throughput. Elon Musk pointed out that these newer models will be produced in vastly higher volumes than current hardware designs to fulfill their long-term global deployment targets: 📈 The upgraded user hardware is manufactured at a much higher volume capacity than existing units. 🌍 The company's ultimate target is to successfully deploy a few hundred million of these next-generation terminals worldwide. "In fact, these are the new Starlink terminals, which we made in much higher volume than the current terminals. Ultimately, we think there's probably going to be a few hundred million Starlink terminals out there." --- 📈 Aspirational timeline for orbital AI compute scaling Elon Musk laid out an ambitious, multi-year execution timeline detailing how the company plans to progressively scale space-based processing power. The roadmap targets an initial run-rate by the end of next year and sets an aggressive pace to increase total operational capacity sequentially through a structured, multi-phase timeline: 1️⃣ The initial target aims to hit an annualized run-rate of 1 gigawatt of space AI compute by the end of next year. 2️⃣ The capacity scales to an annualized rate of 10 gigawatts within the next two and a half years. 3️⃣ The operational envelope expands to reach 100 gigawatts in three and a half years. 4️⃣ The long-term deployment plan scales directly to a full terawatt capacity per year using the output of the terafab. "The goal is to get to roughly an annualized rate of a gigawatt per year by the end of next year in terms of space AI compute. Then aspirationally, we want to scale that by an order of magnitude per year. In two and a half years, hitting an annualized rate of 10 gigawatts a year in space, and in three and a half years, maybe a hundred gigawatts, going beyond that with the terafab to scale to a terawatt per year." --- 🌕 Ultimate scaling via lunar production and mass drivers Elon Musk explained that scaling three orders of magnitude past a single terawatt forces a transition completely off-planet to avoid the logistical penalty of Earth's deep gravity well. The vision relies on establishing manufacturing infrastructure directly on the moon to leverage localized resource loops and zero-atmosphere physics: 🌙 The company plans to establish localized raw production lines on the moon to fabricate solar panels, photovoltaics, and radiators from lunar materials. ⚡ Manufacturing components locally avoids the massive fuel and mass penalties of transporting heavy structural materials from Earth. 🧲 Because the moon has no atmosphere and only one-sixth of Earth's gravity, the facility will utilize an electromagnetic mass driver to launch completed satellites. 🚀 Operating essentially as a linear electric motor rail gun, this mechanism will shoot fully assembled AI satellites straight into deep space without relying on chemical rockets. "The only way that we can really see that you can achieve that is on the moon with a mass driver, essentially where you do local production of photovoltaics, solar panels, and radiators on the moon. Because the moon has no atmosphere and only one-sixth Earth's gravity, you can accelerate the AI satellites into deep space without a rocket. You can basically shoot them into space using an electromagnetic gun, like a rail gun type—it's basically a linear electric motor."

Ming

22,203 次观看 • 3 个月前