OpenAI’s overdraft continued its upward trajectory on Monday when the company signed a multi-year $38 billion contract with AWS to have it run its AI workloads.
The latest spending spree adds to the incremental $250 billion of Azure services it pledged to buy last week, and, of course, to the commitment it has made towards building Stargate data centers with Oracle and other partners.
According to a release from OpenAI, the dollars heading to AWS from the organization, whose annual revenues are projected to be a mere $13 billion, will be used to “run and scale” its AI workloads, beginning immediately.
The release went on to say that AWS will provide the company with “Amazon EC2 UltraServers, featuring hundreds of thousands of chips, and the ability to scale to tens of millions of CPUs for its advanced generative AI workloads.”
Deals based on ‘unproven growth’
In reacting to the announcement, Sanchit Vir Gogia, the chief analyst, founder, and CEO of Greyhound Research, said, “OpenAI’s cloud investments have reached a level that no longer correlates with its revenue profile. Despite claims of scaling for future demand, these multi-cloud arrangements are best understood as a form of insurance rather than expansion.”
The business case here, he said, is not grounded in present usage or profitability, but rather in the anticipation of massive, yet unproven, growth. “OpenAI is distributing its infrastructure load across AWS, Microsoft, Oracle, and Google,” he explained. “This is not about optimizing workloads; it is about ensuring that no single failure point can threaten its ability to operate.”
The aim, said Gogia, “is continuity, not cost efficiency. These deals are forward leaning, relying on revenue forecasts that remain speculative. In that context, OpenAI must continue to draw heavily on outside capital, whether through venture rounds, debt, or a future public offering.”
He pointed out, “the company’s recent legal and corporate restructuring was designed to open the doors to that capital. Removing Microsoft’s exclusivity makes room for more vendors but also signals that no one provider can meet OpenAI’s demands. In several cases, suppliers are stepping in with financing arrangements that link product sales to future performance. While these strategies help close funding gaps, they introduce fragility. What looks like revenue is often pre-paid consumption, not realized margin.”
Execution risks, he said, add to the concern. “Building and energizing enough data centers to meet OpenAI’s projected needs is not a function of ambition alone. It requires grid access, cooling capacity, and regional stability. Microsoft has acknowledged that it lacks the power infrastructure to fully deploy the GPUs it owns. Without physical readiness, all of these agreements sit on shaky ground.”
Lots of equity swapping going on
Scott Bickley, advisory fellow at Info-Tech Research Group, said he has not only been astounded by the funding announcements over the last few months, but is also appalled, primarily, he said, “because of the disconnect to what this does to the underlying technology stocks and their market prices versus where the technology is at from a development and ROI perspective … and from a boots on the ground perspective.”
He added that while the financial pledges involve “huge, staggering numbers, most of them are tied up in ways that are not necessarily going to require all the cash to come from OpenAI. In a lot of cases, there is equity swapping. You have these highly discounted exchanges going on with the hyperscalers, with capacity and this whole circular financing loop taking place.”
The organization’s vision, said Bickley, “is huge, and to me, it’s either they go big or they go bust, because they are not going to make this money back off ChatGPT.”
OpenAI, he said, is “going to make [it] back if they become the new web browser, the new vehicle for search, the new vehicle for ad revenues where AI agents may be able to handle sophisticated purchase requests.”
Another theory, he said, is that OpenAI is also “building a vertical distribution network where they want to control everything — the infrastructure, the application layer, the interface layer, and ironically, the hyperscalers may be funding a company that wants to take them out over time. If they buy enough time to realize that grand vision, then the revenues can get big enough to start paying back on these investments, but those are a lot of big ‘ifs’.”
In addition, said Gogia, “as OpenAI secures early access to large shares of advanced GPUs, other organizations are being forced to adjust. Some are shifting their cloud workloads to alternative vendors, while others are rebalancing infrastructure across regions in search of availability. The vision of cloud as an open, flexible resource is becoming harder to realize, especially for businesses without the scale or capital to compete.”
This shift, he said, “signals a new phase for AI development. What began as a software-driven movement is becoming increasingly shaped by infrastructure and capital. OpenAI’s role is no longer limited to innovation at the application layer. It now serves as a major force in how the physical backbone of AI is being built.”
That level of involvement, said Gogia, “brings weight and responsibility. If the assumptions supporting these expansive projects do not convert into stable revenue, the strain will not stop at OpenAI. It could disrupt suppliers, unsettle financing structures, and affect how enterprises access computing resources.”
Maintaining balance across the ecosystem, he said, “requires more than technical advancement. It will depend on steady execution, practical delivery timelines, and the financial strength to support what has been promised. Ambition alone will not carry this model forward.”