Jensen Huang says robots will have human capabilities this year! The Robots at CES Had.. Other Plans

Summary

This video from Turing Post contrasts Jensen Huang’s bullish predictions about robots achieving human-level capabilities with the comical reality of robots at CES 2026 still struggling with basic tasks. The host uses humorous clips of robots failing at CES to set up a more substantive discussion about the actual progress being made in robotics across various industries, from home assistants to agricultural machines.

The video provides a comprehensive overview of robotics progress driven by generative AI, covering multiple sectors: personal assistant robots like Richie Mini with on-device AI and open-source capabilities, warehouse logistics with adaptive autonomous fleets at Amazon and DHL, factory automation with vision-guided robotic arms, hospital service robots for medication delivery and surgical assistance, agricultural robots for precision farming, and inspection robots for dangerous environments like mines and nuclear facilities.

The episode concludes with clips from both Elon Musk discussing Tesla’s Optimus robots (predicting 10 billion+ robots by 2040) and Jensen Huang providing a philosophical take on robotics development, explaining the progression from locomotion to gross articulation to fine motor skills, and the concept of AI models that can “embody” different forms like humanoids, manipulators, or self-driving cars.

Highlights

”When Will Robots Have Human Capabilities? This Year.”

yt-dlp --download-sections "*0:15-1:15" "https://www.youtube.com/watch?v=PAjb83HHLaU" --force-keyframes-at-cuts --merge-output-format mp4 -o "PAjb83HHLaU-0m15s.mp4"

“When do you think we’re going to get robots that actually have human level capabilities? Oh, this year.” — Jensen Huang, 0:15

”10 Billion Robots by 2040 Is a Low Number”

yt-dlp --download-sections "*4:27-5:20" "https://www.youtube.com/watch?v=PAjb83HHLaU" --force-keyframes-at-cuts --merge-output-format mp4 -o "PAjb83HHLaU-4m27s.mp4"

“10 billion by 2040? That’s not… that’s a low number.” — Elon Musk, 4:38

”You Embody the Car Like an Extension of You”

yt-dlp --download-sections "*5:49-6:40" "https://www.youtube.com/watch?v=PAjb83HHLaU" --force-keyframes-at-cuts --merge-output-format mp4 -o "PAjb83HHLaU-5m49s.mp4"

“When you sit inside a car, somehow you embody the car. You’re able to steer a car like it’s an extension of you. When you picked up a tennis racket, you embodied the tennis racket. Somehow your arm got longer.” — Jensen Huang, 5:57

”Robots That Are Almost Boringly Helpful”

yt-dlp --download-sections "*1:37-2:30" "https://www.youtube.com/watch?v=PAjb83HHLaU" --force-keyframes-at-cuts --merge-output-format mp4 -o "PAjb83HHLaU-1m37s.mp4"

“Don’t forget we already have robots that are almost boringly helpful.” — Turing Post Host, 2:00

Key Points

• Jensen Huang’s Prediction (0:00) - When asked about human-level robot capabilities, Jensen Huang surprised everyone with his optimistic timeline
• CES Robot Fails (0:26) - Humorous compilation showing robots at CES still looking “funny, awkward, and clumsy” despite bold predictions
• Generative AI Driving Progress (1:37) - Last year robots had phenomenal progress driven by generative AI helping with simulation and training
• Richie Mini Robot (2:08) - True star of CES, combining vision, microphones, and on-device GenAI with open-source Hugging Face model ecosystem
• Context-Aware Home Robotics (2:25) - First step toward robots that can see, hear, and interact using real AI in home environments
• Warehouse Revolution (2:39) - Amazon, Agado, and DHL running adaptive autonomous fleets that replan in real time and coordinate with humans
• Factory Automation Threshold (2:56) - Vision-guided robotic arms now handle variance with different parts, perfect placement, and fast reconfiguration
• Hospital Integration (3:10) - Service robots became infrastructure for medication delivery, UV disinfection, and surgical assistance
• Trust Over Capability (3:25) - The crucial shift in hospital robotics is about trust and uptime, not just capability
• Agricultural Decision Makers (3:30) - Autonomous tractors evolved from GPS followers to decision makers with precision spraying and selective harvesting
• Dangerous Environment Robots (3:44) - Inspection robots with improved autonomy now work in pipelines, mines, nuclear facilities, and disaster zones
• Optimus Deployment Scale (4:02) - Elon Musk predicting massive robot deployment, calling 10 billion by 2040 “a low number”
• Locomotion First (5:00) - Jensen Huang explains locomotion is the first challenge, followed by gross articulation, grasping, then fine motor skills
• Reasoning AI Inside Robots (5:29) - Robots will have great reasoning AI models resonant inside them, with cloud access for additional knowledge
• Embodiment Concept (5:52) - Jensen’s philosophical point: humans embody tools like cars and tennis rackets, extending their capabilities
• Multi-Environment AI (6:15) - AI models trained as humanoids could work as manipulators or self-driving cars - transferable intelligence

Mentions

Companies

• Nvidia (0:05) - Jensen Huang’s company, developing robotics technology
• Amazon (2:49) - Running adaptive autonomous robot fleets in warehouses
• Agado (2:49) - Running swarm logistics systems
• DHL (2:49) - Running adaptive robot systems for logistics
• Tesla (4:02) - Developing Optimus humanoid robots
• Hugging Face (2:22) - Model ecosystem used by Richie Mini for custom behaviors
• Turing Post (4:58) - The channel producing this video

