Skip to content
Finance

The New Market Mover: How Private AI Labs' Revenue Became Wall Street's Most Watched Number

A single OpenAI revenue figure knocked billions off chip stocks last week, and Anthropic is about to publish its own numbers. Private AI lab finances are now a public-market event.

Raj ·Founder & Editor
Published in Finance · 4 min read ·
Glowing AI brain above circuit-board chips with a falling red stock chart
Glowing AI brain above circuit-board chips with a falling red stock chart
In this article

    A revenue number that moved the Nasdaq

    On Thursday, October 8, AI stocks fell hard. The trigger wasn't an earnings miss at a listed company or a new rate decision. It was a number from a company that isn't publicly traded at all.

    CNBC confirmed that OpenAI told investors it reached about $50 billion in annualized revenue at the end of September. That was well below the roughly $68 billion figure widely reported late last month. According to a person familiar with the matter, the larger number counted gross revenue from OpenAI's partners, which made it easier to compare with rival Anthropic. The $50 billion number leaves that partner revenue out.

    The market didn't wait for the explanation to settle in. Nvidia fell 3%, Oracle dropped almost 6% and CoreWeave lost nearly 8%. AMD and Broadcom each slid 4%, Intel fell 5%, and Super Micro Computer dropped close to 5%. Reuters, cited by BasisPilot, reported that the U.S. semiconductor index fell 3.4% that day, with Micron down 4.8%.

    The underlying business didn't shrink

    The selloff reacted to how revenue was presented, not to a collapse in demand. OpenAI's investor update also reported 77% total run-rate growth in its third quarter and 107% run-rate growth in its enterprise business, per CNBC's source. Separately, a person familiar with the matter told reporters that the revised figure reflected the removal of cloud-partner revenue and "wasn't an indication of changing demand."

    That was enough to hit stocks anyway. In 2026, private AI labs sit at the center of a huge capital-spending chain. Their revenue pays for the cloud contracts, which pay for the data centers, which pay for the chips. When investors suddenly think the top of that chain is earning less than they assumed, every company below it gets repriced.

    Record chip sales, falling chip stocks

    What made the day odd was that the hard data on AI hardware looked excellent. TSMC, which manufactures most of the world's advanced AI chips, reported September revenue of about NT$511.86 billion, up 54.6% year over year. Its third-quarter revenue came in at NT$1.49 trillion (about $46.7 billion), up 50% from a year earlier and ahead of analysts' NT$1.46 trillion estimate.

    In earlier cycles, numbers like those would have lifted the whole sector. This time chip stocks fell anyway. BasisPilot put the gap this way: there's a difference between wafers shipping today and what AI customers will eventually earn from them. Strong deliveries show that spending is still happening. They don't show that the spending will pay off.

    Rates are making it worse

    The macro backdrop is adding pressure. According to Invezz, the U.S. 10-year Treasury yield stayed above 5.3% that Thursday, and Brent crude rose above $104 as geopolitical tensions pushed energy prices up. High yields hurt long-duration growth stocks more than most, because more of their value depends on profits far in the future. When those future profits rest on private companies whose figures can change with an accounting choice, investors demand a bigger discount.

    Wall Street analysts haven't turned bearish on AI demand, though. Citi raised its AMD price target to $800 from $575, arguing that agentic AI will need far more continuous compute than chatbots do, which would expand demand for CPUs, memory and networking. Mizuho lifted its Intel target to $114 from $92 while keeping a Neutral rating, saying stronger CPU demand is "lifting all boats."

    Next test: Anthropic's numbers

    The next disclosure is due this week. Anthropic plans to present its third-quarter financials for the first time at a meeting with potential investors on Wednesday, as it prepares for what could be the largest IPO in history. The company has filed confidentially with the SEC but hasn't made its S-1 financials public. Reuters has reported a November listing, though the date hasn't been confirmed.

    One question matters most: does Anthropic count partner revenue the way OpenAI's larger figure did, or the way the $50 billion figure doesn't? Last week showed that the answer can move stock prices by several percentage points. Chip stocks were already falling again on Tuesday, extending an overnight selloff driven by worries about how durable AI financing is.

    Then TSMC reports on October 15

    Right after that, TSMC releases its full third-quarter results on October 15 (its conference call is at 2:00 p.m. Taiwan time). Investors will look past revenue, which is largely known already, to gross margins, advanced-packaging demand, progress on leading-edge nodes and guidance. TSMC has raised its full-year U.S.-dollar growth outlook several times this year: from the mid-to-high 20% range in January, to more than 30% in April, to at least 40% in July. If revenue beats but profitability weakens, the market will read that very differently from a beat with steady margins.

    What it means for business, tech and finance readers

    The main lesson from this week is that AI-sector risk now depends partly on disclosure. Some of the most important numbers in tech come from private companies, are shared selectively with investors, and get reported through anonymous sources. The way those numbers are defined, gross or net and with or without partners, can matter as much as the growth behind them.

    For operators, this probably means AI vendors and cloud providers will face more questions about how stable their pricing and capacity are. For investors, it means looking at what a revenue headline actually includes before reacting to it. And for the AI labs heading toward public markets, it means the move from private fundraising to SEC-standard reporting will get close scrutiny, and accounting definitions alone could change the price of a whole sector.

    AI demand looks solid based on the hardware data. What the market is now trying to figure out is how much that demand is worth and who earns it.

    Sources

    AI stocksOpenAIAnthropic IPOTSMCSemiconductors
    Follow via RSS

    Written by Raj

    Founder & Editor

    I write HorizonDesk: practical, no-hype articles on technology, money, business and the vehicles we drive. Everything here is something I’ve researched, tested or wanted explained clearly myself.

    More from HorizonDesk

    See all in Finance