Two confidential filings, a week apart, will force the most richly funded startups in history to prove that intelligence can be a business, not just a story.

For three years, the frontier AI labs answered to almost no one. Private capital arrived in torrents — $122 billion in a single OpenAI round this spring — and the only scoreboard that mattered was the benchmark leaderboard. That era is ending. Sometime this year, one of these companies will ring a bell on a stock exchange, and the questions will change overnight from "how smart is the model" to "where is the profit." The race to that bell is now on, and the two front-runners are running it in opposite directions.

In early June, Anthropic quietly submitted a confidential draft prospectus to the U.S. Securities and Exchange Commission. About a week later, on June 8, OpenAI confirmed it had done the same, with Goldman Sachs, Morgan Stanley, and JPMorgan lined up to lead the offering. Two companies that have spent the better part of a decade insisting the mission came before the money are now, within days of each other, preparing to be judged by public shareholders. The symmetry is neat. The substance underneath could not be more different.

What the AI IPO race is really testing

An initial public offering is a forcing function. It replaces the forgiving logic of venture capital — where a compelling narrative and a steep growth curve can excuse years of losses — with the unsentimental arithmetic of quarterly disclosure. Revenue, margins, cash burn, and the path to profit all become public, permanent, and comparable. For companies that have raised at valuations no operating business has ever justified, that transparency is the real event, not the ticker symbol.

Consider the numbers each company brings to the door. OpenAI closed a $122 billion funding round on March 31, valuing the company at $852 billion, with SoftBank co-leading and Amazon committing $50 billion while Nvidia and SoftBank each put in $30 billion. At the time, OpenAI said it was generating about $2 billion in revenue a month and had booked $13.1 billion the prior year. It was also, by its own account, still burning cash and not yet profitable.

Anthropic arrives with a smaller valuation but a startlingly different income statement. According to CNBC, Anthropic was on track for roughly $10.9 billion in second-quarter revenue — more than double the $4.8 billion it posted in the first quarter — and, remarkably, a first operating profit of around $559 million. Where OpenAI has told investors it does not expect to turn profitable until the end of the decade, after spending on the order of $600 billion on computing infrastructure, Anthropic appears to have reached the black years earlier, and at a fraction of the scale.

That is the fork in the road. Two companies, one race, two philosophies.

The scale bet versus the discipline bet

OpenAI's wager is on scale and dominance. Spend whatever it takes on compute, capture the largest share of consumers and enterprises, and trust that profitability will follow once the market consolidates around a winner. It is the Amazon playbook applied to intelligence: lose money for as long as the capital markets will fund it, build an unassailable position, and monetize from the mountaintop. The $852 billion valuation is a bet that OpenAI reaches that mountaintop first.

Anthropic's wager is on discipline. Grow fast, yes — doubling revenue in a quarter is not caution — but reach profitability early enough that the business is defensible on its own terms before public markets ever demand it. An operating profit ahead of an IPO is a statement: this is not a science project that happens to sell software; it is a company that happens to do frontier science. In a market suddenly nervous about whether AI economics ever close, that statement is worth a great deal.

Which theory Wall Street rewards will shape how the entire industry raises and spends money for years. If OpenAI debuts at or above a trillion dollars despite its losses, the lesson every founder learns is that scale forgives everything. If Anthropic's early profit earns it a premium multiple, the lesson is that discipline is a moat. The two IPOs are, in effect, a referendum on which story the AI boom is — and the outcome will echo through every term sheet in the sector.

What it means for each stakeholder

For public investors, this is the first real chance to price frontier AI. Until now, only a handful of insiders and sovereign funds set these valuations in private rounds where price discovery is thin and information is asymmetric. An IPO drags the numbers into daylight and lets the broad market vote. That vote will either validate the private marks or expose them — and given valuations that dwarf any precedent in software, the gap between the two could be violent in either direction.

