Nvidia и OpenAI: союз на 20 миллиардов долларов (и это не предел)

Imagine you're selling shovels during a gold rush, and then you decide to invest your profits in the biggest mine nearby. That's exactly what Nvidia is doing right now. According to sources, Jensen Huang's company is close to investing an astronomical $20 billion in OpenAI.

In the context of the modern industry, this looks like an attempt to build a perpetual motion machine: Nvidia sells chips to OpenAI, which trains models on them, and then Nvidia returns part of its profits back to the company as investments so OpenAI can buy even more chips. This funding round resembles an arms race, where the stakes have long since exceeded reasonable limits. A few years ago, $20 billion would have seemed like the annual budget of a small country, but today it's simply an entry ticket to the highest league of artificial intelligence.

OpenAI needs money now more than ever. Developing GPT-5 and training the Sora video generator require not just talent, but endless farms of H100 graphics processors and the upcoming Blackwell. Why is Nvidia taking this step right now?

The answer lies in Sam Altman's ambitions. The OpenAI CEO has repeatedly hinted that he wants to reduce dependence on external hardware suppliers and has even considered creating his own network of chip manufacturing plants. For Nvidia, such a prospect is a direct existential threat.

By pouring billions into OpenAI, the "green" company is not just buying a stake in the business—they're buying loyalty. It's soft power meant to convince Altman that building his own factories is long and expensive, when ready-made solutions from his best friend in a leather jacket are right at hand. Don't forget about Microsoft either.

For a long time, Redmond was the only major benefactor of OpenAI. Now the shareholder structure is turning into a real "Avengers" club of the tech world. If Apple and Nvidia join the round, we'll have a unique alliance that will control virtually every level of the stack: from hardware and operating systems to end-user consumer services.

For smaller startups, that's bad news. The barrier to entry into the industry is now measured not by code quality, but by access to capital on a scale unavailable to many public corporations. Of course, behind the scenes of this deal loom the shadows of regulators.

In the US and Europe, antitrust authorities have already begun asking uncomfortable questions about the links between chip makers and model developers. A $20 billion investment is not just financial support—it's the creation of a vertically integrated giant that can strangle competition in its infancy. But it seems that in Silicon Valley right now, the rule is: build the monopoly first, sort out the legal issues later.

The bottom line: Nvidia is definitively transitioning from the status of "hardware supplier" to "industry architect." Will anyone be able to challenge the OpenAI and Nvidia pairing under these conditions, or are we witnessing the final stage of AI market formation for decades to come?