Products & Technologies

• Richie Mini (2:08) - Open-source personal assistant robot with vision, microphones, and on-device GenAI
• Optimus (4:02) - Tesla’s humanoid robot, predicted to perform microsurgery better than any human surgeon
• Generative AI (1:45) - Technology driving phenomenal progress in robotics through simulation and training
• Autonomous Mobile Robots (2:39) - Warehouse robots that moved from scripted routes to adaptive fleets
• Vision-Guided Robotic Arms (2:56) - Factory automation crossing perception and dexterity thresholds

People

• Jensen Huang (0:05) - CEO of Nvidia, predicting human-level robot capabilities
• Elon Musk (4:10) - Discussing Optimus deployment scale, predicting over 10 billion robots by 2040

Surprising Quotes

“When do you think we’re going to get robots that actually have human level capabilities? Oh, this year.” — 0:15

“Don’t forget we already have robots that are almost boringly helpful.” — 2:00

“It’s Elon Musk. He’s rarely on time, but he always delivers.” — 4:10

“10 billion by 2040? That’s not… that’s a low number.” — 4:38

“When you sit inside a car, somehow you embody the car. You’re able to steer a car like it’s an extension of you. When you picked up a tennis racket, you embodied the tennis racket. Somehow your arm got longer.” — 5:57

Transcript

0:00 Hello my fellow humans. During his conference at CES, Jensen Huang, CEO of Nvidia, was asked when robots will have truly human level capabilities. And his reply was really surprising.

0:15 “When do you think we’re going to get robots that actually have human level capabilities?” “Yeah. Oh, this year. Yeah, this year. Because I know how fast the technology is moving.”

0:26 Because unfortunately robots at CES were not aware that Jensen Huang was so bullish. [Video montage of robot fails at CES with dramatic music]

1:37 Well, robots indeed still look funny and awkward and clumsy. But jokes aside, last year robots had a phenomenal progress driven by generative AI that help with simulation and training. And I have no doubt that what we’ve seen at CES is already old news because don’t forget we already have robots that are almost boringly helpful, which is exactly the personal assistants.

2:08 Robots like Richie Mini, who was the true star this year, now combine vision, microphones, and on-device GenAI, so they can recognize people, respond to voice, and run custom behaviors from the Hugging Face model ecosystem. This is the first step toward context-aware home robotics that can see, hear, and interact using real AI. Plus, Richie Mini is open sourced, so the community can develop whatever use cases they need it for.

2:39 Warehouses and logistics. Autonomous mobile robots move from scripted routes to adaptive fleets. They now replan in real time, coordinate with humans and handle mixed inventory. Amazon, Agado, and DHL are running systems that look less like automation and more like swarm logistics.

2:56 Factories - vision guided robotic arms cross the threshold in perception and dexterity. They can now handle variance, different parts in perfect placement, fast reconfiguration. This is why micro assembly and inspection lines are getting shorter, faster, and cheaper to redeploy.

3:10 Hospitals - service robots stopped being pilots and became infrastructure. Medication delivery, UV disinfection, internal logistics, even surgical assistance are now integrated into hospital workflows.

3:25 This shift is crucial. It’s amazing. It’s not capability. It’s trust and uptime.

3:30 Agriculture - autonomous tractors and harvesters move from GPS followers to decision makers. Precision spraying, selective harvesting, night operation, lower chemical use. This is robotics colliding with real world economics.

3:44 Dangerous environments - inspection robots have been helping in dangerous environments for a long time. Now they gain autonomy, endurance, and better sensing. Pipelines, mines, nuclear facilities, disaster zones. These systems now go where humans should not go at all. Not just where it is inconvenient.

4:02 And have you heard the latest about Optimus? Watch this short episode when Elon Musk said how many Optimuses and how fast they’re going to deploy. It’s Elon Musk. He’s rarely on time, but he always delivers.

4:13 “I mean, there’s a million of these things to figure out, but who’s going to have access to the first Optimus that does far far better micro surgery than any surgeon on Earth, but you’ve only manufactured the first 10,000 of them? How do you do it out?”

4:27 “I don’t think people understand how many robots there’s going to be.” “Yeah.” “Well, there’s a window - Saudi said 10 billion by 2040.” “You still on that path?” “Uh that’s not… that’s a low number.”

4:40 This is very exciting time. And I’m passing the mic back to Jensen Huang who has always something interesting to say and he will be perfect closing for this episode. Thank you for watching. Let me know what you think about robotics. Remember to like, subscribe, share and leave your comments. Humanly, yours from Turing Post.

5:00 “We’re developing technology in that area and I know the rest of the industry is doing so as well. Locomotion is hard. Um, but that’s making incredible progress. And so I think locomotion is the first, gets all gross articulation and grasping will be the second, and third will be fine motor skills - in that kind of, you know, not in that sequence but in the degree of difficulty.”

5:26 “The rest of it - cognition skills - is evolving very quickly as you know, and so we’re going to have really really great reasoning AI models that are resonant inside the robot. And so the robot will be able to reason very quickly, and anything that it wants to, you tap the AI in the cloud for additional knowledge you can. So you’re going to see some pretty amazing things.”

5:49 “And then the one thing that I will say in addition to that is these AI models are humanoid in nature. But don’t forget - when you’re humanoid, but when you sit inside a car, somehow you embody the car. You’re able to steer a car like it’s an extension of you. When you picked up a tennis racket, you embodied the tennis racket. Somehow your arm got longer.”

6:13 “And so the ability for AI to become multi-environment, meaning we train an AI model to be a humanoid, but it turns out it’s a perfectly good, you know, manipulator. It’s a perfectly good self-driving car. That day is probably going to come, too. And so I think that the next several years is going to be really exciting.”