For OpenAI, the confidential filing buys optionality without commitment: it can prepare for a listing while keeping timing, pricing, and even the final decision open. But the burn is the shadow over everything. Every quarter of losses becomes a public data point, and the "profit by 2030" timeline, once an internal projection, becomes a promise the market will hold it to. Going public means the runway is now visible to everyone counting it.

For Anthropic, profitability is the differentiator, and the reported Decart acquisition talks — a roughly $6 billion move aimed at cutting the cost of running its models — show it defending that edge deliberately. The risk is expectations. Reach the black early and investors will expect you to stay there while still spending enough on compute to keep up with a rival that treats losses as a strategy. Discipline is a harder promise to keep than growth.

For the rest of the market, these two filings reset the clock for everyone behind them. A successful debut re-rates every private AI company and reopens the exit window that has been effectively shut. A disappointing one — or a valuation that craters after the lockup expires — chills the whole asset class. The scores of AI startups raising today are, whether they admit it or not, hostages to how these two IPOs trade.

The takeaways for builders

Two lessons stand out for anyone building or funding in this cycle.

First, private valuations and public value are different currencies, and the exchange rate is set by strangers. A number agreed to by SoftBank and Amazon in a closed round is a negotiating outcome. A number set by a public market is a verdict. Confusing the two — building your plans, hires, and comp on the private mark — is how companies get caught out when the two finally meet.

Second, growth and profitability are not a single spectrum but a genuine choice, and the choice has consequences you inherit for years. OpenAI and Anthropic are proving that two companies in the same market, at the same moment, can rationally pick opposite strategies. The market does not reward the strategy; it rewards the one that fits the moment. Betting the company on scale in a patient market is wise; betting on scale when the mood turns to discipline is fatal, and the mood is not yours to control.

The bigger picture

The AI boom has been financed almost entirely on faith — faith that models keep improving, that adoption keeps compounding, and that profits arrive before the money runs out. Confidential filings are where faith meets fact. When the S-1s eventually go public and the redacted numbers turn real, the industry will get its first honest look at whether the economics of frontier AI actually work, or whether the valuations were a collective act of imagination.

Two companies are walking through that door within a week of each other, carrying opposite proofs. One says spend your way to dominance. The other says earn your way to trust. By this time next year, the market will have told us which it believes — and everyone still private will be building on whichever answer comes back.

Frequently Asked Questions

Have OpenAI and Anthropic filed for an IPO?

Both companies have confidentially submitted draft IPO paperwork to the SEC. Anthropic filed its confidential prospectus in early June 2026, and OpenAI confirmed its own confidential filing on June 8, 2026, with Goldman Sachs, Morgan Stanley, and JPMorgan leading, according to CNBC. Confidential filings do not commit either company to a specific date, price, or the decision to go public at all.

What is OpenAI's valuation heading into the AI IPO race?

OpenAI closed a $122 billion funding round on March 31, 2026, at a post-money valuation of $852 billion, per CNBC and Bloomberg. At that time the company reported roughly $2 billion in monthly revenue and $13.1 billion for the prior year, while still burning cash and remaining unprofitable.

How is Anthropic's financial position different?

According to CNBC, Anthropic was on track for about $10.9 billion in second-quarter 2026 revenue — more than double its first-quarter figure — and a first operating profit near $559 million. That contrasts with OpenAI, which has told investors it does not expect to reach profitability until around 2030.

Why does the AI IPO race matter beyond these two companies?

How OpenAI and Anthropic trade as public companies will set the benchmark for pricing every other AI startup. A strong debut reopens exit windows and validates private valuations; a weak one could chill funding across the sector. The two filings also test opposite strategies — scale-and-burn versus early profitability — and Wall Street's verdict will influence how the whole industry raises and spends.

Editor's note — sources: CNBC on OpenAI's confidential IPO filing; CNBC on OpenAI's $852B funding round; CNBC on Anthropic's Q2 2026 revenue and first operating profit; Fortune / Bloomberg on Anthropic's reported Decart acquisition talks. OpenAI's ~2030 profitability timeline and infrastructure spending are as reported to investors. All figures are as reported by the cited outlets.

Share this article: