OpenAI loses $21 billion in 2025 while revenue climbs to $13 billion

OpenAI’s audited 2025 financials reveal a $20.92 billion loss from operations on $13.07 billion of revenue, underscoring a spending surge that dwarfs its growth as the company prepares for an IPO【2】.

Revenue growth vs. expense explosion

OpenAI’s top line jumped 253 % year‑over‑year, from $3.7 billion in 2024 to $13.07 billion in 2025, and monthly revenue was nearing $2 billion by year‑end【2】. Yet the cost side accelerated faster: research and development rose from $7.81 billion to $19.18 billion, while sales and marketing surged from $1.11 billion to $5.73 billion. The “cost of revenue” line—largely compute spend for inference—climbed from $2.65 billion to $7.5 billion, reflecting the heavy compute bill of serving more prompts【2】.

Accounting quirks and cash flow implications

The headline net loss of $38.53 billion includes a $41.55 billion non‑cash charge tied to the fair‑value adjustment of convertible interests and warrant liabilities, a one‑time accounting item that inflates the loss figure【4】. Stripping that out leaves the operating loss of $20.92 billion as the core metric investors will scrutinize. The operating loss, while still massive, improved on a percentage basis—from 237 % of revenue in 2024 to 160 % in 2025—suggesting marginal efficiency gains as scale increases【2】.

Competitive and regulatory backdrop

OpenAI’s spending spree comes as rivals such as Anthropic are edging toward profitability, with Anthropic reportedly close to an operating profit in the current quarter【1】. At the same time, a 42‑state coalition has subpoenaed OpenAI over chatbot safety concerns, adding regulatory risk to the IPO narrative【3】. The company also secured $867 million from SoftBank and $303 million from Microsoft in 2025, and paid Microsoft over $10.5 billion for R&D, likely for large‑model training【1】.

What to watch

• IPO filing timeline – OpenAI has confidentially submitted a draft S‑1; any SEC comments on the disclosed losses could affect pricing.
• Regulatory developments – The ongoing multi‑state investigation into ChatGPT’s behavior may force product changes that impact cost structure.
• R&D spend trends – Future quarters will reveal whether OpenAI can curb its compute burn as model efficiency improves.

The leaked numbers paint a picture of a company racing ahead on revenue while its expense base expands even faster, raising questions about the sustainability of its growth model and the appetite of public investors for a business that still burns tens of billions each year.