At CES 2025, Nvidia unveiled its bold vision for the future of AI, showcasing a $3,000 desk-sized supercomputer, robotics innovations, and foundational models that promise to transform industries and unlock a $3 trillion opportunity. Welcome to the next phase of AI.
Nvidia (NVDA) CEO Jensen Huang took to the stage of CES 2025 to reveal new RTX 50 GPUs promising better performance. He also revealed new Cosmos foundation' models that generate photo-realistic video, which can be used to train robots and self-driving cars.
However, the show's real star was a $3,000 AI supercomputer the size of a Mac Mini capable of handling 200B parameters. Nvidia proudly declared that the new computer, dubbed "Project Digits," is 1,000 times more powerful than the laptop in your bag, making it a direct competitor of Mac Studio.
what??
undefined el.cine (@EHuanglu) January 7, 2025
NVIDIA just dropped Project DIGITS, a $3,000 personal AI supercomputer that’s small enough to look like a Mac Mini but packs 1,000x the power of your average laptop.
Handles AI models with up to 200 BILLION parameters.
This is incredible.. pic.twitter.com/z4JOeFD2JI
A tiny supercomputer with big AI ambitions
The tiny AI PC is powered by the GB10 Grace Blackwell Superchip, featuring 128GB of unified memory and up to 4TB of NVMe storage. It delivers up to 1 petaflop of AI performance using FP4 precision and runs on Linux-based NVIDIA DGX OS, supporting frameworks like PyTorch and Jupyter notebooks.
Its unified memory architecture resembles Apple's M-series chips but integrates NVIDIA's software stack and GPU power. While ideal for running large LLMs locally, some are skeptical whether the memory bandwidth could keep up with concerns it may be bottlenecked in applications.
Although not released until May, you can register your interest early. Access to this much computing power that is small enough to fit on your desk or carry anywhere in your bag will undoubtedly change how we work with AI. But it also feels like a watershed moment where AI and data developers are encouraged to think bigger about what they can create.
Nvidia’s path to a $4 trillion market cap
Nvidia's race to become the first $4 trillion market cap company is underway. The digital metamorphosis from chip maker to AI superpower is there for all to see in Huang's vision for a future where ideas are much bigger than a tiny supercomputer.
Huang declared that robotics and AI agents represent a trillion-dollar opportunity in the physical AI space. An increase in Nvidia share prices suggests that investors agree.
Agents are a trillion dollar opportunity
undefined Ryan McNutt (@ryanmcnutty33) January 7, 2025
One of the largest convergences of technologies is about to change everything.
We are on the exponentialpic.twitter.com/JneqT647bu
Nvidia aims to make AI computing more accessible and localized by addressing a growing need for high-performance yet portable AI systems that can operate independently of the cloud. While still in its early stages, the company is primarily targeting developers.
Developers with access to a tiny AI supercomputer and foundational AI models on Nvidia's new Cosmos platform can generate data and build their own models for free. These models, trained on 20 million hours of video data, aim to do for the physical world what large language models like ChatGPT have done for natural language processing. Paired with Nvidia's Omniverse platform, developers can simulate real-world conditions for applications ranging from robotics to autonomous vehicles.
Humanoid robotics and autonomous vehicles
Fans of the Terminator movie franchise will know the next stage of AI will inevitably include the rise of machines. Not one to disappoint, Huang also introduced the Isaac GR00T Blueprint for robotics. The new toolset is designed to accelerate the development of humanoid robots. This platform lets developers quickly train robots using pre-recorded motions, reducing the traditionally time-intensive process of programming robot movements.
Jensen, at the CES 2025 stage with 14 humanoid robots standing in the background, announced NVIDIA Isaac GR00T Blueprint.
undefined The Humanoid Hub (@TheHumanoidHub) January 7, 2025
It's a simulation workflow for synthetic motion generation, enabling developers to create large datasets for training humanoids using imitation learning. pic.twitter.com/F8BOZLrktV
Strategic partnerships with Toyota, Aurora, and Continental reveal plans to power next-generation autonomous vehicles and self-driving trucks. For example, Toyota will utilize Nvidia's Drive AGX hardware and software for its upcoming autonomous vehicle line, while Aurora and Continental plan to deploy Nvidia's technology for thousands of driverless trucks starting in 2027.
At a time when many businesses are trying to determine the ROI of expensive AI projects, Nvidia is bucking the trend with ambitious targets for its automotive industry to grow to $6 billion in revenue by 2026.
Developers hold the key to Nvidia’s AI-powered future
It's important to remember that many shiny new technologies unveiled on the show floor at CES never see the light of day. If we reign in our enthusiasm, robotics and automotive will contribute a modest share of Nvidia's overall revenue – $449 million in its most recent quarter compared to $30.8 billion from its AI data center chips.
Nvidia faces mounting competition from major customers like Amazon and Microsoft, which are developing their own in-house AI chips for data centers. If we zoom out from the hype, Nvidia is attempting a strategic diversification to prove the company is more than just an AI chip maker on stage at one of the biggest tech events in the world.
With lofty ambitions of becoming an AI superpower, the company has its sights firmly set on unlocking new markets. But to do this, it needs to enable developers to simulate and build physical AI systems at scale. As Microsoft's Steve Ballmer passionately shouted twenty years ago, it's not technology that creates a new future but developers, developers, developers.
At CES, Nvidia has shared a vision for a future filled with AI robotics, self-driving vehicles, and personalized AI tools. But the question remains whether Nvidia can overcome the challenges of cost, scalability, and competition to unlock the full potential of this multitrillion-dollar opportunity or whether the only thing more futuristic than NVIDIA's new tech unveiled at CES 2025 was Jensen Huang's shiny $9,000 jacket.
Your email address will not be published. Required fields are markedmarked