With SpaceX now publicly trading, attention in the IPO market has shifted decisively toward the two leading artificial intelligence laboratories: OpenAI and Anthropic. Both companies have confidentially filed paperwork with the SEC, setting the stage for what are likely to be two of the largest public offerings in history. The questions surrounding these listings go far beyond the companies themselves — they amount to a referendum on whether the entire AI vision for the economy is commercially sound.
What SpaceX's Debut Signals for AI Listings
SpaceX is not a direct comparison to OpenAI or Anthropic; it is a fundamentally different kind of company rather than a true foil. Yet there is one meaningful similarity: an IPO reveals how investors perceive a company's overall vision. In SpaceX's case, the strong interest, the oversubscription, and the high valuation all reflected investor belief in Elon Musk's broad vision for the company and its multiple revenue streams.
By analogy, the OpenAI listing will function primarily as a test of investor appetite for the entire AI vision as applied to the economy. A useful historical parallel is Google's 2004 IPO. That offering was effectively a test of investor confidence in the internet itself — a wager on whether the internet would become the fundamental distribution channel for the economy, for marketing, and for commerce. Google went public at roughly seven times revenue, and its value has since grown by a factor of 196.
OpenAI, by contrast, is expected to come to market at approximately 40 times its annualized revenue. That elevated multiple is itself an indication of how much confidence investors are placing in the idea that AI will become the foundational "intelligence layer" of the economy.
What Investors Are Actually Buying
A central question is what, precisely, an investor acquires when buying into OpenAI. At its core, the answer is frontier models — the most advanced AI systems available — which are clearly important to the direction the world is heading.
But OpenAI's strategic differentiator versus Anthropic is its ambition to own the entire infrastructure stack. OpenAI has roughly $1.4 trillion in commitments tied to data centers, compute, and energy. Anthropic, by comparison, sits somewhat behind on this front and has largely chosen the "rental" route for its compute rather than building and owning it.
This distinction frames the bet. Buying OpenAI means buying the full stack — model plus underlying infrastructure — whereas Anthropic represents a more capital-light approach. Whether OpenAI's enormous CapEx commitments ultimately pay off is genuinely uncertain. The company is currently generating revenue nowhere near the scale of its $1.4 trillion in commitments, which raises real questions about whether it can remain solvent and whether the strategy will work out. Even so, the fundamental asset being purchased remains the frontier models.
The Contrarian Case for Massive Infrastructure Spending
When OpenAI first announced these vast infrastructure commitments, the widespread reaction was that the move made no sense — that it was, in many people's view, among the least sensible decisions imaginable. The contrarian interpretation, however, is that the spending was in fact a smart and deliberately strategic move, and that OpenAI was ahead of the curve.
The reasoning runs as follows. If a company intends to deliver a frontier-grade product to hundreds of millions — and eventually billions — of people worldwide, along with many enterprises, it will require enormous scale: scale in compute, scale in power, and scale in the raw ability to actually serve these models. That cannot be achieved without controlling a vast supply of compute and energy.
By making these commitments over the past year or two, OpenAI appears to have anticipated exactly this need. Acting early allowed the company to lock in better rates than would be available by trying to purchase compute today — assuming such compute could even be obtained, given how difficult it currently is to buy any of this hardware. The volume of capital involved is undeniably large, and managing it against the revenue it must eventually generate remains an open challenge. But the underlying logic holds: to be a frontier model provider, a company needs far more compute than it has today. We are still in the early days of AI adoption, particularly as the world moves into the agentic era, which will demand even more compute and energy. On this reading, OpenAI saw that requirement coming and positioned itself accordingly.
The Metrics That Will Justify a Trillion-Dollar Valuation
A key debate is whether these trillion-dollar valuations can be justified at all. Evaluating that justification is not simply a matter of examining revenue or projected revenue. Instead, it hinges on a few critical questions:
First, does AI become fundamental infrastructure for the economy? In other words, is it used daily by consumers, and does it genuinely become the foundational intelligence layer of the economy — or does its value ultimately rest more narrowly on subscription revenue?
Second, what does the cost landscape look like? AI today carries very high capital costs in data centers and power. The crucial detail to scrutinize once OpenAI's numbers become public will be the company's projections for how these costs decline over time. The deeper issue is not merely whether people will use ChatGPT, but whether AI is actually a profitable business across the entire supply chain.
In short, the S1 filing must demonstrate not just usage and demand, but a credible path to profitability spanning the full chain of production, from chips and energy through to the end product.
OpenAI Versus Anthropic: Where the Competitive Advantage Lies
Markets will inevitably compare the two companies, and secondary market pricing currently suggests investors are favoring Anthropic. Assessing their relative competitive advantages reveals a few key factors.
OpenAI holds the stronger consumer brand, largely because ChatGPT is so widely known. Despite that, there is considerable momentum building toward Anthropic. The fundamental reason is Anthropic's stronghold among enterprise clients for its Claude models. Many large organizations — including Fortune 100 companies — appear to be favoring Claude at the enterprise level, and those relationships are extremely valuable. Enterprise subscriptions for such clients can run as high as roughly $1 million per year, making this a high-value segment.
That enterprise momentum is what gives Anthropic an edge. Ultimately, however, the competitive outcome will also depend on who benefits most from the overall AI supply chain — who controls data centers, processing usage, and the broader infrastructure. Brand strength favors OpenAI; enterprise traction favors Anthropic; and the resolution of the infrastructure question will help determine which advantage proves decisive.
The Open Questions
Neither company has set an official date for its IPO. What is clear is that these offerings will be defining, watershed moments for AI within the capital markets. The listings will test, on a public stage, whether AI truly becomes daily-use infrastructure, whether its steep capital costs can fall enough to yield durable profits, and whether owning the full stack or renting it proves to be the wiser path. The answers will shape not only the fortunes of OpenAI and Anthropic, but the market's verdict on the entire AI economy.
