Paul van Gerven
21 March

Not content with being a ‘commodity’ chip supplier, Nvidia maneuvers to become an AI platform company.

Dubbed “AI Woodstock” by one analyst, Nvidia unveiled a wide range of fresh plans and new products at the company’s annual GTC conference held Monday in a massive sports arena in San Jose. As its GPUs have become the ubiquitous engines that power the generative AI revolution, the spectacularly successful company introduced new hardware (see inset below) but also made it abundantly clear that it’s not just chips it wants to sell.

As profitable as the sale of chips has been – on the wings of tripled sales, Nvidia’s market value rose 2 trillion dollars recently – the real money is further downstream. A February survey by the American Census Bureau showed 12 percent of US firms of all sizes use some form of AI or plan to do so within the next six months. Wedbush analyst Dan Ives estimated that for every dollar spent on an Nvidia GPU, there could be 10-12 dollars in additional spending across related software, IT services and infrastructure tech.

Presenting itself as a one-stop AI shop of sorts, Nvidia announced the expansion of its software offerings with tools to help build AI applications that run on the chips. “We’re effectively an AI foundry. We’ll do for you and the industry on AI what TSMC does for us building chips,” Nvidia CEO Jensen Huang said during his 2-hour keynote presentation at the GTC.

Huang touted Nvidia Inference Microservices, an API-based platform specifically designed to facilitate the creation of genAI applications using his company’s software and hardware tools. Currently, his biggest customers are cloud computing giants such as Amazon and Microsoft and builders of genAI models such as OpenAI, all of which are looking to reduce their reliance on Nvidia. By offering tailored solutions, this customer base can be expanded to firms across a wide range of industries. Additionally, “these new microservices allow for the creation of an entirely new revenue stream and business strategy for Nvidia because they can be licensed on a per GPU/per hour basis,” Techanalysis Research chief Bob O’Donnell points out.

Bits&Chips event sponsor registration-early bird

Nvidia has rolled out sets of tools for automotive, robotics, healthcare and 6G research, and many other domains will likely follow. Huang: “So long as there’s some structure where we can apply some patterns, we can learn the patterns. And if we can learn the patterns, we can understand the meaning. When we understand the meaning, we can generate it as well.”

Main picture credit: Nvidia