OpenAI Needs A Trillion Dollars In The Next Four Years

Source: Where’s Your Ed At Date: 2025-09-26 URL: https://www.wheresyoured.at/openai-onetrillion/

Summary

Ed provides the specific math on why OpenAI’s infrastructure promises are impossible: each gigawatt of data center capacity costs ~$32.5B (infrastructure + GPUs); to receive NVIDIA’s promised $100B investment, OpenAI would need to spend ~$325B; OpenAI has made contradictory commitments totaling hundreds of billions across multiple parties, all due within five years. Industry analyst Gil Luria confirmed “there isn’t enough capital for all of this.” Ed characterizes Sam Altman as “a career liar” whose extravagant promises receive uncritical media coverage.

Implications

This is a core AI financial sustainability signal — the specific arithmetic of why the infrastructure buildout is unfeasible. The $32.5B/gigawatt figure is a useful benchmark for evaluating any AI infrastructure announcement. The contradictory commitments pattern (promising the same resources to multiple parties) is a classic fundraising fraud marker that Ed documents here before it becomes mainstream coverage. The “career liar” characterization is strong — watch whether subsequent events (delayed products, missed milestones, revised timelines) corroborate the pattern Ed documents.