Loading video...

Video Failed to Load

There was a problem loading this video. This could be due to a temporary network issue or the video might be unavailable.

#generativeAI models have grown 10x in parameters per year for the last five years 🤯 That's why we're announcing Cloud TPU v5p (in preview)—the most powerful, scalable, and flexible TPU to date. P.S. Gemini was trained at scale on Cloud TPUs, too! #GeminiAI

Google Cloud Tech

1,308,137 subscribers

142,604 views • 2 years ago •via X (Twitter)

Education News & Politics Science & Technology #generativeAI #GeminiAI

Anya Rossi• Live Now

Private livecam show

4 Comments

Google Cloud Tech2 years ago

And because a chip is only as powerful as the full stack that surrounds it, we're also introducing AI Hypercomputer! Read about this accelerated supercomputer architecture featuring performance-optimized compute, storage, and networking →

leshalem2 years ago

@googlecloud Google builds good products, but they end up destroying it with too much protocols. How can login in to Gemini take several protocols asking lots of things from users. Very boring user journey. Google always want to dictate users by imposing agreements and policies.

Thao Nguyen🇻🇳2 years ago

well, TPU v5p (in preview)—the most powerful, scalable, and flexible TPU to date.

_lex2 years ago

@Google Did you guys fake the demo? Rumor has it, you spoon fed the questions and pre planned the whole skit.. Is this true. When will you respond to faking the demo?

Related Videos

✨ #GOOGL Key Milestone ✨ shared on Power Lunch Market Navigator Back in July, Alphabet hit a $191 high, but bearish signals kept it in check. Now, those signals (like key moving averages) have flipped to support, paving the way for a potential bullish breakout! 🚀 A weekly close above $191 could push the stock toward $220, based on Fibonacci levels. 📈 Keep an eye on this trend! 👀 ✨A key theme for 2025 is AI Agents✨ Smarter AI Agents Alphabet’s Trillium TPU is a cloud differentiator. TPUs are used to train large language models, and fast and more energy-efficient training capabilities position Alphabet to compete with AWS, and Azure. The trillium TPU increases performance by 4x and is 67% more energy efficient. Salesforce states that increase training speed – which means this will be key for AI agents. ✅It requires increased training speed, preferably an energy-efficient one. ✅Key cloud differentiator with Alphabet’s Trillium TPU ✅Performance is increased by 4 times in comparison to its predecessor and is 67% more energy efficient ✅On Dec 11 Google announced Gemini 2.0 which was trained on these TPUs ✅Combine this training with cloud customers and Google is now positioned to compete with AWS and Azure.

✨ #GOOGL Key Milestone ✨ shared on Power Lunch Market Navigator Back in July, Alphabet hit a $191 high, but bearish signals kept it in check. Now, those signals (like key moving averages) have flipped to support, paving the way for a potential bullish breakout! 🚀 A weekly close above $191 could push the stock toward $220, based on Fibonacci levels. 📈 Keep an eye on this trend! 👀 ✨A key theme for 2025 is AI Agents✨ Smarter AI Agents Alphabet’s Trillium TPU is a cloud differentiator. TPUs are used to train large language models, and fast and more energy-efficient training capabilities position Alphabet to compete with AWS, and Azure. The trillium TPU increases performance by 4x and is 67% more energy efficient. Salesforce states that increase training speed – which means this will be key for AI agents. ✅It requires increased training speed, preferably an energy-efficient one. ✅Key cloud differentiator with Alphabet’s Trillium TPU ✅Performance is increased by 4 times in comparison to its predecessor and is 67% more energy efficient ✅On Dec 11 Google announced Gemini 2.0 which was trained on these TPUs ✅Combine this training with cloud customers and Google is now positioned to compete with AWS and Azure.

