Nvidia and OpenAI: where did 100 billion dollars go?

Remember those sensational headlines where Nvidia and OpenAI were practically merging in financial ecstasy against the backdrop of a hundred billion dollars? It seemed that Jensen Huang and Sam Altman were building a new empire, where one side provided infinite computational power and the other — infinite intelligence. But reality turned out to be far more mundane: the enormous pile of money that was whispered about in Silicon Valley back rooms and splashed across the front pages of business publications simply vanished from the radar. Investors who had already factored this growth into their strategies are now scratching their heads in bewilderment, trying to understand whether it was a clever maneuver or just wishful thinking.

To understand why this deal 'dissolved' right now, you need to look at the context of recent months. OpenAI is rapidly transforming from a cozy research startup into an aggressive corporate machine. Sam Altman no longer wants to simply buy chips — he wants to build his own factories, design his own architectures, and completely control the supply chain. Nvidia, for its part, has long ceased to be merely a 'shovel seller' for gold prospectors. The company has become the most expensive player on the planet, and such critical dependence on a single, albeit the most media-savvy client, is starting to look like a strategic vulnerability rather than an advantage.

The 100 billion dollar deal could have been either the pinnacle of this alliance or a gilded cage for both sides. If Nvidia had injected such funds, it would have effectively bought itself the right to a decisive voice in the future development of GPT and all subsequent iterations of neural networks. However, Altman, apparently, is not ready to hand over the keys to his kingdom. His ambitions to create a global network of data centers and semiconductor manufacturing capacity directly conflict with Nvidia's business model. Why would Jensen Huang feed someone who tomorrow could become his most dangerous competitor in the specialized hardware market?

Moreover, you cannot ignore the regulatory factor. The antitrust authorities of the United States and Europe are currently examining the AI industry through a very powerful microscope. A merger between the world's largest chip manufacturer and the undisputed leader in language models would have triggered such a wave of investigations and lawsuits that the companies would have been mired in legal bureaucracy for decades. It's simpler to quietly part ways and pretend that no official documents were signed than to explain to regulators why you're not trying to monopolize the future of human civilization. This is a classic example of how 'too big' companies become hostages to their own scale.

For the rest of the market, this is an extremely troubling signal. If the two main engines of the industry couldn't agree on money, it means the era of 'blind optimism' and endless checks is coming to an end. Now every dollar of investment will be checked for return on investment, and every partnership deal will be examined for hidden knives in the back. Nvidia will continue to sell its Blackwell to all takers, and OpenAI will have to seek new allies, perhaps among those who don't see them as a direct threat to their core business. We are entering a phase of harsh pragmatism, where old friends become competitors faster than model weight classes are updated.

Key Takeaway: The romance is over, a harsh reshuffling of spheres of influence has begun. Can OpenAI maintain its leadership without exclusive support from the planet's leading chipmaker?