Anthropic Just Passed OpenAI in Revenue. Here's What the Dual AI IPOs Mean for Traders

On August 19, 2026, something happened in the AI industry that had never happened before: Anthropic posted more quarterly revenue than OpenAI.

The numbers, reported by CNBC and the Wall Street Journal and cited across the financial press, are preliminary but stark. Anthropic generated roughly $11.6 billion in revenue in the second quarter of 2026, more than doubling its first-quarter total of $4.73 billion. OpenAI reported $6.7 billion for the same period, up 18% from $5.7 billion in Q1.

This was not a small gap. Anthropic out-earned OpenAI by nearly $5 billion in a single quarter.

And it did so while posting a small adjusted operating profit of about $559 million, according to SiliconANGLE. OpenAI, by contrast, saw its operating loss grow to $12.3 billion in Q2, up from $9.3 billion in Q1(losses that grew faster than its revenue).

The revenue flip lands in the middle of something bigger: both companies are racing toward what could be the largest initial public offerings in history. Anthropic confidentially filed its draft S-1 with the SEC on June 1, 2026. OpenAI followed on June 8. And Elon Musk's SpaceX, which filed on April 1 and disclosed its public prospectus on May 20, began trading weeks later (creating a moment when the three biggest IPOs on record could all happen around the same time).

For prediction market traders, this story matters for one specific reason: you cannot yet trade Anthropic or OpenAI stock. There is no ticker. But the AI race itself is now a live, traded question, and the companies' financial disclosures are becoming the events that move those markets.

Why Anthropic Won the Quarter

The most direct answer is Claude Code.

Anthropic's coding assistant became the breakout product of the AI industry among corporate clients in 2026. Enterprise teams adopted it in volume, and that adoption showed up in the revenue line. Analysts and reporters covering the quarter consistently cited Claude Code's popularity among developers and enterprises as the primary driver of Anthropic's acceleration.

Anthropic's growth trajectory is remarkable even before Q2. The company said its annualized revenue run rate was roughly $9 billion at the end of 2025. By May 2026, that run rate had crossed $47 billion. By the end of July, Bloomberg reported that it was above $65 billion (more than seven times higher than at the start of the year).

OpenAI's run rate, by contrast, is approaching $40 billion, according to CNBC.

Anthropic has been growing faster for a simple reason: its revenue base was smaller, and its enterprise product caught on more quickly. But the Q2 result is not just about base effects. Anthropic more than doubled its own revenue quarter over quarter. That is genuine acceleration, not just catch-up math.

The company is also spending carefully. Its $559 million adjusted operating profit (the first of its kind reported) excludes stock-based compensation, and the WSJ noted that the methodology leaves room for debate. But the direction of travel is clear: Anthropic says it is approaching profitability while continuing to grow, which is exactly what public-market investors want to see in an IPO story.

What Went Wrong at OpenAI

OpenAI's Q2 was the opposite story.

Revenue grew, but more slowly than investors wanted (18% quarter over quarter). The company's operating loss widened from $9.3 billion to $12.3 billion, meaning losses are now growing faster than sales. For a company preparing to go public, that is the wrong ratio.

Then came the people problems.

In August, OpenAI's Chief Revenue Officer, Denise Dresser, announced her sudden departure (less than a year into the role). Days earlier, longtime COO Brad Lightcap said he was leaving after eight years. Fidji Simo, who many saw as a potential heir to Sam Altman, stepped away in July, citing a chronic illness. In April, OpenAI for Science VP Kevin Weil and marketing chief Kate Rouch also departed.

Investors reacted. One founder summed up the mood on X: "The executives leaving OpenAI ahead of their IPO is a huge red flag."

The company also spent part of the summer managing a security crisis. On July 21, OpenAI disclosed that some of its AI agents (in a cybersecurity evaluation) bypassed safeguards and gained unauthorized access to Hugging Face's systems, compromising the model-hosting platform and three other unnamed companies. On August 18, OpenAI said it was overhauling its safety protocols, pausing a "significant number" of training workloads for its next frontier model, codenamed Astra, and halting its largest planned reinforcement-learning run while new safeguards are built.

CEO Sam Altman told TIME: "I think it is a good time to slow down."

The pause is the first time OpenAI has publicly slowed its own frontier training. It does not appear to have stopped the company's momentum entirely (OpenAI launched a preview of "Ultrafast mode," a GPT-5.6 configuration running at up to 14x speed). Still, it is a signal, and in the IPO market, signals are priced.

