OpenAI's annualized revenue surpasses $20 billion, but The New York Times remains pessimistic: bankruptcy possible by 2027

OpenAI recently delivered an impressive report card, with annualized revenue (ARR) surpassing $20 billion, growing nearly tenfold in just two years. ChatGPT has also transformed from a simple chatbot into a widely embedded digital infrastructure for work and daily life. However, The New York Times worries that OpenAI’s spending rate far exceeds revenue growth, and it may face the risk of running out of funds before 2027.

OpenAI Revenue Surges, ChatGPT Moves Toward Comprehensive Business Tool

OpenAI Chief Financial Officer Sarah Friar recently revealed that the company’s annualized revenue has exceeded $20 billion, far above the approximately $2 billion in 2023. She pointed out that this growth stems from a shift in ChatGPT’s usage scenarios, evolving from an early consumer AI tool to a daily work platform across education, writing, programming, marketing, and finance.

As AI penetrates organizational and team operations, OpenAI’s business model is also evolving, expanding from individual subscriptions to enterprise solutions, and introducing API usage-based billing for developers, making company revenue proportional to the work completed with the tools.

(OpenAI launches ChatGPT Health, can the chatbot serve as a personal health assistant?)

Revenue and Computing Power Scale Simultaneously

On the other hand, Friar also stated that OpenAI’s computing capacity has grown about threefold over the past few years, reaching approximately 1.9 GW in 2025, compared to only about 0.2 GW in 2023, with expansion speed nearly synchronized with revenue growth.

She emphasized that revenue essentially reflects “how much real-world work can be served,” which depends entirely on available computing power. As a result, the company has gradually shifted from a single cloud provider to multiple partners. Earlier this year, OpenAI signed a partnership worth up to $10 billion with chip startup Cerebras Systems, focusing on inference infrastructure.

NYT Cold Water: AI is Not SaaS, Burning Money Might First Destroy Itself

However, some are skeptical of its growth model. The New York Times last week cited multiple studies indicating that even in the most optimistic scenarios, the generative AI industry faces significant financial pressure. The report estimates that OpenAI could burn about $8 billion in cash by 2025, with cumulative losses possibly reaching $40 billion by 2028.

Senior Fellow Sebastian Mallaby of the Council on Foreign Relations predicts that OpenAI will run out of funds within the next 18 months: “Generative AI is fundamentally different from the traditional SaaS model. Each additional user requires more computing power and energy costs, unlike traditional software that benefits from near-zero marginal costs at scale.”

The End of Artificial General Intelligence: The ‘A’ in AGI as Ads’ ‘A’

Another concern is user structure. Currently, most users remain on free versions. Once the platform introduces ads, limits usage, or raises charges, future user retention remains uncertain. In an environment where AI tools are increasingly diverse, switching costs are actually much lower than many imagine.

Many media outlets also jokingly say that OpenAI has finally achieved AGI (Ads-Generated Income, the income generated from advertising).

Nevertheless, Mallaby also points out that if AI further evolves to include agent functions (AI agents), with deep understanding of personal preferences, decision-making logic, and emotional context, platform stickiness will greatly increase, and the business model may truly stabilize.

High Growth and High Financial Pressure: What’s Next for OpenAI?

With CEO Sam Altman’s strong fundraising ability, OpenAI has successfully raised about $40 billion, setting a record for private company fundraising. However, Mallaby still warns that, unlike tech giants with stable profit models, OpenAI is still in a high-investment, low-certainty stage.

Today, the strong growth in annualized revenue only proves market demand exists, but the costs of computing power, capital consumption, and unproven long-term business models also force OpenAI to balance growth speed with financial pressure.

This article “OpenAI’s annualized revenue surpasses $20 billion, NYT remains skeptical: may go bankrupt by 2027” first appeared on Chain News ABMedia.