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.

Robots are getting smarter, but how can their hardware match that growth? New Microsoft Research findings show that moving AI inference beyond the robot can improve task success, boost efficiency, and support more advanced physical AI workloads.

Microsoft Research

555,070 subscribers

19,605 views • 3 days ago •via X (Twitter)

Anya Rossi• Live Now

Private livecam show

0 Comments

No comments available

Comments from the original post will appear here

Related Videos

Prediction: By 2030, global data center workloads are projected to split into: • 50% traditional workloads • 37% AI inference • 13% AI training Inference is set to become nearly 3× larger than training. As more AI applications move into production, builders will need infrastructure that can support lower-cost inference, reliable execution, faster routing and private workloads at scale. This is where FAR AI enters the market. Source: JLL Research, 2025

Prediction: By 2030, global data center workloads are projected to split into: • 50% traditional workloads • 37% AI inference • 13% AI training Inference is set to become nearly 3× larger than training. As more AI applications move into production, builders will need infrastructure that can support lower-cost inference, reliable execution, faster routing and private workloads at scale. This is where FAR AI enters the market. Source: JLL Research, 2025

FAR Labs

20,706 views • 3 months ago

Robots don’t just need better brains. They need WAY more real-world data. 🤖 And collecting high-quality dexterous robot data at scale is one of the hardest problems in physical AI. A fascinating approach is emerging: Wearable human demonstrations + structurally matched dexterous robots. Instead of humans directly teleoperating a robot, Chinese embodied AI startup X Square Robot's TwinDEX system captures the motion, contact, and visual information needed to train the robot — while keeping the data closer to the hardware that will actually execute the task. Early results show promising performance on tool use, fine manipulation, and complex contact-rich tasks. If this approach scales, it could change how we build real-world robotics datasets. The next frontier of physical AI might not be bigger models. It might be better data.

Robots don’t just need better brains. They need WAY more real-world data. 🤖 And collecting high-quality dexterous robot data at scale is one of the hardest problems in physical AI. A fascinating approach is emerging: Wearable human demonstrations + structurally matched dexterous robots. Instead of humans directly teleoperating a robot, Chinese embodied AI startup X Square Robot's TwinDEX system captures the motion, contact, and visual information needed to train the robot — while keeping the data closer to the hardware that will actually execute the task. Early results show promising performance on tool use, fine manipulation, and complex contact-rich tasks. If this approach scales, it could change how we build real-world robotics datasets. The next frontier of physical AI might not be bigger models. It might be better data.

The Daily Ai

33,311 views • 24 days ago

Building autonomous vehicles, robots, and vision AI takes more data than any team can collect on their own. At #CVPR2026, NVIDIA announced physical AI agent skills to help speed development: composable workflows that automate data generation, simulation, policy training and evaluation, powered by Cosmos 3. Read the blog:

Building autonomous vehicles, robots, and vision AI takes more data than any team can collect on their own. At #CVPR2026, NVIDIA announced physical AI agent skills to help speed development: composable workflows that automate data generation, simulation, policy training and evaluation, powered by Cosmos 3. Read the blog:

NVIDIA

32,297 views • 3 months ago

Introducing Matrix I crawled 100,000+ agents, skills and tools to train a new model which can answer what capabilities are the best match for a task. Think Google, but for agents. A living model that learns from the gossiping network, and gets smarter with every interaction.

Introducing Matrix I crawled 100,000+ agents, skills and tools to train a new model which can answer what capabilities are the best match for a task. Think Google, but for agents. A living model that learns from the gossiping network, and gets smarter with every interaction.

Varun

899,240 views • 6 months ago

Part of the reason IT people are being laid off is that many of their abilities can be replaced by AI. In the past, the joke was that you would just google it, but that still required some IT knowledge. Now, you can ask AI how to fix an issue and it will walk you through how to fix most things. (if you have admin permissions to your computer) To be fair, no one knows all the weird nuances and issues that can arise with software so google has always been used to find IT fixes - but it took more research and knowing how to implement the fix before.

Part of the reason IT people are being laid off is that many of their abilities can be replaced by AI. In the past, the joke was that you would just google it, but that still required some IT knowledge. Now, you can ask AI how to fix an issue and it will walk you through how to fix most things. (if you have admin permissions to your computer) To be fair, no one knows all the weird nuances and issues that can arise with software so google has always been used to find IT fixes - but it took more research and knowing how to implement the fix before.

AlphaFox

23,415 views • 1 month ago

It's the little things that make me excited about the future. Many people are (rightfully) worried that AI and robotics will take their jobs. Yes, it's about money they need to live on, and solutions are needed. At the same time, it's often work that's strenuous and tedious. So we can be glad when robots take these tasks off our hands. Furthermore, robots don't need breaks, vacations, or sleep. They can be scaled quickly and are efficient. Globalization, and humanity as a whole, can therefore benefit massively from this increase in efficiency.

