OpenAI quietly dropped a bombshell on Monday: it's filed a confidential S-1 with the SEC, taking its first formal step toward going public. The timing is telling. This comes barely ten days after arch-rival Anthropic made the same move, which means the AI industry's two heavyweight contenders are now sprinting toward the public markets simultaneously — with SpaceX reportedly waiting in the wings at a $1.75 trillion valuation. If this feels like the dot-com era decided to have a sequel, you're not wrong.

The "confidential" part matters. Filing under the SEC's confidential review process lets OpenAI kick the tires on going public without spilling its financial guts to the world just yet. No pricing, no raise target, no public balance sheet — not until they're ready. OpenAI itself acknowledged it's in no rush, noting in a blog post that "it may be a while" before any actual listing, given that some things are still easier to execute as a private company. Translation: there's still internal restructuring and strategic maneuvering that goes down smoother without quarterly earnings calls breathing down your neck.

The Mission Statement Nobody Asked For (But Regulators Might Notice)

Here's where it gets interesting. Around the same time as the confidential filing, OpenAI published a sprawling philosophical manifesto about AGI, humanity's future, and its own noble intentions. Normally, companies entering a quiet period — the legally sensitive window around an IPO — tread very carefully around public communications. OpenAI, apparently, did not get that memo. Or more likely, they got it and decided the current regulatory climate made it optional.

The SEC under the Trump administration has adopted a noticeably lighter touch toward tech companies than its predecessors. OpenAI seems to be reading that room clearly. Whether that's savvy or reckless probably depends on which lawyer you ask.

The Valuation vs. The Reality Check

Let's talk numbers, because this is where the story gets genuinely complicated. OpenAI was last valued at $852 billion in a funding round that raised $122 billion — the largest single raise in Silicon Valley history, $3 billion of which came directly from retail investors. That's an eye-watering number. It's also the kind of number that comes with equally eye-watering expectations attached.

Here's the uncomfortable math. OpenAI projects it will spend approximately that same $122 billion on compute alone in 2028. Even assuming it doubles revenue year-over-year, the company still expects to burn roughly $85 billion that year. You read that correctly. Public market investors are being asked to price a business that, by its own internal forecasts, won't generate more cash than it consumes for at least four more years. That's not a business plan — that's a very expensive bet on the future of intelligence itself.

OpenAI's CFO Sarah Friar has reportedly flagged concerns about whether the company can sustainably support its data center spending commitments. When your own finance chief is raising red flags before the IPO roadshow, that's not a footnote — that's a headline.

Anthropic Looks Better on Paper — For Now

Anthropic, by contrast, has been telling a much tidier story. The company claims it's approaching its first quarterly profit, which in this industry is roughly the equivalent of summiting Everest in flip-flops. That said, "approaching profitability" and "sustainably profitable" are very different things, and Anthropic isn't exactly running lean. A recent $65 billion funding round plus a reported $36 billion in chip-allocated debt means the burn rate, while arguably more controlled, isn't modest by any stretch.

Still, secondary market investors have noticed the relative contrast. On Forge Global, Anthropic recently hit a $1 trillion valuation — briefly eclipsing OpenAI's $880 billion mark on the same platform. Year-to-date appreciation tells an even starker story: Anthropic is up roughly 123% versus OpenAI's 11.3%. That's not a gap; that's a chasm.

What Actually Matters for Builders

If you're building on top of either of these platforms, the IPO race isn't just financial theater — it has real implications for product stability, pricing, and API reliability. Public companies face quarterly pressure that private ones don't. That pressure has a way of trickling down into decisions about compute allocation, model deprecation timelines, and how aggressively they monetize the developer ecosystem.

The structural challenge both companies share is one the whole industry is quietly wrestling with: training large language models costs more than the revenue those models currently generate. SpaceX's AI spending trajectory illustrates the same dynamic. At scale, inference and training costs are genuinely brutal, and no amount of benchmark press releases changes the underlying unit economics.

OpenAI also recently missed its own internal targets for user growth and revenue — not by a catastrophic margin, but enough that it's getting reported. Missing targets heading into an IPO is the kind of thing that gets amplified by public market scrutiny in ways that private company boards can paper over.

The Bottom Line

OpenAI going public isn't a surprise — it was always a matter of when, not if. But the timing, the financial profile, and the competitive pressure from Anthropic make this a more complicated offering than the headline valuation suggests. The company is asking investors to believe that spending at astronomical scale today will translate into category-defining returns tomorrow. Maybe it will. The underlying technology is genuinely impressive, and the market position is real.

But if you're evaluating this as an investor rather than a believer, you'll want to read the S-1 very carefully when it finally goes public. The gap between a company's mission statement and its cash flow statement tends to be where the interesting footnotes live.