OpenAI's IPO Dream Is Colliding With Cold Reality

The Boom That Leaves Its Author Behind

There is a certain irony in the AI investment supercycle: the company that arguably started it all is struggling to keep up with it. Global AI capital expenditure is on track to hit $725 billion this year, up 77% from $410 billion in 2025. Meta, Microsoft, Alphabet, and Amazon are collectively pouring a quarter of a trillion dollars into infrastructure. And yet OpenAI, the name most associated with the modern AI era, cannot close an $18 billion financing round and is burning through cash at a rate that makes its 2026 IPO target look optimistic at best.

This episode of Good Revenue puts that tension at the center of a broader economic picture that is more complicated than the stock market's daily moves suggest.

A Macro Backdrop That Should Not Be Ignored

Before getting to the corporate drama, it is worth noting the wider context Neeta flags. Since the Iran war began at the end of February, the Global Supply Chain Pressure Index tracked by the New York Federal Reserve has climbed to its highest point since the COVID-era peak of 2022. Saudi Aramco is warning that global gas and jet fuel stockpiles could reach critically low levels heading into summer if the Strait of Hormuz remains disrupted.

Financial Times sector data shows a clear bifurcation: AI, tech, and oil and gas are outperforming, while virtually every other sector is under pressure. For sophisticated investors, these supply chain signals are a reason for caution even as headline indices hold up. The fundamentals and the market narrative are not fully aligned.

Winners, Losers, and the Business Model Question

Among the hyperscalers, Google Cloud is the standout, posting a 63% revenue increase and pulling ahead of both Microsoft and Amazon in cloud growth. Microsoft's results are more complicated. Profitability and margins are being squeezed by the raw cost of compute, and the company's response has been to shift Azure AI customers to usage-based pricing.

This is a logical move from a cost-recovery standpoint, but it introduces a friction that Neeta identifies clearly: enterprise CFOs hate unpredictable bills. Consumption-based models are notoriously difficult to forecast quarter to quarter. If competitors offer more stable pricing structures, Microsoft could find itself in a position where the business model fix creates a new customer retention problem.

Meta, meanwhile, is profitable but in the doghouse with investors. The company is historically lean on capital expenditure, and its shareholders have not forgotten the metaverse, a multi-billion dollar bet that produced almost nothing. Heavy AI spending reopens that wound, even if the underlying AI economics are more defensible.

OpenAI: The Weight of Its Own Ambitions

OpenAI's situation is the most precarious. The company struck a $500 billion deal with Broadcom designed to reduce its dependence on NVIDIA for GPU supply, but the associated $18 billion financing tranche has not closed. Internal growth targets set for the end of 2025 were missed, and five months later they still have not been met. The projected cash burn through 2029 exceeds $115 billion. Reporting from The Information describes open conflict between the CFO, who does not believe the company is ready for a public offering, and CEO Sam Altman, who is pushing ahead.

Altman is also personally distracted. The Elon Musk lawsuit, which entered its third and final week as this episode was recorded, carries stakes that go well beyond reputational damage. Musk is seeking to reverse OpenAI's corporate structure back to a nonprofit, remove Altman and co-founder Greg Brockman, and claw back the $134 billion in value that principals and investors would receive from an IPO. Microsoft, which has invested $13 billion in the company since 2019, is a party to the suit. Satya Nadella testified that Musk never raised concerns about the Microsoft investment, a claim Musk disputes directly. Whatever the jury and judge decide, the outcome will materially affect the IPO timeline and the fortunes of everyone involved.

Apple, Intel, and the Government's Quiet Industrial Bet

On the quieter end of the news cycle, Apple has reportedly closed a chip supply agreement with Intel. The details are thin: no confirmed volumes, no clarity on whether the chips are destined for iPhones or other products. But the deal matters for a reason that goes beyond Apple's AI strategy under its new CEO John Ternus.

The U.S. government took a 10% ownership stake in Intel last year, an intervention that would have been almost unthinkable in prior administrations. Washington has since been actively encouraging NVIDIA, SpaceX, and Apple to direct business toward Intel. All three have now done so in some form. Whether this amounts to a genuine revival of Intel's fortunes, or just a political gesture, remains to be seen. But it signals that AI infrastructure is increasingly an area where government and corporate strategy are intersecting in ways that will shape competition for years ahead.

The $725 billion supercycle is real. The question is who actually benefits from it when the invoices come due.