Jessica Inskip

28,234 views • 1 year ago

A masterclass on Google's TPU v8 Networking. Two TPU chips? Pssh. We already knew workload-specific silicon was here. But two scale-up networking topologies? That's the actual Google TPU news. Workload-specific interconnects. Think about that. New Semi Doped with Vikram Sekar and Austin Lyons. Copper? Yep. Optics? Yep. What we cover: - TPU splits in two: 8t training, 8i inference. - Virgo: 47 Pb/s scale-out fabric, 100% OCS. - Boardfly scale-up: copper PCB + AECs inside racks, OCS between groups. 16 hops → 7. - Training uses 3D torus (Rubik's Cube). - Inference doesn't. Workload-specific topologies now. - Dragonfly traces to a 2008 paper by Kim, Dally, Scott, Abts. Abts went on to build Groq's interconnect before Nvidia. Chapters: 0:00 Intro 0:21 Two TPUs for two workloads 2:31 HBM, SRAM, and Axion CPUs 7:22 Why networking is the new bottleneck 17:14 Virgo: rebuilding scale-out on optics 25:24 3D torus Rubik's Cube scale-up for training 34:50 Boardfly: scale-up for MoE inference 42:07 Workload-specific everything $GOOGL

A masterclass on Google's TPU v8 Networking. Two TPU chips? Pssh. We already knew workload-specific silicon was here. But two scale-up networking topologies? That's the actual Google TPU news. Workload-specific interconnects. Think about that. New Semi Doped with Vikram Sekar and Austin Lyons. Copper? Yep. Optics? Yep. What we cover: - TPU splits in two: 8t training, 8i inference. - Virgo: 47 Pb/s scale-out fabric, 100% OCS. - Boardfly scale-up: copper PCB + AECs inside racks, OCS between groups. 16 hops → 7. - Training uses 3D torus (Rubik's Cube). - Inference doesn't. Workload-specific topologies now. - Dragonfly traces to a 2008 paper by Kim, Dally, Scott, Abts. Abts went on to build Groq's interconnect before Nvidia. Chapters: 0:00 Intro 0:21 Two TPUs for two workloads 2:31 HBM, SRAM, and Axion CPUs 7:22 Why networking is the new bottleneck 17:14 Virgo: rebuilding scale-out on optics 25:24 3D torus Rubik's Cube scale-up for training 34:50 Boardfly: scale-up for MoE inference 42:07 Workload-specific everything $GOOGL

Semi Doped

92,064 views • 2 months ago

TARS AI was on the ground at Google Cloud Summit Nord in Düsseldorf, surrounded by builders, AI engineers, and cloud leaders. From autonomous agents to GenAI infrastructure, the takeaway was clear, the market is moving from pilots to large-scale deployment. For TARS AI, it was a sharp signal, growing demand for agent-first products and clear alignment on scaling AI systems for real-world adoption. Highlights, insights, and snapshots from the Google Cloud Summit 👇

TARS AI was on the ground at Google Cloud Summit Nord in Düsseldorf, surrounded by builders, AI engineers, and cloud leaders. From autonomous agents to GenAI infrastructure, the takeaway was clear, the market is moving from pilots to large-scale deployment. For TARS AI, it was a sharp signal, growing demand for agent-first products and clear alignment on scaling AI systems for real-world adoption. Highlights, insights, and snapshots from the Google Cloud Summit 👇

TARS AI 🤖

161,078 views • 11 months ago

For over a decade, Intel and Microsoft Azure have partnered to build trust in the cloud. At Microsoft Build 2026, we spoke with Azure CTO Mark Russinovich on how Azure and Intel TDX work together to support confidential cloud workloads at scale with near-zero performance overhead and minimal effort. Watch the full interview:

For over a decade, Intel and Microsoft Azure have partnered to build trust in the cloud. At Microsoft Build 2026, we spoke with Azure CTO Mark Russinovich on how Azure and Intel TDX work together to support confidential cloud workloads at scale with near-zero performance overhead and minimal effort. Watch the full interview:

Intel Business

19,756 views • 15 days ago

🔥 NVIDIA GTC 2026: The Vera Rubin platform just dropped, introducing 10x inference throughput per watt, at 1/10th the cost per token vs. Blackwell. Jensen Huang called it: every company will now run a "token factory." Here's the thing: Aethir has been running distributed token factories since day one, with nearly 440,000+ GPU containers across 94 countries, delivering 1.5B+ compute hours to enterprises, powered by ATH. When the world's most powerful AI chips land, you need a cloud that can actually deploy them at scale, without the centralized cloud limitations and bottlenecks. That's Aethir’s advantage over centralized compute providers. ⚡

