Can the Stock Market Swallow Anthropic, SpaceX and OpenAI?

On a Tuesday morning in late May, the line outside a single Anthropic office in San Francisco stretched around the block for a free lunch talk. That's not the story. The story is that 22% of the people in that line were not engineers. They were bankers. And those bankers were not there for the sandwiches.

They were there to price a ticket that doesn't exist yet. Anthropic hasn't filed an S-1. SpaceX laughed at the idea in its last all-hands. OpenAI's board is still figuring out what a for-profit even means. Yet the pressure to go public is already reshaping all three companies from the outside in.

The conventional wisdom says that going public is the natural endgame for any successful startup. But that wisdom dates from a world where capital was scarce and liquidity was a reward for patience. In 2026, the reverse is true: capital is drowning in private markets, and the real prize is staying private long enough to finish what you started.

Let me give you a number: $1.5 trillion. That's the combined valuation these three companies would command if they went public today, based on secondary market trades tracked by Caplight and Forge Global. To put that in perspective, that's larger than the entire market cap of the S&P 500 Energy sector.

The problem isn't whether the stock market can swallow them. The problem is whether they can survive being swallowed.

Here's the cognitive reversal most people miss. Public markets don't actually want these companies to succeed at their stated missions. Public markets want predictable quarterly earnings growth. Anthropic's mission is to build safe AGI. SpaceX's mission is to colonize Mars. OpenAI's mission, at least on paper, is to ensure AGI benefits all of humanity. None of those missions are compatible with a 13-F filing every quarter.

Consider SpaceX. In 2024, it posted $8.7 billion in revenue. But Starship development costs were $5 billion annually, and Starlink's capital expenditure is still eating cash. A public SpaceX would face immediate pressure to cut Starship spending and milk Starlink for margins. The bean counters would win. Mars would wait.

This is not speculation. It's historical pattern. When Google went public in 2004, it had a 10-year horizon. By 2010, Wall Street had forced it to prioritize Search revenue over moonshots. Waymo was funded grudgingly, not strategically. The same pattern played out at Amazon, which only got away with its long-term bets because Bezos owned 17% of the voting stock and could ignore shareholders for two decades.

None of these three companies have that luxury. OpenAI's CEO Sam Altman owns less than 2% of the company. Dario Amodei's stake in Anthropic is likely smaller. Elon Musk owns about 42% of SpaceX, but even he has fiduciary duties to other investors that constrain his worst impulses.

The irony here is thick enough to cut. These companies are building the most consequential technologies since the atomic bomb—general intelligence, interplanetary transport, God-knows-what-OpenAI-is-really-building—but their biggest existential threat isn't regulation, competition, or technical failure. It's the S&P 500 inclusion criteria.

Let me explain how the hidden mechanics work, because they're not obvious. The moment a company files for IPO, the SEC requires a quiet period that restricts what executives can say. For a company like Anthropic, where safety research is public and the CEO regularly discusses alignment risks, this is a cultural disaster. Once public, they become subject to Regulation FD (Fair Disclosure), which means every major update must be shared with all investors simultaneously. No more small dinner conversations with safety researchers. No more changing course mid-flight because the board saw something scary.

But the real killer is the quarterly earnings call. These calls are not about information. They are about expectation management. A public company that announces a major safety pause that delays product revenue by six months will see its stock drop 30% in a day. The board will not let that happen twice. The pause becomes three months. Then one month. Then a blog post that says "we've decided to proceed with deployment while continuing our research."

This is not a conspiracy theory. It's the basic incentive structure of public markets, documented in over 200 academic papers on short-termism. McKinsey found in 2023 that 87% of executives say they would sacrifice long-term value creation to meet quarterly earnings targets. That number has gotten worse, not better.

The counterargument I hear most often is that these companies are unique enough to demand special treatment. Alphabet and Meta both created dual-class share structures to retain founder control. OpenAI could theoretically do the same. But the Nasdaq and NYSE have tightened rules on dual-class listings since the 2021 SPAC boom. And the real problem is that even with voting control, a founder can't stop a shareholder lawsuit if the stock drops 40% after a missed quarter. Musk learned this the hard way with Tesla.

So what actually happens? I think the most likely outcome is that none of these three companies go public in the traditional sense. Instead, watch for a new structure I'll call the "perpetual private mega-corporation."

SpaceX already looks like one. It raises capital through tender offers every 12-18 months, giving employees liquidity without a public listing. In April, it closed a $3.2 billion secondary round at a $250 billion valuation with no IPO date mentioned. The investors were sovereign wealth funds and pension funds who explicitly agreed to a 7-year lockup on resale. That's the new normal.

OpenAI is more complicated. Its capped-profit structure is already a hybrid, but it needs enormous capital for compute infrastructure. The rumored $400 billion SoftBank deal earlier this year wasn't for equity—it was a special purpose vehicle with debt-like returns. OpenAI is essentially inventing a new asset class to avoid the public markets.

Anthropic is the most vulnerable, ironically because it's the most principled. Its "Long-Term Benefit Trust" structure is supposed to protect its safety mission, but the trust has no voting power over financial decisions. If the company needs cash and the only option is going public, the trust becomes a decoration. I'm watching their Series G terms closely.

The forward view is this: within three years, the SEC will be forced to create a new listing category for "National Interest Companies" that exempts them from quarterly reporting requirements in exchange for enhanced disclosure on long-term risks. Or the companies will simply refuse to list, and the stock market will watch the most important value creation of the century happen entirely outside its walls.

Neither option is clean. But the one thing I'm certain of is that the old IPO playbook is dead for the companies that matter most. The stock market doesn't get to decide the future of intelligence and space travel. It just doesn't know that yet.