Startups & Funding

Shanghai Exchange plans futures for AI tokens as GPU derivatives emerge

CME, ICE, and Shanghai are racing to create token and GPU futures markets.

Deep Dive

China’s Shanghai Futures Exchange is developing a derivatives market for AI tokens, according to Reuters. This follows separate announcements from CME Group and Intercontinental Exchange (owner of the NYSE) that they are working on futures contracts for renting GPUs. The GPU spot market is already robust: AI Mining Co. reports median hourly prices for Nvidia H100 GPUs range from $1.40 to $4.27 across 13 marketplaces, while H200s average $2.34–$5.00 per hour across 10. Over the past week, average H100 prices stayed between $2.79 and $3.33. These futures would allow companies to lock in compute costs amid surging demand.

Token-based pricing is also becoming standard. OpenAI charges $5 per million input tokens and $30 per million output tokens for its latest GPT-5.5 API. Cloud providers like Amazon Bedrock offer per-token billing. By targeting AI tokens rather than raw GPU compute, the Shanghai exchange’s product would be directly tied to how AI companies price their services. This gives businesses, investors, and data center operators a way to hedge against rising compute costs as hundreds of billions pour into AI infrastructure from cloud providers, private equity, and neocloud companies.

Key Points
  • Shanghai Futures Exchange is designing a derivatives market for AI tokens; CME and ICE are working on GPU rental futures.
  • GPU spot prices: H100 median $1.40–$4.27/hr across 13 marketplaces; H200 avg $2.34–$5.00/hr across 10.
  • OpenAI charges $5/1M input tokens and $30/1M output tokens for GPT-5.5; Amazon Bedrock also offers per-token billing.

Why It Matters

Token and GPU futures let businesses hedge compute costs as AI infrastructure spending explodes.