🔥 NVIDIA GTC 2026: The Vera Rubin platform just dropped, introducing 10x inference throughput per watt, at 1/10th the cost per token vs. Blackwell. Jensen Huang called it: every company will now run a "token factory." Here's the thing: Aethir has been running distributed token factories since day one, with nearly 440,000+ GPU containers across 94 countries, delivering 1.5B+ compute hours to enterprises, powered by ATH. When the world's most powerful AI chips land, you need a cloud that can actually deploy them at scale, without the centralized cloud limitations and bottlenecks. That's Aethir’s advantage over centralized compute providers. ⚡

Aethir Ecosystem

14,116 views • 3 months ago

Here's the news- I'm moving on from my head of community role at the Data on Kubernetes Community. Too many people to thank and I'm extremely grateful for having been able to watch this community grow the last 2.5 years. I'll stay working in the cloud native ecosystem- let's stay in touch!

Here's the news- I'm moving on from my head of community role at the Data on Kubernetes Community. Too many people to thank and I'm extremely grateful for having been able to watch this community grow the last 2.5 years. I'll stay working in the cloud native ecosystem- let's stay in touch!

Bart Farrell

10,001 views • 3 years ago

🚨 BREAKING: Secretary of Defense Pete Hegseth has just CANCELED the US Department of Defense's reliance on "cheap Chinese labor" for their cloud services, an Obama-era program... ...WHY ON EARTH WAS THE US RELYING ON CHINA FOR OUR MILITARY COMPUTER INFRASTRUCTURE?!?! "It turns out that some tech companies have been using cheap Chinese labor to assist with DoD cloud services. This is obviously unacceptable, especially in today's digital threat environment." "Now, this was a legacy system created over a decade ago during the Obama administration...that's why today I'm announcing that China will no longer have any involvement whatsoever in our cloud services effective immediately." It is INSANE that Pete Hegseth had to do this.

🚨 BREAKING: Secretary of Defense Pete Hegseth has just CANCELED the US Department of Defense's reliance on "cheap Chinese labor" for their cloud services, an Obama-era program... ...WHY ON EARTH WAS THE US RELYING ON CHINA FOR OUR MILITARY COMPUTER INFRASTRUCTURE?!?! "It turns out that some tech companies have been using cheap Chinese labor to assist with DoD cloud services. This is obviously unacceptable, especially in today's digital threat environment." "Now, this was a legacy system created over a decade ago during the Obama administration...that's why today I'm announcing that China will no longer have any involvement whatsoever in our cloud services effective immediately." It is INSANE that Pete Hegseth had to do this.

Q STORM RIDER

575,774 views • 2 months ago

🚨 BREAKING: Secretary of Defense Pete Hegseth has just CANCELED the US Department of Defense's reliance on "cheap Chinese labor" for their cloud services, an Obama-era program... ...WHY ON EARTH WAS THE US RELYING ON CHINA FOR OUR MILITARY COMPUTER INFRASTRUCTURE?!?! "It turns out that some tech companies have been using cheap Chinese labor to assist with DoD cloud services. This is obviously unacceptable, especially in today's digital threat environment." "Now, this was a legacy system created over a decade ago during the Obama administration...that's why today I'm announcing that China will no longer have any involvement whatsoever in our cloud services effective immediately." It is INSANE that Pete Hegseth had to do this.

🚨 BREAKING: Secretary of Defense Pete Hegseth has just CANCELED the US Department of Defense's reliance on "cheap Chinese labor" for their cloud services, an Obama-era program... ...WHY ON EARTH WAS THE US RELYING ON CHINA FOR OUR MILITARY COMPUTER INFRASTRUCTURE?!?! "It turns out that some tech companies have been using cheap Chinese labor to assist with DoD cloud services. This is obviously unacceptable, especially in today's digital threat environment." "Now, this was a legacy system created over a decade ago during the Obama administration...that's why today I'm announcing that China will no longer have any involvement whatsoever in our cloud services effective immediately." It is INSANE that Pete Hegseth had to do this.

