OpenAI has dramatically increased its spending forecast, projecting a cumulative cash burn of $665 billion by 2030 to train and operate its AI models. This figure represents an increase of roughly $111 billion over previous internal estimates, signaling that the immense computational costs of the AI race are accelerating faster than even the industry's leaders anticipated. The company, according to a report by The Information citing internal financial documents, does not expect to become cash-flow positive until 2030, a timeline that lags behind key rival Anthropic, which is targeting breakeven as early as 2028.

The revised projections show a steep climb in annual cash outflow, with an expected $25 billion burn in 2026 and $57 billion in 2027—a combined $30 billion more than prior forecasts for those two years. This follows a previous major upward revision last fall, when OpenAI raised its expected cumulative cash burn through 2029 by roughly $80 to $115 billion. The primary driver of this cost explosion is inference, the day-to-day expense of running AI models for users. In 2025, these inference costs quadrupled as demand for ChatGPT and API services exceeded expectations, forcing the company to secure more expensive computing capacity on short notice.

This surge in operational costs is squeezing profitability. OpenAI's adjusted gross margin dropped to 33 percent in 2025, down from 40 percent the year before and significantly below its internal target of 46 percent. Looking ahead, the company now expects margins to land between 52 and 67 percent in the future, missing its previous ambitious goal of reaching 70 percent by 2029. Training expenses alone are projected to reach nearly $440 billion by the end of the decade, underscoring the staggering capital required to develop next-generation models.

Despite the soaring costs, OpenAI's revenue is also growing rapidly, just not fast enough to keep pace with expenditures. The company more than tripled its revenue to $13.1 billion in 2025, beating its own forecast by $100 million. It now projects $30 billion in revenue for 2026 and around $62 billion for 2027, with overall revenue projections through 2030 roughly 27 percent higher than previous estimates. The consumer subscription business remains the largest revenue driver, expected to generate around $150 billion by 2030. OpenAI reports 910 million weekly active users, a new high, though this fell short of the one billion target it wanted to hit by the end of 2025. Internal memos from CEO Sam Altman indicate ChatGPT's growth has resumed at a rate of 10 percent month-over-month, and the company is targeting 2.75 billion weekly active users by 2030.

On the enterprise side, OpenAI aims for a massive expansion, pushing revenue from ChatGPT Enterprise and other B2B offerings to $70 billion by 2030, up from just $2 billion in 2025. Revenue from its API segment is projected to hit $47.5 billion. In contrast, hardware and other new products are projected to contribute a more modest $15 billion by 2030. The company expects its first revenue from hardware and "new products" in 2026, starting at $100 million and growing to $1.5 billion the following year, with the first device potentially shipping in spring 2027. Products in development reportedly include a smart speaker.

The financial trajectory outlined in these documents paints a picture of a company in a high-stakes, capital-intensive sprint. OpenAI is betting that massive upfront investment in model training and inference infrastructure will secure a dominant, profitable market position by the end of the decade. However, the repeated upward revisions to cost forecasts and the delayed path to profitability highlight the extreme financial volatility and uncertainty inherent in scaling frontier AI. The gap between OpenAI's timeline and Anthropic's earlier breakeven target also suggests intensifying competition not just on model capability, but on financial discipline and operational efficiency. For the broader AI industry, these figures serve as a stark benchmark for the scale of investment required to compete at the highest level, potentially reshaping the landscape for startups, investors, and cloud providers.