OpenAI spent $34 billion in 2025 as losses mount ahead of planned IPO

OpenAI Financials Expose Massive Gap Between Revenue and Costs

OpenAI financials obtained by independent journalist Ed Zitron and verified by the Financial Times paint a striking picture of a company racing to dominate the AI market while burning through capital at unprecedented rates 1

1

2

2

. The ChatGPT maker generated $13.07 billion in revenue during 2025, a dramatic jump from $3.7 billion in 2024 1

1

. Yet OpenAI spending hit $34 billion last year, creating a chasm between income and expenditure that defines the current economics of frontier AI development 3

3

.

By year's end, monthly revenue was approaching $2 billion, and the company now serves more than 900 million weekly users, though only about 50 million subscribe to paid tiers 1

1

. Despite this growth, the OpenAI cost structure reveals that scale alone does not guarantee efficiency in the AI market.

OpenAI Research and Development Spending Dominates Budget

Audited financial figures show OpenAI research and development spending climbed from $7.81 billion in 2024 to $19.18 billion in 2025—a figure that exceeded the company's total annual revenue 1

1

4

4

. Of that amount, $10.59 billion went directly to Microsoft, likely covering payments to its infrastructure partners and the enormous cost of training new models 1

1

.

AI infrastructure costs extend beyond model training. Running models at scale proves equally expensive, with OpenAI's cost of revenue climbing from $2.65 billion in 2024 to $7.5 billion in 2025 1

1

. Every user interaction with ChatGPT carries a compute cost, and inference costs at current usage levels add up quickly. Sales and marketing expenses also surged from $1.11 billion to $5.73 billion as the company aggressively expanded its user base 1

1

.

Understanding OpenAI Revenue and Losses Through Accounting Adjustments

The headline net loss figure of nearly $39 billion appears especially stark, but requires context 1

1

2

2

. A substantial portion stems from a one-time accounting adjustment tied to changes in investor valuations following OpenAI's transition from a non-profit to a for-profit structure. Under US accounting rules, investors received convertible interest rights that went on the company's ledger as liabilities, creating a $41.55 billion loss related to the conversion 2

2

.

Stripping out these non-cash charges brings the operating loss closer to $8 billion, a figure that better reflects underlying operations 1

1

3

3

. Even this adjusted number represents a significant increase from the $8.78 billion operating loss in 2024 1

1

. The Financial Times cited a person familiar with OpenAI's financial situation who confirmed the $8 billion figure as more representative of actual operational losses 2

2

.

OpenAI Planned IPO Targets Trillion-Dollar Valuation

The leaked documents arrive as OpenAI prepares for its planned IPO, having filed confidentially with the US Securities and Exchange Commission 3

3

. The company targets a valuation of up to $1 trillion, which would rank among the largest public listings in history 3

3

. This comes after OpenAI raised $122 billion in March at a valuation of $852 billion, marking one of the largest funding rounds in the sector 1

1

.

One backer told the Financial Times that underwriting the round meant assuming an eventual public valuation north of $1.2 trillion 3

3

. Investor confidence remains strong despite the losses, backed by SoftBank, Nvidia, and more than two dozen others 3

3

. The company has told investors it hopes to become profitable by 2030, though achieving this goal will depend partly on lowering training and inference costs 1

1

.

Market Pressures and Strategic Shifts Shape Future Outlook

Several factors could pressure ChatGPT revenue growth going forward. Enterprise customers are pushing back against token-based pricing and demanding clearer returns on their AI investments 1

1

. Competition is intensifying as rivals like Anthropic put pressure on pricing, which could squeeze margins if subscription prices decline 1

1

.

These realities are already shaping internal decisions. OpenAI has moved away from several initiatives, including shutting down its Sora video model earlier this year, while leadership signals a tighter focus on core products aimed at developers and business users 1

1

. The $34 billion spent last year represents an early installment on OpenAI's commitment of roughly $600 billion to AI infrastructure through 2030 3

3

. The fundamental question remains whether revenue growth can continue to accelerate fast enough to eventually match the infrastructure required to support it.