Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
Nvidia is a “technology company”, not a “consumer” or “enterprise” company, as emphasized by CEO Jensen Huang. What do you mean, exactly? Nvidia does not want consumers to spend hundreds or thousands of dollars on the new, expensive RTX 50 series GPU? Don’t they want more companies to buy their AI training chips? Nvidia is the kind of company with many fingers in many pies. To hear Huang say, if the crust of those pies are the company’s chips, then AI is the filling.
“Our technological influence will impact the future of consumer platforms,” Huang, dressed in his typical black jacket and hot chest of AI hype, said in a Q&A with reporters a day after it opened of the CES. But how does a company like Nvidia fund everything those epic AI experiments? The training chips H100 AI made Nvidia such a technological powerhouse in the last two years, with a few stumbles along the way. But Amazon and other companies are trying to create alternatives to cut into Nvidia’s monopoly. What should happen if the competition cuts the spee short?
“We have to meet customers where they are,” Huang said. Part of that is helping companies build “agent AI,” AKA multiple AI models capable of performing complex tasks. That includes several AI toolkits made to throw companies a bone. While the H100 made Nvidia great, and RTX keeps gamers coming back, it wants its new one $3,000 “Project Digits” AI processing hub to open “a new universe” for those who can use it. Who will use it? Nvidia said it’s a tool for researchers, scientists, and maybe students — or at least those who find themselves sitting on $3,000 in their $1.50 bowl of instant ramen that they’re eating for dinner for the fifth night in a row. .
Nvidia make sure you know about the RTX 5090’s 3,352 TOPS of AI performance. So, Huang’s company dropped details on several software initiatives — both gaming and non-gaming related. None of his statements were more confusing than his “world foundation” AI models. These models should be able to train in real-life environments, which could be used to help autonomous vehicles or robots navigate their environment. It’s a lot of future technology, and Huang admitted that he failed to deliver better to a crowd that was mainly coming to see cool new GPUs.
“(The world foundation model) understands things like friction, inertia, grasping, object presence and elements, geometric and spatial intelligence,” he said. “You know, the things that children know. They understand the physical world in a way that language models don’t know.
Huang opened CES 2025 on January 6 with a keynote that packed the Michelob Ultra arena in Las Vegas’ Mandalay Bay casino. There was certainly a large part of gamers who had come to see the latest cards of the RTX 50 series in the flesh, but more were there to see how a profitable company like Nvidia is moving forward. RTX and Project Digits drew screams and shouts from the crowd. Spending half his time talking about his world foundation model, the audience didn’t seem all that enthused.
It shows how annoying AI messaging can be, especially for a company that owes much of its popularity to the attentive population of PC gamers. There’s been so much talk about AI that it’s easy to forget that Nvidia was in this game years before ChatGPT arrived on the scene. Nvidia’s in-game AI upscaling technology, DLSS, has been around for almost six years, improving all the time, and is now one of the best AI-upscalers in games, although it is limited by its exclusivity to Nvidia cards. It was good before the advent of generative AI. Now, Nvidia promises that Transformer models will also enhance upscaling and ray reconstruction.
On top, the promoted multi-frame gene could possibly grant four times the performance for the 50 series GPUs, at least if the game supports it. It is a good one for those who can afford the new RTX 50 series. The RTX 5090 tops out at $2,000. The gamers who benefit most from frame gen are those who can only afford a lower GPU. Huang declined to offer any hints about an RTX 5050 or 5060, joking “We’ve announced four cards, and you want more?”
The global foundation model is just a prototype, like much of Nvidia’s new AI software on display to the public. The real questions are, when will it be ready for primetime, and who will end up using it? Nvidia has shown strange NPC AI, chatbots in play, AI nursesand an audio generator last year. This year, he wants to flourish with his world foundation model, plus a bunch of AI “microservices”, including a strange animated talking head that should serve as your PC’s always-on assistant. Maybe some of these will stick. In cases where Nvidia hopes AI will replace nurses or audio engineers, we hope that doesn’t happen.
Huang considers Nvidia “a small company” with 32,000 employees worldwide. Yes, this is less than half of the staff that Meta has, but you can not think that it is small in terms of the influence of the market for AI training chips. Because of its market position, it has an overwhelming influence in the technology industry. The more people use AI, the more people need to buy their AI-specific GPUs, plus any of their other AI software. If everyone buys their own in-house AI treatment chipthey don’t need external data centers and external chatbots. Nvidia, like any technology company, just needs to find a use for AI beyond replacing all our jobs.
