U.S. technology markets lurched lower Thursday, then stabilised Friday, as confusion over OpenAI's actual revenue sent shockwaves through global chipmaking and cloud infrastructure. The Nasdaq fell 1.25% after the Financial Times reported that OpenAI's annualised revenue stood at $50 billion (€44.6bn) at the end of September — roughly $20 billion (€17.8bn) below what the company had projected the previous month. The volatility exposed a harder truth: the artificial intelligence boom, which has justified trillion-dollar valuations and massive capital expenditure across Europe's technology supply chains, rests on accounting uncertainties and investor assumptions that may not survive serious scrutiny.
Nasdaq 100 futures recovered 0.8%, and S&P 500 futures gained 0.4% in early trading Friday, suggesting some stabilisation. But confidence in major technology holdings took a real hit. Oracle, which supplies computing power to OpenAI, fell 5.5%. CoreWeave, a cloud specialist with multibillion-dollar contracts to OpenAI, dropped nearly 8%. Microsoft, Amazon, Alphabet and Meta each declined more than 1%. Nvidia, the dominant chipmaker, fell almost 3%. AMD slid around 4%. Broadcom, which is building a custom chip with OpenAI, fell 4.35%.
The arithmetic matters. It shows how fragile consensus valuations have become. OpenAI's reported $50 billion in annualised revenue counted only direct customer payments to the company itself. The earlier $70 billion (€62.5bn) estimate included sales of OpenAI's models through cloud platforms like Amazon Web Services and Google Cloud — the same accounting method Anthropic, OpenAI's chief rival, uses. Anthropic reportedly reached $65 billion (€58bn) in annualised revenue at the end of July. Because the companies count revenue differently, a direct comparison is nearly impossible. What's clear is that market participants had been working with incomplete or inconsistent information.
The Valuation Question
OpenAI had a valuation of $852 billion (€760bn) in March. The company is now reportedly seeking at least $30 billion (€26.7bn) in new funding at a $1.4 trillion (€1.25tn) valuation — doubling its enterprise value in months. Neither figure rests on audited financial statements. SoftBank Group, which owns about 13% of OpenAI, saw its Tokyo-listed shares fall as much as 7.3% Friday before recovering some losses. Major institutional investors' exposure to unaudited, shifting revenue calculations raises a governance issue European regulators should monitor.
The broader AI hardware sector showed more resilience. Samsung estimated Thursday that its third-quarter operating profit reached 107.4 trillion won, or $80.2 billion (€71.5bn) — a record for any technology company if confirmed, surpassing Nvidia's most recent quarterly operating income of $63.7 billion (€56.8bn). Taiwan's TSMC, the world's largest contract chipmaker, reported that September sales rose 54.6% year on year, with third-quarter revenue reaching a record NT$1.49 trillion, or $46.8 billion (€41.8bn). Those reports suggest demand for AI infrastructure remains robust, even if the financial architecture supporting it is shakier than assumed.
What Comes Next
Investors will get a clearer picture in the coming days. ASML, the Dutch chip-equipment maker that supplies the technology underpinning global semiconductor manufacturing, is due to report on 14 October. TSMC will report on 15 October. Results from both companies will indicate whether the AI infrastructure boom is translating into sustained purchasing patterns or whether Thursday's volatility signals a broader slowdown in capital deployment.
Potential public listings by Anthropic as soon as November and by OpenAI in 2027 could force both companies to publish audited accounts. Until then, European investors and policymakers must face an uncomfortable reality: the AI boom that was supposed to drive European competitiveness and justify massive government subsidies for chip manufacturing rests partly on valuations built from incomplete or methodologically inconsistent financial reporting. The market correction—modest as it was—shows that confidence, once shaken, takes more than optimism to restore.
Why This Matters:
Europe's technology and industrial strategy now depends heavily on the assumption that AI infrastructure investment will drive growth, chip manufacturing capacity, and competitive advantage against China and the United States. OpenAI's revenue confusion matters because it exposes how little standardisation exists in how AI companies report their finances to investors. When a $1.4 trillion valuation rests partly on accounting methods that differ from a rival's by $20 billion—and major institutional investors hold stakes in unaudited companies—the risk of correction increases. European governments have committed substantial public resources to chip fabrication (including ASML's supply chain) and AI infrastructure. If the demand signals supporting those investments come from inconsistent or optimistic accounting, the returns on those bets may not materialise as planned. The revenue confusion isn't just a market event; it's a governance and fiscal accountability issue.