According to Silicon Republic, OpenAI just signed a massive $38 billion deal with Amazon Web Services that gives them immediate access to AWS compute infrastructure featuring “hundreds of thousands” of Nvidia GPUs. The seven-year partnership will use AWS clusters with Nvidia GB200 and GB300 chips to train ChatGPT’s next-generation models, with potential expansion to “tens of millions” of CPUs. This comes right after OpenAI’s corporate restructuring that gave Microsoft a $135 billion stake and confirmed the company’s $500 billion valuation. CEO Sam Altman revealed they’ve already spent around $1 trillion on infrastructure, and Reuters reported they’re laying groundwork for a potential $1 trillion IPO.
Sponsored content — provided for informational and promotional purposes.
The multi-cloud chess game
Here’s what’s really interesting about this deal. OpenAI already works with Microsoft, Google, Oracle, and CoreWeave for cloud needs. So why add AWS to the mix? Basically, they’re playing the field to avoid getting locked into any single provider. And at $38 billion over seven years, this isn’t some small experiment – it’s a massive commitment that shows just how compute-hungry these AI models have become.
Think about it: hundreds of thousands of Nvidia chips just to get started. That’s an insane amount of processing power. But when Altman says they need “massive, reliable compute” for frontier AI, he’s not kidding. These models are getting so large that even tech giants need to pool resources across multiple cloud providers.
Restructuring and runway
The timing here is everything. This AWS deal comes right after OpenAI reshuffled its corporate structure and confirmed that $500 billion valuation. They’re clearly cleaning house and lining up their ducks before what looks like an inevitable IPO. A trillion-dollar valuation would make them one of the most valuable tech companies ever, right up there with the absolute giants.
But here’s the thing – can they actually justify these numbers? They’re spending like crazy, with that $1 trillion infrastructure figure Altman dropped last week. That’s more than some countries’ GDPs. The pressure to deliver groundbreaking AI that actually generates real revenue must be immense.
The reliability question
There’s one little detail that makes me raise an eyebrow. The article mentions that AWS had an outage last month that took down “several dozen websites” including banks and government services. When you’re running AI workloads that cost billions, can you afford those kinds of disruptions? I’m sure OpenAI has redundancy built in across their multiple cloud providers, but still – it makes you wonder about putting all your eggs in different baskets that might all have their own issues.
At the end of the day, this deal shows that the AI arms race is accelerating, not slowing down. We’re talking about investments at a scale we’ve rarely seen in tech history. The question isn’t whether AI will transform industries – it’s whether any company, even one valued at half a trillion dollars, can afford to stay in the game.
