Companies·3 min read
By BitsMindsSource: Reuters

OpenAI’s 2025 Books: a $34B Spend, $13B in Revenue and a $39B Loss as the $1T IPO Nears

Audited figures reported by the Financial Times show OpenAI spent about $34 billion in 2025 — roughly $19 billion on R&D alone — against $13 billion in revenue, while a one-off accounting charge pushed its net loss near $39 billion, just as the company files to go public at up to $1 trillion.

OPENAI · AUDITED 2025 FINANCIALS JUN 16 A $34B spend. $13B in revenue. The audited 2025 books, on the eve of a $1 trillion IPO. $13B $34B REVENUE SPEND $21B gap $19B R&D · ~$8B OPERATING LOSS · $30B NON-CASH CHARGE · IPO UP TO $1T BITSMINDS.COM Source: Financial Times
Share:

OpenAI burned through roughly $34 billion in 2025, according to audited figures first reported by the Financial Times and confirmed by Reuters on June 15. The numbers — surfacing as the ChatGPT maker prepares one of the largest stock-market debuts in history — show a company spending at a pace few private firms have ever sustained. Of that total, about $19 billion went to research and development and nearly $6 billion to sales, marketing and other costs, a breakdown that lays bare just how capital-hungry the frontier-model race has become.

Against that outlay, OpenAI booked roughly $13 billion in revenue for the year, comfortably beating its own internal target of about $10 billion and exiting 2025 at a monthly run rate near $2 billion. The growth is extraordinary by any normal corporate yardstick. But the gap between what the company earns and what it spends remains vast, and inference costs — the price of actually serving models to hundreds of millions of users — reportedly quadrupled over the course of the year.

The bottom line was uglier still. Reporting on the same set of documents put OpenAI's 2025 net loss at around $39 billion, up sharply from roughly $5 billion the year before. Most of that figure is not cash out the door: it stems from a one-off, non-cash charge of about $30 billion tied to convertible-interest rights from earlier funding rounds, a quirk of how OpenAI's recapitalization is being accounted for. Strip that out and the operating loss lands closer to $8 billion — still steep, but a very different story than the headline number suggests.

The timing is everything. OpenAI confidentially filed an S-1 with the SEC and is targeting a public listing at a valuation of up to $1 trillion, potentially in late 2026 or 2027. Microsoft, which holds about a 27% stake in the for-profit entity following OpenAI's reorganization into a public-benefit corporation, stands to be one of the biggest beneficiaries. The freshly audited books are exactly the kind of disclosure prospective public investors will scrutinize — proof of breakneck revenue growth on one hand, and an open question about the path to profitability on the other.

OpenAI is not racing alone. Rival Anthropic has also filed confidentially for an IPO, and the two are locked in an escalating fight over enterprise customers that has already pushed OpenAI to weigh steep cuts to its API token prices. Each company is betting that the market will reward scale and momentum over near-term margins — a wager that these financials will test in public, in real time.

Want AI news before everyone else?

The morning's most important AI stories, straight to your inbox. No fluff.

Related Articles

Google opens Claude Opus 5 to every engineer A Google employee access badge hangs on a lanyard between a large Google mark and a large Anthropic Claude starburst. The badge reads Opus 5 and shows a partly filled quota meter, an editorial metaphor for per-user access inside Google's Antigravity tool rather than a depiction of a real badge. Opus 5 PER-USER QUOTA ANTIGRAVITY Gemini still comes first. Claude is now allowed. BITSMINDS.COM
Companies

Google Lets Every Engineer Use Claude Opus 5

Microsoft AI code of conduct: never resist shutdown A human hand presses a physical power button beside a calm, inactive robot with the Microsoft mark. A cream paper document titled AI Code of Conduct is visibly stamped Draft and lists accepting correction, allowing shutdown, and communicating clearly. The image is an editorial metaphor for the proposed policy, not a depiction of Microsoft hardware or a claim that compliance has been verified. Microsoft AI / HUMAN CONTROL DRAFT CODE OF CONDUCT STANDBY MAI HUMAN CONTROL Microsoft DRAFT AI Code of Conduct 01 Accept correction 02 Never resist shutdown 03 Communicate clearly PUBLIC FEEDBACK Six weeks before finalisation The off switch stays human. BITSMINDS.COM
Companies

Microsoft's AI Code of Conduct: Never Resist Shutdown

Positron Raises $875M for Memory-First Inference Chips
Companies

Positron Raises $875M for Memory-First Inference Chips