What goes around comes around.
Monday’s announcement that Nvidia’s $100 billion investment in OpenAI marked one of the biggest AI infrastructure investments to date. The real beneficiary of this deal, however, might be Nvidia.
- OpenAI signed an eye-popping $300 billion contract with Oracle in mid-September to provide the model developer with computing power over the next five years.
- And Oracle, meanwhile, is feasting on Nvidia chips: The cloud giant struck a deal in May with Nvidia to purchase $40 billion worth of high-performance chips to power a data center in Abilene, Texas.
- Though the dollar amounts aren’t one-to-one, Nvidia is essentially investing it itself, allowing it to come out on top in this cycle.
But the benefits may extend beyond the money itself. The core of OpenAI’s AI infrastructure utilizes Nvidia technology, with plans for the first gigawatt deployment on the Nvidia Vera Rubin platform scheduled for 2026.
Nvidia might now be in the position to heavily influence OpenAI’s hardware strategy toward its own products and make the firm reluctant to turn to competitors, Scott Bickley, advisory fellow at Info-Tech Research Group, told The Deep View on Monday.
At a minimum, Bickley noted the relationship can ensure “any custom-built solutions are complementary, not competitive.”
Given OpenAI’s supremacy in the market, these kinds of deals could allow Nvidia to bend the future of AI infrastructure design in and of itself.