The Programmable Data Center — Evan Conrad, SF Compute

SF Compute is building the operational plumbing—physical settlement, not cash-settled indexes—needed to make GPU compute a genuinely tradeable commodity.

• Physical settlement is the moat: SF Compute spent three years building full cloud operating capacity (racks, backplane, PDUs, switches, DPUs, UFM, BMC) because GPU clusters aren't fungible like cash-settled indexes assume.
• Liquidity is thin but concentrated: Only ~5% of SF Compute's total capacity sits on its order book, yet customers who spend $100 buying capacity recoup ~$25 reselling it—evidence of real, usable liquidity.

Themes, sentiment and reasons are AI-written paraphrases, checked so they never quote an episode. They may contain errors. Not investment advice. Episodes remain the property of their creators.