By ByteBot
Amazon is in talks to invest up to $10 billion in OpenAI—one of the largest AI investments ever. But here’s the catch: the deal requires OpenAI to use Amazon’s Trainium chips and AWS infrastructure, funneling the investment money right back to Amazon. It’s vendor financing dressed up as venture capital, and it’s raising questions about whether the AI boom is sustainable or a bubble waiting to burst.
The Circular Deal Explained
Here’s how the circular financing works. Amazon transfers $10 billion to OpenAI’s account. OpenAI then commits to spending that money on Amazon’s Trainium AI chips and AWS data center capacity. Amazon books the spending as cloud revenue. The investment money goes in a loop: Amazon → OpenAI → back to Amazon.
This isn’t OpenAI’s first commitment to AWS. In November, OpenAI signed a $38 billion cloud computing deal with Amazon. The new $10 billion investment would layer on top of that existing relationship.
Critics warn this echoes the dot-com bubble. In March 2000, the Nasdaq fell 77% after similar circular deals between tech companies artificially inflated revenue. It took 15 years for the Nasdaq to recover.
Supporters counter that AI demand is real, unlike dot-com speculation. UBS analysts note that current AI vendor financing doesn’t show the same warning signs as past bubbles. However, the comparison is hard to ignore when investment money flows in circles.
Amazon’s Dual AI Strategy
Amazon isn’t betting on just OpenAI. The company already invested $8 billion in Anthropic—OpenAI’s main competitor—with the same chip requirements. Anthropic committed to use AWS Trainium and Inferentia chips for training future models.
That’s roughly $18 billion across two rival AI labs, both required to use Amazon chips. Amazon is playing both sides, hedging its bets to ensure it wins regardless of which AI lab dominates.
The real goal isn’t owning AI models. It’s validating Amazon’s Trainium chips and locking in AWS infrastructure revenue. Anshel Sag, analyst at Moor Insights & Strategy, explains: “Having OpenAI use Amazon’s Trainium chips sends a signal to enterprise customers that Trainium is safe, viable, and future-proof.”
If OpenAI—the most prominent AI lab—uses Trainium instead of Nvidia GPUs, it validates Amazon’s chip strategy. Enterprise customers will follow.
For OpenAI, the deal provides computing capacity and negotiating leverage. Charles Fitzgerald, a cloud infrastructure investor and former Microsoft employee, notes: “They can go back to Nvidia or Microsoft or Oracle and say, ‘If you don’t give us better terms, we’ll just use Amazon.'”
Nvidia capacity remains constrained, and Microsoft’s infrastructure is heavily committed to existing workloads. Therefore, OpenAI needs alternatives.
The Microsoft Complication
Here’s the twist: even with a $10 billion investment, Amazon can’t market OpenAI’s advanced models. Microsoft holds exclusive API rights to those models through the 2030s, part of a restructured partnership announced in October.
Microsoft owns 27% of OpenAI—larger than Amazon’s stake would be. OpenAI also committed to purchase an additional $250 billion of Azure services from Microsoft. Despite Amazon’s investment, Microsoft remains the primary partner.
This reveals Amazon’s true goal. The company isn’t buying access to OpenAI’s models. Instead, it’s buying infrastructure revenue and chip validation. Amazon gets recurring AWS and Trainium sales. Microsoft keeps the models.
Industry-Wide Circular Financing
OpenAI sits at the center of a web of circular deals totaling hundreds of billions of dollars. Beyond Amazon, OpenAI has signed similar arrangements with Nvidia, Oracle, CoreWeave, and AMD.
Nvidia is investing $100 billion in OpenAI, with OpenAI committing to deploy millions of Nvidia chips. Oracle signed a $300 billion cloud deal with OpenAI—one of the largest cloud contracts ever. Meanwhile, OpenAI invested $350 million in CoreWeave and committed up to $22.4 billion in cloud spending, while Nvidia owns 7% of CoreWeave.
The money flows in loops. Nvidia invests in OpenAI. OpenAI uses Oracle cloud services. Oracle buys Nvidia chips for data centers. Consequently, Nvidia’s investment money returns as chip sales.
All major AI companies are financially tied together. That creates systemic risk. If one major player fails, the effects could cascade across the industry.
Investors scarred by the dot-com bubble recognize the pattern. Vendor financing and circular deals were hallmarks of that era. Furthermore, the question is whether AI demand is real enough to sustain this financial structure—or whether it’s artificially inflating revenue.
Fitzgerald warns that training workloads represent “the worst business” for cloud providers due to high costs and rapid obsolescence. As a result, Amazon may be investing billions in the least profitable segment of AI infrastructure.
What This Means
For developers, this creates more cloud options for running AI workloads. AWS joins Azure and Oracle as infrastructure providers for frontier models. Moreover, competition could drive down costs.
However, it also raises lock-in questions. If you build on OpenAI’s APIs, which cloud provider ultimately controls your access? The AI labs increasingly depend on infrastructure partners, and those partnerships come with chip and capacity requirements.
For the AI industry, the pattern is clear: cloud giants control the future of AI, not the AI labs themselves. Frontier models require multi-billion dollar infrastructure commitments that only Amazon, Microsoft, Google, and Oracle can provide. Additionally, custom chips like Trainium and Google’s TPU are challenging Nvidia’s dominance.
Circular financing is becoming the norm. Traditional venture capital can’t fund AI at the scale required. Cloud providers step in as de facto investors, with strings attached.
The big unanswered question: is this sustainable, or a bubble? AI demand appears real, unlike dot-com speculation. Nevertheless, circular deals artificially inflate revenue and tie the industry’s fortunes together. Time will tell whether this financial structure supports innovation or collapses under its own weight.
One thing is certain: Amazon’s $10 billion bet on OpenAI isn’t really about OpenAI. It’s about making sure the money—and the AI industry—flows through AWS.
