Benchmarking GPUs: The Players Building a Commodity Market for Compute
AccBalanced · x · 2026-09-05
The original thread maps out what's needed before GPU compute can be traded like a commodity, and who is building it:
- Physical GPU testing: SiliconMark by Silicon Data tracks how a card's performance changes over time
- System-level benchmarks: MLPerf (MLCommons), InferenceX (SemiAnalysis) and Artificial Analysis compare how fast systems run workloads
- Price indexes: rental indexes by Silicon Data and compute price indexes by Ornn Exchange and others measure current GPU rental prices
- Forward curves: e.g. Silicon Data's, projecting future GPU rental prices
- Tradable contracts: CME Group with Silicon Data, and ICE Markets with Ornn Exchange, are turning prices into tradeable contracts
Many more players are entering the space, the author notes.
Related event: Competing GPU compute indexes diverge wildly as power futures emerge(3 posts)→
More from Venture
- FreePi Bets on Ad-Supported Coding Help to Undercut $20/Month AI Coding Subscriptions — heyneighbor · 2026-10-21
- Rumor: Did OpenAI just buy dots.com? — gaganghotra_ · 2026-10-21
- Solo dev's trend-writing agent ran his site for 2 months: 4K impressions, 33 clicks — Status_Blacksmith523 · 2026-10-21
- How to build the wrong product: ask everyone, ship later, heed non-payers — weswinder · 2026-10-21
- Corporate token spending growth collapses from 42% to 0.2% as hyperscaler ROIC falls to 20% — SumitGup · 2026-10-21
- Survey: Big firms burn 28% of software budgets on forced upgrades, costing $12.87M a year — rohanpaul_ai · 2026-10-21