Sam Altman Seeks to Make OpenAI Profitable

The New York Times reports that OpenAI is intensifying efforts to monetize its products after years of rapid expansion, and that executives hope to triple revenue in the coming year (New York Times). The Times also reports an internal expectation to spend about $100 billion more over the next four years to build and deploy models (New York Times). Forbes reports OpenAI recorded $20 billion in revenue for 2025 and says infrastructure agreements with Oracle, AMD, and Broadcom amount to roughly $1.4 trillion of committed compute spend over eight years (Forbes). The Times notes the company has started serving ads inside ChatGPT, reversing an earlier public stance attributed to CEO Sam Altman (New York Times). Search Engine Journal quotes Mr. Altman from the Big Technology Podcast, where he acknowledged that training costs outpace revenue growth and described aggressive training spend as the trade-off for longer-term returns (Search Engine Journal). The original RSS summary reports that Mr. Altman has culled projects and is pursuing more disciplined strategic choices.

Industry-pattern observations: Large foundation-model vendors commonly face a predictable cost curve where training represents lumpy, up-front capital intensity while inference can scale into a steady revenue stream if adoption is broad. The Search Engine Journal transcript of Altman's interview highlights this technical-economic tension: he framed the company's trajectory as front-loaded training investment that inference monetization must eventually "subsume" (Search Engine Journal). Practitioners and infrastructure teams should read these reports as confirmation that compute procurement, cost efficiency of inference, and workload scheduling remain the main levers for AI economics in 2026.

Reporting by the New York Times and Forbes places OpenAI's financial choices in the center of the broader industry debate about sustainability for large-model development. The Forbes account of massive multi-vendor compute commitments underscores how cloud and silicon suppliers are underwriting long-term model capacity even while commercial revenue is still catching up (Forbes). Public-facing moves such as introducing ads into ChatGPT are framed in coverage as pragmatic revenue experiments with user-trust tradeoffs, not as purely technical changes (New York Times).

For practitioners: key signals include quarter-to-quarter revenue growth versus inference volume, margin trends on API and enterprise contracts, reported progress on ad monetization metrics, and any public disclosures about cost-per-inference reductions. Financial actions to monitor in reporting include fundraising rounds or large capital commitments; Forbes reports OpenAI has been in talks to raise as much as $100 billion in new capital (Forbes). Observers should also track product telemetry and developer experience changes that indicate whether adoption is broadening the inference base.