WWG1WGA🇺🇸♥️

130,389 views • 2 months ago

🚨 BREAKING: Secretary of Defense Pete Hegseth has just CANCELED the US Department of Defense's reliance on "cheap Chinese labor" for their cloud services, an Obama-era program... ...WHY ON EARTH WAS THE US RELYING ON CHINA FOR OUR MILITARY COMPUTER INFRASTRUCTURE?!?! "It turns out that some tech companies have been using cheap Chinese labor to assist with DoD cloud services. This is obviously unacceptable, especially in today's digital threat environment." "Now, this was a legacy system created over a decade ago during the Obama administration...that's why today I'm announcing that China will no longer have any involvement whatsoever in our cloud services effective immediately." It is INSANE that Pete Hegseth had to do this.

🚨 BREAKING: Secretary of Defense Pete Hegseth has just CANCELED the US Department of Defense's reliance on "cheap Chinese labor" for their cloud services, an Obama-era program... ...WHY ON EARTH WAS THE US RELYING ON CHINA FOR OUR MILITARY COMPUTER INFRASTRUCTURE?!?! "It turns out that some tech companies have been using cheap Chinese labor to assist with DoD cloud services. This is obviously unacceptable, especially in today's digital threat environment." "Now, this was a legacy system created over a decade ago during the Obama administration...that's why today I'm announcing that China will no longer have any involvement whatsoever in our cloud services effective immediately." It is INSANE that Pete Hegseth had to do this.

Eric Daugherty

1,037,215 views • 1 year ago

Project Genesis is the closest thing we’ve seen to the Manhattan Project for science since 1939, what will this mean for America's AI dominance? - AWS planning $50B in U.S. government AI infrastructure + a 2.2 GW Indiana megacenter. - Google’s Ironwood TPU lands 4× performance and million-token multimodal context. - Amazon deploying 500,000 Tranium 2 chips for Anthropic-scale workloads. - Google is now rivaling NVIDIA by letting companies like Meta train frontier models directly on Google Cloud instead of buying GPUs.

Project Genesis is the closest thing we’ve seen to the Manhattan Project for science since 1939, what will this mean for America's AI dominance? - AWS planning $50B in U.S. government AI infrastructure + a 2.2 GW Indiana megacenter. - Google’s Ironwood TPU lands 4× performance and million-token multimodal context. - Amazon deploying 500,000 Tranium 2 chips for Anthropic-scale workloads. - Google is now rivaling NVIDIA by letting companies like Meta train frontier models directly on Google Cloud instead of buying GPUs.

Peter H. Diamandis, MD

41,936 views • 7 months ago

At Georgia Power, we're proud to power the people who power our state. That's why we froze base rates last year—so you have more predictability, stability, and money in your pocket. Georgia Power. We're here for Georgia. We're here for you. 🤝 Learn more:

At Georgia Power, we're proud to power the people who power our state. That's why we froze base rates last year—so you have more predictability, stability, and money in your pocket. Georgia Power. We're here for Georgia. We're here for you. 🤝 Learn more:

Georgia Power

8,792,665 views • 4 months ago

Former Mayor of Nevada City and Former California Governor candidate Reinette Senum says there is a 200% increase in cloud seeding in America coming over the next few years “Most of the cloud seeding programs are happening from the Colorado Rockies all the way west. When a storm is coming in, these pilots, they get hired to make $750 an hour to do this. — So they release the silver iodide as well as a chaff. And it's like a military chaff. It's a aluminum coated glass — and that is for the irrigation districts, that's for the power companies — So that is what's happening for a long time. They denied it saying we don't cloud seed, and then one day like, oh yes, we do cloud seed, but we've been doing it for decades and it's completely proven safe. No, actually has not been proven safe. In fact, we know it's not safe. And by the way, those cloud seeding projects and programs last year, they announced they want like a 200% increase over the next few years of cloud seeding.”

