Docs: OpenAI's net loss grew from $5B in 2024 to ~$39B in 2025, a source says due to a $30B non-cash accounting charge linked to its previous structure
Audited figures reveal mounting costs from model development, infrastructure and rapid expansion
Context & Ripple Effects
The audited figures add detail to an arc already visible in OpenAI’s disclosures: revenue had been rising rapidly, while R&D remained a major source of cash use. The new documents put 2025 spending at $34B, including $19B in R&D and nearly $6B in sales and marketing.
They also arrive against reports that OpenAI has made spending pledges exceeding $1T over five years. The reported loss therefore matters less as a standalone operating-cost measure than as evidence of how accounting effects and expansion costs can diverge in assessments of the company’s financial position.
First-order effects
- The roughly $39B reported net loss will draw immediate scrutiny to OpenAI’s audited accounts, but the cited $30B non-cash charge means it should not be read as equivalent to 2025 cash burn or operating spend.
- OpenAI’s reported $34B annual spending, particularly its R&D and sales-and-marketing outlays, makes the scale of investment required to support its growth more concrete for investors and counterparties.
Second-order effects
- Financing discussions and partners tied to OpenAI’s long-term spending plans are likely to focus more closely on the distinction between reported losses, non-cash restructuring effects, revenue growth, and recurring operating costs.
- The figures raise the bar for OpenAI to turn rapidly growing revenue into a funding narrative that can support continued model development, customer acquisition, and infrastructure commitments at the same time.
Third-order effects
- If leading AI developers continue to pair fast revenue growth with exceptionally large R&D and expansion budgets, financial capacity—not just model quality—will become a more important determinant of who can compete at the frontier.
- The episode also points toward more sophisticated scrutiny of AI-company disclosures: headline net losses may be a poor proxy for operating economics when corporate-structure and other non-cash charges are material.
The trend: Frontier AI is becoming a capital-intensity contest in which accounting structure, funding access, and the ability to sustain infrastructure and R&D spending increasingly shape competitive position.