OpenAI expects to spend $50 billion on computing power in 2026 to support its artificial-intelligence software, according to co-founder and President Greg Brockman. Brockman disclosed the figure during his testimony in OpenAI’s lawsuit with Elon Musk — the largest single-year compute commitment OpenAI has formally acknowledged.
The Headline Number
- $50 billion in compute spend for 2026
- Disclosed under oath in the Musk litigation
- Largest single-year compute commit OpenAI has formally acknowledged
- Anchors Microsoft’s and now AWS’s data-centre capacity planning
Why The Disclosure Matters
- Brockman’s testimony establishes a formal record of OpenAI’s compute trajectory
- Provides a concrete number for Big Tech’s $700B aggregate AI capex figure
- Confirms OpenAI’s compute appetite is structural, not promotional
- Litigation context limits OpenAI’s ability to walk it back later
The Personal Stake Detail
Separately, Brockman disclosed his personal OpenAI stake is now worth approximately $30 billion. He had originally wanted “a billion.” The figure puts him in the top tier of single-individual AI wealth concentrations.
The Capex Math
- OpenAI: ~$50B compute
- Microsoft: ~$190B all-in capex (2026)
- Alphabet: $180-190B (2026)
- Amazon: ~$200B (2026)
- Meta: $90B+ (2026)
- Combined Big Tech AI capex: ~$700B
The AWS Bedrock Connection
Today’s separate Bedrock announcement — OpenAI launching on AWS — gives OpenAI a multi-cloud distribution path for that $50B compute. Splitting between Azure and AWS reduces vendor concentration risk and broadens enterprise customer access.
What Comes Next
- The Musk litigation continues; expect more disclosures
- Q3 + Q4 OpenAI revenue numbers vs. that compute spend
- Watch for: training-vs-inference compute split disclosure
- Capex curve durability under any economic slowdown scenario
Follow Vibes Uncut Media for continuing AI capex coverage.
Leave a Reply