Former Mayor of Nevada City and Former California Governor candidate Reinette Senum says there is a 200% increase in cloud seeding in America coming over the next few years “Most of the cloud seeding programs are happening from the Colorado Rockies all the way west. When a storm is coming in, these pilots, they get hired to make $750 an hour to do this. — So they release the silver iodide as well as a chaff. And it's like a military chaff. It's a aluminum coated glass — and that is for the irrigation districts, that's for the power companies — So that is what's happening for a long time. They denied it saying we don't cloud seed, and then one day like, oh yes, we do cloud seed, but we've been doing it for decades and it's completely proven safe. No, actually has not been proven safe. In fact, we know it's not safe. And by the way, those cloud seeding projects and programs last year, they announced they want like a 200% increase over the next few years of cloud seeding.”

Wall Street Apes

97,600 views • 7 months ago

It dawned on me last night after I took the oath of office to usher in major reform at the USAGM that it was four years to the date that I put out my video announcing that I was walking away from a 30-year career in journalism. In the past four years, more people have woken up to the lies of the Fake News. And a new wave of independent journalists have come on the scene to spread the truth. I see bright days ahead.

It dawned on me last night after I took the oath of office to usher in major reform at the USAGM that it was four years to the date that I put out my video announcing that I was walking away from a 30-year career in journalism. In the past four years, more people have woken up to the lies of the Fake News. And a new wave of independent journalists have come on the scene to spread the truth. I see bright days ahead.

Kari Lake

104,760 views • 1 year ago

.Jon Callaghan (Founder, True Ventures) on the scale of the AI boom: “Our numbers are 2 to 4 trillion of CapEx going into the AI wave over the next five years. In every prior wave (Web2, mobile, cloud) the value created at the application layer on top of that CapEx was 5 to 10x.” “We’re talking about markets that haven’t been invented yet. Net new consumer behaviors. Net new enterprise behaviors. At an unprecedented level.” For founders and investors, he calls it “the vintage of a lifetime.”

.Jon Callaghan (Founder, True Ventures) on the scale of the AI boom: “Our numbers are 2 to 4 trillion of CapEx going into the AI wave over the next five years. In every prior wave (Web2, mobile, cloud) the value created at the application layer on top of that CapEx was 5 to 10x.” “We’re talking about markets that haven’t been invented yet. Net new consumer behaviors. Net new enterprise behaviors. At an unprecedented level.” For founders and investors, he calls it “the vintage of a lifetime.”

TBPN

23,562 views • 10 months ago

“You have this violent underclass in D.C who have terrorized the city for the last five years…” Alex Marlow and 🇺🇸 Mike Davis 🇺🇸 on President Trump announcing major plans to clean up Washington D.C. and fix cities. Watch below, and subscribe:

“You have this violent underclass in D.C who have terrorized the city for the last five years…” Alex Marlow and 🇺🇸 Mike Davis 🇺🇸 on President Trump announcing major plans to clean up Washington D.C. and fix cities. Watch below, and subscribe:

The Megyn Kelly Show

22,385 views • 11 months ago

I asked Jensen: “2 out of the top 3 models in the world, Claude and Gemini, were trained on TPU. What does that mean for Nvidia going forward?” After a long technical back and forth about what the right accelerator for AI looks like (see full episode), Jensen lays down the gauntlet: Jensen: "Nvidia's computing stack is the best performance per TCO in the world, bar none. Dylan's InferenceMax is sitting out there for everybody to use, and TPU won't come, Trainium won't come. Nobody wants to show up." “I'm not offended by other people using something else and trying things. If they don't try these other things, how would they know how good ours is?” “There are always big claims, but look at the number of ASICs that have been canceled. Just because you're going to build an ASIC, you still have to build something better than Nvidia. It's not that easy building something better than Nvidia. Nvidia's got to be missing something seriously, because of our scale and our velocity.”

I asked Jensen: “2 out of the top 3 models in the world, Claude and Gemini, were trained on TPU. What does that mean for Nvidia going forward?” After a long technical back and forth about what the right accelerator for AI looks like (see full episode), Jensen lays down the gauntlet: Jensen: "Nvidia's computing stack is the best performance per TCO in the world, bar none. Dylan's InferenceMax is sitting out there for everybody to use, and TPU won't come, Trainium won't come. Nobody wants to show up." “I'm not offended by other people using something else and trying things. If they don't try these other things, how would they know how good ours is?” “There are always big claims, but look at the number of ASICs that have been canceled. Just because you're going to build an ASIC, you still have to build something better than Nvidia. It's not that easy building something better than Nvidia. Nvidia's got to be missing something seriously, because of our scale and our velocity.”