It's the little things that make me excited about the future. Many people are (rightfully) worried that AI and robotics will take their jobs. Yes, it's about money they need to live on, and solutions are needed. At the same time, it's often work that's strenuous and tedious. So we can be glad when robots take these tasks off our hands. Furthermore, robots don't need breaks, vacations, or sleep. They can be scaled quickly and are efficient. Globalization, and humanity as a whole, can therefore benefit massively from this increase in efficiency.

Chubby♨️

25,528 views • 6 months ago

MCP! MCP! MCP! Everyone's chanting it. But what is it? MCP is how you give your AI new skills. Install UpRock MCP and your favorite LLM can: → Search the live web → Understand videos → Run deep research → Retrieve real-time intelligence Works with ChatGPT, Claude and Gemini. One install. Smarter AI.

MCP! MCP! MCP! Everyone's chanting it. But what is it? MCP is how you give your AI new skills. Install UpRock MCP and your favorite LLM can: → Search the live web → Understand videos → Run deep research → Retrieve real-time intelligence Works with ChatGPT, Claude and Gemini. One install. Smarter AI.

UpRock

16,231 views • 2 months ago

🎉 JUST IN: GPT-5.6 and Kimi K3 are now live on GlobalGPT! Try two new powerful models for smarter reasoning, faster writing, better coding, and smoother AI workflows. From complex research to content creation, coding, summarizing, and brainstorming — switch between models and find the one that fits your task best. 👇 #GPT56 #KimiK3 #AI #ChatGPT #GlobalGPT

🎉 JUST IN: GPT-5.6 and Kimi K3 are now live on GlobalGPT! Try two new powerful models for smarter reasoning, faster writing, better coding, and smoother AI workflows. From complex research to content creation, coding, summarizing, and brainstorming — switch between models and find the one that fits your task best. 👇 #GPT56 #KimiK3 #AI #ChatGPT #GlobalGPT

marium

13,036 views • 2 months ago

A Gaussian Splat can become a world where Robots and AI agents can act. In our latest OVER Research experiment, we placed a robot inside a real-world 3D capture, with a VLM making decisions based on what it sees. At every step, the robot holds a pose in the reconstruction, gets a newly rendered view of the environment, takes an action, moves, and sees the world again from its new position. Why does this matter? Because 3D captures can become more than reconstructions to explore. They can become environments where embodied AI and robots can navigate, act, be evaluated and eventually train across real-world spaces at scale. Capture a place once. Then turn it into a world where AI and Robots can act. The full experiment, including what we discovered once we actually put the loop to the test:

A Gaussian Splat can become a world where Robots and AI agents can act. In our latest OVER Research experiment, we placed a robot inside a real-world 3D capture, with a VLM making decisions based on what it sees. At every step, the robot holds a pose in the reconstruction, gets a newly rendered view of the environment, takes an action, moves, and sees the world again from its new position. Why does this matter? Because 3D captures can become more than reconstructions to explore. They can become environments where embodied AI and robots can navigate, act, be evaluated and eventually train across real-world spaces at scale. Capture a place once. Then turn it into a world where AI and Robots can act. The full experiment, including what we discovered once we actually put the loop to the test:

Over the Reality 🌐

14,975 views • 25 days ago

AI has leapfrogged years of research on 3D rendering, and can make photorealistic images and videos that are far beyond the technology of CGI. Is 3D graphics a dead field? Video made with my ray tracer:

AI has leapfrogged years of research on 3D rendering, and can make photorealistic images and videos that are far beyond the technology of CGI. Is 3D graphics a dead field? Video made with my ray tracer:

Donald Mitchell

54,612 views • 1 month ago

MAKING AXIS TASKS FEEL EASY 👀🤖 I made a quick video showing how to complete an Axis Robotics task step by step. The task gives you a clear goal, steps, controls and camera views to guide you. In this one, the goal is simple, Open the trash bin, pick up the onion, and place it inside. 🧅 Just follow the instructions, control the robot arm and complete the task. What I like is that even beginners can understand what they need to do. But the interesting part is what happens with these interactions after we complete them. Each completed task can contribute valuable robot training data, helping Physical Ai systems learn how to handle different objects and environments. Small tasks from the community can become useful data for teaching robots how to interact with the real world. That’s what makes Axis interesting to me.

