According to CNBC, Jim Cramer has publicly implored Amazon not to engage in a potential $10 billion investment in OpenAI. The reported deal would see Amazon invest that massive sum in exchange for OpenAI agreeing to use Amazon’s custom Trainium AI chips for its computing needs. Cramer, speaking on CNBC’s “Squawk on the Street” and during his Club’s Morning Meeting on Wednesday, called such arrangements “sham-like” and “not real,” comparing them directly to the speculative deals that fueled the dotcom bubble before its crash over two decades ago. He specifically criticized the circular nature of the deal, suggesting Amazon is so desperate to sell its chips that it would effectively pay a customer $10 billion to buy them. This comes as OpenAI has been on a deal spree in 2025, securing $1.4 trillion in total infrastructure commitments from various tech giants.
Cramer’s Core Criticism
Here’s the thing: Cramer isn’t just yelling about a big number. His warning cuts to a specific, risky practice. He’s calling out what looks like a circular, “you scratch my back, I’ll scratch yours” capital arrangement. The fear is that Amazon‘s cloud division (AWS) books a huge, headline-grabbing “sale” of its chips, while its corporate treasury writes a check to the “customer.” It makes the AI infrastructure arms race look even hotter on paper, but does it represent genuine, organic demand? Or is it just money being moved around to inflate the perception of a market? Cramer’s point is that this feels exactly like the kind of interconnected, speculative financing that blew up in 2000. When the music stops, who’s left holding the bag?
The Broader AI Arms Race Context
Now, this isn’t happening in a vacuum. Every cloud giant—Amazon, Microsoft, Google—is desperate to lock in the next generation of AI winners as flagship customers. They’re all building custom chips (like Trainium or Google’s TPUs) to reduce reliance on Nvidia and to create sticky, long-term platform dependencies. So from a pure competitive strategy standpoint, Amazon cutting a huge check to land OpenAI makes a brutal kind of sense. It’s a loss-leader on steroids. But that’s Cramer’s whole warning! The dotcom bubble was also fueled by strategies that made “sense” in the moment—spend wildly to gain market share, damn the profits. The market tolerated it until, very suddenly, it didn’t. And when the pivot to profitability was demanded, countless companies couldn’t make the turn.
Stakeholder Impact and The Bubble Question
So who does this affect if it goes through? For other AI startups, it sets a wild precedent. Your funding round might not just be from VCs, but from a cloud provider paying you to be a client. For the chip market, it further heats up the battle between custom silicon and merchant silicon like Nvidia’s. For enterprises buying AI services, it might not change much in the short term, but it concentrates even more power and capital into a few hyperscalers. But the big, looming question Cramer is asking is for investors: Is this sustainable? A $1.4 trillion commitment from OpenAI is an almost incomprehensible sum. It’s leveraged betting on a scale that screams “peak.” If the expected AI revenue tsunami is even slightly delayed or smaller than forecast, the entire house of cards gets shaky. Cramer’s essentially saying the market will eventually punish this excess, just like it did in 2000. The only question is when.