Dwarkesh Patel

419,021 views • 3 months ago

What I loved about the launch of Gemini 3 last week was seeing so many products (from Google and others in the ecosystem) improve their experience with it on Day 1, truly innovation at scale. Gemini is a manifestation of our decade long AI first strategy, I see it as a through line for everything - from Search to YouTube to Cloud to Waymo etc I talked about this and much more with Logan Kilpatrick on our Release Notes podcast last week, full episode now on YouTube:

What I loved about the launch of Gemini 3 last week was seeing so many products (from Google and others in the ecosystem) improve their experience with it on Day 1, truly innovation at scale. Gemini is a manifestation of our decade long AI first strategy, I see it as a through line for everything - from Search to YouTube to Cloud to Waymo etc I talked about this and much more with Logan Kilpatrick on our Release Notes podcast last week, full episode now on YouTube:

Sundar Pichai

222,213 views • 7 months ago

Jenny Johnson, CEO of Franklin Templeton on Trump Accounts: "If you invest $1,000 a year for 10 years at age 20, when you retire... you will have more money than a person who starts to invest at age 30 and invests the same amount per year for 30 years—that's the power of compounding—and so if you look at Trump Accounts, we're now starting at birth and being able to compound. I think it's incredibly powerful."

Jenny Johnson, CEO of Franklin Templeton on Trump Accounts: "If you invest $1,000 a year for 10 years at age 20, when you retire... you will have more money than a person who starts to invest at age 30 and invests the same amount per year for 30 years—that's the power of compounding—and so if you look at Trump Accounts, we're now starting at birth and being able to compound. I think it's incredibly powerful."

Rapid Response 47

591,023 views • 9 days ago

The power of generative models — now embodied in humanoids. Announcing DreamControl –– After a year-long research effort at General Robotics — we present a scalable framework for whole-body humanoid control that fuses diffusion priors with reinforcement learning to unlock real-world scene interaction. Diffusion + RL → natural whole-body skills on real robots. DreamControl enables humanoids to move beyond locomotion demos → performing natural, human-like skills such as –– Picking & lifting objects, Opening drawers & doors, Precise punching, kicking, and jumping, Bimanual manipulation tasks Our key innovation: a diffusion prior over human motion that guides RL, eliminating the need for massive teleoperation datasets, and producing motions that look human while transferring to real hardware. Trained purely in simulation, deployed on the Unitree G1 humanoid, DreamControl policies run in real time, bridging sim-to-real with unprecedented naturalness. We leverage a novel hybrid edge + cloud infrastructure that runs RL-trained policies on the edge backed by powerful AI models running in the cloud This is the next step in General Robotics’ journey toward general-purpose humanoid assistants that interact, adapt, and assist autonomously. Paper: Blog: 1/n

The power of generative models — now embodied in humanoids. Announcing DreamControl –– After a year-long research effort at General Robotics — we present a scalable framework for whole-body humanoid control that fuses diffusion priors with reinforcement learning to unlock real-world scene interaction. Diffusion + RL → natural whole-body skills on real robots. DreamControl enables humanoids to move beyond locomotion demos → performing natural, human-like skills such as –– Picking & lifting objects, Opening drawers & doors, Precise punching, kicking, and jumping, Bimanual manipulation tasks Our key innovation: a diffusion prior over human motion that guides RL, eliminating the need for massive teleoperation datasets, and producing motions that look human while transferring to real hardware. Trained purely in simulation, deployed on the Unitree G1 humanoid, DreamControl policies run in real time, bridging sim-to-real with unprecedented naturalness. We leverage a novel hybrid edge + cloud infrastructure that runs RL-trained policies on the edge backed by powerful AI models running in the cloud This is the next step in General Robotics’ journey toward general-purpose humanoid assistants that interact, adapt, and assist autonomously. Paper: Blog: 1/n

Ashish Kapoor

118,133 views • 10 months ago