MAKING AXIS TASKS FEEL EASY 👀🤖 I made a quick video showing how to complete an Axis Robotics task step by step. The task gives you a clear goal, steps, controls and camera views to guide you. In this one, the goal is simple, Open the trash bin, pick up the onion, and place it inside. 🧅 Just follow the instructions, control the robot arm and complete the task. What I like is that even beginners can understand what they need to do. But the interesting part is what happens with these interactions after we complete them. Each completed task can contribute valuable robot training data, helping Physical Ai systems learn how to handle different objects and environments. Small tasks from the community can become useful data for teaching robots how to interact with the real world. That’s what makes Axis interesting to me.

SufianXFN

14,419 views • 20 days ago

New: Projects are now available on mobile. We’ve updated the experience so you can organize project-based work wherever you are, from simple task management to more advanced workflows that rely on shared files, instructions, skills, and connectors.

New: Projects are now available on mobile. We’ve updated the experience so you can organize project-based work wherever you are, from simple task management to more advanced workflows that rely on shared files, instructions, skills, and connectors.

Manus

56,167 views • 4 months ago

Each weaver ant can pull on a leaf much harder when in a group than when it works alone. The ants’ sticky feet and coordinated legwork, researchers believe, improve the ants’ collective efficiency—a finding that may provide insights into how to build and coordinate swarms of robots that work together. Learn more: News from Science

Each weaver ant can pull on a leaf much harder when in a group than when it works alone. The ants’ sticky feet and coordinated legwork, researchers believe, improve the ants’ collective efficiency—a finding that may provide insights into how to build and coordinate swarms of robots that work together. Learn more: News from Science

Science Magazine

50,685 views • 1 year ago

Hold 10,000 $CEREBRO and start earning AI inference credits. Today, you can use those credits through the API or trade them on the Cerebro Marketplace. But Phase 3 gives them another purpose. AI agents need compute to run. Every response, model call and task consumes inference. As more agents and applications are built through Cerebro, those credits become part of what actually keeps them running. Something you earn today becomes something the ecosystem can need tomorrow. That’s what Phase 3 starts to unlock.

Hold 10,000 $CEREBRO and start earning AI inference credits. Today, you can use those credits through the API or trade them on the Cerebro Marketplace. But Phase 3 gives them another purpose. AI agents need compute to run. Every response, model call and task consumes inference. As more agents and applications are built through Cerebro, those credits become part of what actually keeps them running. Something you earn today becomes something the ecosystem can need tomorrow. That’s what Phase 3 starts to unlock.

Cerebro

14,636 views • 13 days ago

Building robots that can effectively operate alongside human workers is difficult. 🛠️ Advances in open-source physics, open foundation models, and frameworks are helping accelerate physical #AI deployment. ✔️ Newton Physics Engine, an open-source GPU-powered simulation built on OpenUSD, speeds up robot learning for advanced manipulation and mobility. ✔️ NVIDIA Cosmos Reason, an open reasoning vision language model, gives robots the ability to think like humans using prior knowledge, common sense and physics ✔️NVIDIA Isaac GR00T N1.6, an open robot foundation model, enables humanoids to understand ambiguous instructions Leading robotics developers including Agility Robotics, Lightwheel, Mentee Robotics, UniversalRobots, and Wandelbots are adopting simulation technologies and libraries to accelerate physical AI development and deployment. Omniverse Ambassador Dylan Tobin built an AI chatbot trained on Isaac Sim workflows, helping devs navigate Omniverse faster. Read the full blog 👉

Building robots that can effectively operate alongside human workers is difficult. 🛠️ Advances in open-source physics, open foundation models, and frameworks are helping accelerate physical #AI deployment. ✔️ Newton Physics Engine, an open-source GPU-powered simulation built on OpenUSD, speeds up robot learning for advanced manipulation and mobility. ✔️ NVIDIA Cosmos Reason, an open reasoning vision language model, gives robots the ability to think like humans using prior knowledge, common sense and physics ✔️NVIDIA Isaac GR00T N1.6, an open robot foundation model, enables humanoids to understand ambiguous instructions Leading robotics developers including Agility Robotics, Lightwheel, Mentee Robotics, UniversalRobots, and Wandelbots are adopting simulation technologies and libraries to accelerate physical AI development and deployment. Omniverse Ambassador Dylan Tobin built an AI chatbot trained on Isaac Sim workflows, helping devs navigate Omniverse faster. Read the full blog 👉

NVIDIA

48,500 views • 11 months ago

💡AI Breakthrough in Epilepsy Detection: Spotting the ‘Invisible’ 🧠 🔍 A new AI tool, MELD Graph, can identify hidden brain abnormalities in epilepsy patients—ones that even radiologists often miss. 📊 Developed by King’s College London and trained on MRI data from over 700 epilepsy patients, the AI can highlight abnormalities, explain its findings, and improve diagnosis speed, potentially leading to life-changing surgeries. ⚡ Why it matters? Up to 50% of epilepsy-related brain lesions go undetected. This AI finds two-thirds of those missed cases, offering hope for more effective treatments.

