What looked like a record number — US$70 billion in annualized revenue — turned into a warning. On October 8, a Financial Times report revealed that OpenAI had informed its investors that its annualized recurring revenue (ARR) was approximately US$50 billion at the end of September, about US$20 billion below market estimates. This discrepancy does not represent lost sales, but reveals something more fundamental: the tension between how fast generative AI grows and how fast it actually generates cash.
The central issue is not whether OpenAI is growing — it clearly is. According to Reuters, the company reported on September 29 that its annualized revenue had risen more than 70% since July, with enterprise sales more than doubling in the same period. Third-quarter consumer revenue exceeded the full-year consumer revenue of 2025. But there is a chasm between multiplying monthly billing by 12 and having recognized, contractual, predictable revenue. The FT report showed internal documents where OpenAI projected US$36 billion in total revenue for 2026 and, for the period 2026–2030, a cumulative cash burn of US$278 billion. This funding gap explains the company's aggressive fundraising: it raised US$122 billion in March at a US$852 billion valuation and now seeks an additional US$30 billion, potentially at a US$1.4 trillion valuation.
Markets reacted as if the entire AI infrastructure ecosystem were at risk. Oracle shares fell more than 5% before recovering 5.3% the following day — Gil Luria, a technology analyst at DA Davidson, noted that OpenAI accounts for roughly half of Oracle's compute backlog. Nvidia lost more than 2%, Broadcom almost 4%, CoreWeave also declined, and the Philadelphia Semiconductor Index plunged 3.39%. The Nasdaq fell 1.25%, the S&P 500 0.47%, while the Dow Jones rose 0.10%, reflecting a more defensive market posture.
What many analysts are underestimating is the context. OpenAI itself reported US$2 billion in annualized revenue in 2023 and reached US$20 billion by the end of 2025 — a tenfold increase in two years. Its compute capacity jumped from 0.2 to 1.9 gigawatts. In 2026, the 70% growth in run rate reflects this massive infrastructure expansion finally being monetized. But the "annualized run rate" metric has a fundamental flaw: it multiplies one month's billing by 12 and assumes that month's pace holds. In September, OpenAI was in a growth phase, but that does not guarantee October, November or December will follow the same trajectory. Reuters explicitly warned that this metric can be misleading and does not equate to recognized annual revenue, contracted recurring revenue, profit, or cash collected.
The competition also weighs heavily. Anthropic, a direct rival, reported a twelvefold increase in 2025 to nearly US$4.6 billion in revenue — but also a net loss of US$42 billion. The pattern in the AI industry is clear: explosive revenue coexists with massive losses, and the ultimate winner will be the one that sustains operations longest before the IPO, not necessarily the one that bills most in the short term. OpenAI itself is in the process of hiring revenue executives — the name Dali Rajic, former Wiz executive, appeared as Denise Dresser's replacement as chief revenue officer, a move signaling commercialization priority over product-only focus.
What makes this case especially interesting is the asymmetry of information. Investors received public estimates of US$70 billion, internal documents point to US$50 billion, and the company itself did not provide a comment. A spokesperson did not respond to requests for comment. The ambiguity is not accidental: OpenAI is trying to sell the narrative of accelerated growth to attract US$1.4 trillion in valuation while preparing for its IPO, but it needs to be honest with investors about the model's sustainability.
OpenAI's IPO may bring answers. When the company lists on Wall Street — and both it and Anthropic would be leading players in the coming weeks — investors will have access to audited financial statements. Until then, every partial report feeds the expectation game that sustains the largest investment bubble in technology history.
Sources: CNBC, Axios, Crypto Briefing
✓ Independent sources cross-checked and verified before publishing