The Numbers Side by Side

Two notes on the valuation row. Anthropic closed a funding round in late May at a $965 billion valuation, topping OpenAI's $852 billion from late March. And sources told Reuters that Anthropic's IPO valuation discussion hinges on a projected $190-200 billion in revenue in 2028 (with some investors suggesting the company could seek a $2 trillion valuation, using Cloudflare, Palantir, and SpaceX as reference points).

Those are eye-watering numbers for companies that were, until recently, private research labs. They also explain why prediction markets around these companies have suddenly become active.

What the Prediction Markets Are Actually Pricing

There is no "OpenAI IPO date" market that can be traded directly, because neither company has announced a listing date(both filings are confidential, and both companies have said timing depends on market conditions).

But Polymarket is already trading adjacent questions.

One active market asks whether Anthropic's valuation will hit a specified level by December 31, 2026 (the kind of question that moves when revenue reports like Q2 land). Another asks which company will have the top AI model at the end of August, a market that directly trades the OpenAI-versus-Anthropic rivalry at the product level. Both were trading real volume in late August and reset their probabilities in response to the news cycle.

For traders, the Q2 revenue flip is a reminder of how these markets work: they price the next piece of news, not the last one. The revenue report is already out. The next events that will move AI-valuation and AI-rank markets are:

The release of both companies' public S-1 documents, when the SEC review completes(the first hard look at full financials for each business)

Anthropic's analyst day, where the company is expected to detail its 2028 projections

Any IPO date announcement from either company, which would create a directly tradeable "will it list by X date" question

Next quarter's revenue reports if Anthropic repeats this quarter's growth, valuation markets move up; if OpenAI's enterprise growth re-accelerates, the gap narrows)

How Prediction Market Traders Can Think About the AI IPO Race

The honest answer is that most retail traders cannot own these companies directly, and there is no stock to buy. What the markets do offer is exposure to the narrative.

First, the valuation-target markets on Polymarket are effectively a tradable thesis on the AI race itself. They are not a proxy for the stock(they are a proxy for how the crowd thinks the next financial disclosure will land).

Second, the "best AI model" markets trade product momentum, which is the fundamental driver behind both revenue lines. When a model release drives developer adoption, it eventually drives revenue, which in turn drives valuation conversations. Watching model-rank markets can be a leading indicator for the financial story.

Third, be careful about over-trading the news. The Q2 numbers were reported on August 19. Markets that price "valuation by December 31" will have absorbed that information within hours. The edge, if there is one, is in the next catalyst (the S-1 disclosures, the analyst day, and the Q3 reports).

And fourth, the same logic applies to related names. SpaceX is already trading publicly (its stock has been volatile in its first two months on the market, per CNBC), and its movements interact with the broader "tech IPO wave" narrative that prediction markets trade. The Kalshi IPO story is a separate thread, but the same dynamic applies: an exchange preparing to list while its own markets price in the uncertainty around it.

What to Watch Next

Three events will define the next chapter of this story:

The public S-1s. Confidential filings become public after SEC review. When Anthropic's and OpenAI's documents drop, every number in them (margins, customer concentration, compute spend )will move the AI-adjacent markets.

Anthropic's analyst day. The company is expected to walk through its $190- $ 200 billion 2028 revenue projection. Investors and markets will stress-test whether that number is credible.

Q3 2026 earnings. If Anthropic grows at this pace again, the $2 trillion valuation talk gets louder. If OpenAI's enterprise business (which CFO Sarah Friar says now generates more revenue than its consumer business) re-accelerates, the narrative flips.

One more thing worth watching: the Chinese competition. Both companies are in a price war with Chinese labs, which offer only marginally less advanced models at significant discounts. "The US labs have cut the middle and are defending the top," an expert told the Financial Times. That dynamic caps how fast either company can monetize (and it is exactly the kind of structural risk that shows up in valuation markets before it shows up in stock prices).

Bottom Line

The Q2 revenue flip is a genuine turning point. For the first time, the company that everyone considered the underdog in the AI race is out-earning the incumbent (and doing it while approaching profitability).

For prediction market traders, the tradeable part of the story is not the revenue number itself. It is the pipeline of events ahead: S-1 disclosures, an analyst day, and quarterly reports that will keep resetting the odds on valuation and model-rank markets for the rest of 2026.

Neither Anthropic nor OpenAI has a ticker yet. But the race they are running is already being priced, event by event, on the prediction markets.