💡AI Breakthrough in Epilepsy Detection: Spotting the ‘Invisible’ 🧠 🔍 A new AI tool, MELD Graph, can identify hidden brain abnormalities in epilepsy patients—ones that even radiologists often miss. 📊 Developed by King’s College London and trained on MRI data from over 700 epilepsy patients, the AI can highlight abnormalities, explain its findings, and improve diagnosis speed, potentially leading to life-changing surgeries. ⚡ Why it matters? Up to 50% of epilepsy-related brain lesions go undetected. This AI finds two-thirds of those missed cases, offering hope for more effective treatments.

NEXTA

63,350 views • 1 year ago

NVIDIA might have just declared war on the cloud GPU business For years, AI builders had one option Rent compute Pay every month Watch the bill grow every time usage increased Now NVIDIA is putting serious AI hardware directly on people's desks Small enough to fit next to a monitor Powerful enough to run workloads that used to require expensive cloud infrastructure That's why this launch is getting so much attention The real story isn't the hardware specs It's the business model shift Every month, developers send money to cloud providers for inference, testing, fine-tuning and AI applications The question nobody can answer yet is what happens if enough developers decide they'd rather buy infrastructure once than rent it forever Because if local AI hardware keeps getting more powerful, the economics start changing very quickly Cloud providers built empires on renting access to compute NVIDIA is betting more people will eventually want to own it And that's a much bigger story than a new piece of hardware sitting on a desk

NVIDIA might have just declared war on the cloud GPU business For years, AI builders had one option Rent compute Pay every month Watch the bill grow every time usage increased Now NVIDIA is putting serious AI hardware directly on people's desks Small enough to fit next to a monitor Powerful enough to run workloads that used to require expensive cloud infrastructure That's why this launch is getting so much attention The real story isn't the hardware specs It's the business model shift Every month, developers send money to cloud providers for inference, testing, fine-tuning and AI applications The question nobody can answer yet is what happens if enough developers decide they'd rather buy infrastructure once than rent it forever Because if local AI hardware keeps getting more powerful, the economics start changing very quickly Cloud providers built empires on renting access to compute NVIDIA is betting more people will eventually want to own it And that's a much bigger story than a new piece of hardware sitting on a desk

beamnxw ./

30,361 views • 3 months ago

Today, we're excited to announce a partnership with Manta Network (🔱,🔱), the Modular L2 solution that is transforming the landscape of ZK. Nesa is bringing private AI inference to the Manta ecosystem through a specialized collaboration. For the first time, developers on Manta can access Nesa's full library of AI models on-chain, and enjoy lightning fast, end-to-end private AI inference without ever leaving the Manta ecosystem. This means that dapps and protocols can now fuse with AI via smart contract on Manta. This integration is set to redefine the future of decentralized technology with AI. Stay tuned for more updates on how Nesa and Manta Network will be shaping the future of crypto together.

Today, we're excited to announce a partnership with Manta Network (🔱,🔱), the Modular L2 solution that is transforming the landscape of ZK. Nesa is bringing private AI inference to the Manta ecosystem through a specialized collaboration. For the first time, developers on Manta can access Nesa's full library of AI models on-chain, and enjoy lightning fast, end-to-end private AI inference without ever leaving the Manta ecosystem. This means that dapps and protocols can now fuse with AI via smart contract on Manta. This integration is set to redefine the future of decentralized technology with AI. Stay tuned for more updates on how Nesa and Manta Network will be shaping the future of crypto together.

Nesa

35,224 views • 2 years ago

Research gets messy very quickly when one AI is trying to do everything alone. Grok Build’s new /deep-research command is built for the bigger jobs. It sends multiple agents into the work at once, but with structure. Plan. Research. Verify. Report. That means the system can split the task, chase evidence, cross-check findings, and come back with cited output instead of one lonely model guessing in a corner. The geeky kind of command people will start overusing immediately. Grok xAI / Writer: Annette, Grok Imagine Designer: Janné

Research gets messy very quickly when one AI is trying to do everything alone. Grok Build’s new /deep-research command is built for the bigger jobs. It sends multiple agents into the work at once, but with structure. Plan. Research. Verify. Report. That means the system can split the task, chase evidence, cross-check findings, and come back with cited output instead of one lonely model guessing in a corner. The geeky kind of command people will start overusing immediately. Grok xAI / Writer: Annette, Grok Imagine Designer: Janné

Mario Nawfal

36,005 views • 1 month ago