OpenAI Projects $278B Cash Burn Through 2030 at Possible $1.2T Valuation

The eye-watering figure — $278 billion in projected cash burn from 2026 through 2030 — quantifies the scale of OpenAI's compute-and-talent buildout against a possible $1.2 trillion valuation. Reported by the FT, the number reframes OpenAI as one of the most capital-intensive companies in history, with the burn driven overwhelmingly by compute commitments to train and serve frontier models like GPT-6 Astra.

The mechanics are the crux: this is not operating loss in the traditional sense but a deliberate bet that securing compute now — via multi-year datacenter and chip deals — establishes an insurmountable lead. It directly connects to NVIDIA's Huang forecasting chip sales doubling in 2027; OpenAI is a primary source of that demand, and the two numbers are two sides of the same capex story.

Competitively and politically, the disclosure lands as Altman prepares to brief the UN Security Council on September 25 — an extraordinary elevation of a private AI CEO to a body normally reserved for heads of state, reflecting how central AI has become to security discourse. It also sharpens the question critics raise about the entire AI buildout: who funds $278B, and what returns justify it.

Skeptics see bubble risk — a valuation and burn rate predicated on continued exponential demand that could stall. Bulls counter that revenue and enterprise adoption are climbing fast enough to service the burn. Watch OpenAI's next funding round terms, whether the $1.2T valuation materializes, and what Altman actually proposes to the Security Council.