OpenAI $122B Funding Exhausted by 2028

This financing plan means OpenAI is still buying time, not self funding escape velocity. The March 31, 2026 round brought in $122B at an $852B post money valuation, but internal projections cited in the page imply that cash is consumed within roughly two years because model training, inference, and reserved infrastructure capacity are absorbing capital faster than current operating cash flow can replenish it. OpenAI was at an estimated $25B annualized revenue in February 2026 and about $50B by September 2026, yet the business still needs outside money to keep expanding compute.1 2 3 4

• The burn is driven by physical infrastructure, not normal SaaS spend. OpenAI has been reallocating GPUs away from side bets like Sora, which was estimated to burn $15M per day, and toward coding and enterprise products where each extra user still pulls real inference cost through the system.2
• The strategic trade is scale now, financing later. By April 2026, OpenAI had lined up major compute supply through Amazon, Oracle, Nvidia, Google, and Cerebras, which helps it meet demand spikes, but it also turns future growth into large fixed commitments that must be monetized fast enough to avoid another capital crunch.3
• The closest comp is Anthropic, and the contrast shows the market logic. OpenAI raised $122B in 2026 versus Anthropic at $95B, while OpenAI was estimated at $50B annualized revenue by September 2026 and Anthropic at $65B by July 2026. Frontier labs are being financed like infrastructure companies because revenue growth alone is not covering compute expansion.1 3 5

The next phase is about turning reserved compute into higher quality revenue, especially coding, enterprise workflows, and distribution through cloud and channel partners. If OpenAI can keep shifting usage toward products with stronger expansion and repeat spend, the funding story moves from emergency capital to planned capacity buildout, and the path to durable cash generation gets much shorter.2 3 4