AI token futures trading set to launch soon

The next major market could revolve around LLM tokens, with financial institutions racing to build supporting infrastructure.
According to Reuters, the Shanghai Futures Exchange is developing a derivatives market for AI tokens. This follows separate announcements from CME Group and the Intercontinental Exchange, owner of the NYSE, that they are working on futures contracts for GPU rentals.
While the GPU market is still maturing, the wide range of companies using, selling, and renting GPUs has created a robust spot market for GPU rental, typically billed by the hour. AI Mining Co., which tracks daily GPU rental pricing across 28 marketplaces and cloud providers, reports median prices for Nvidia H100 GPUs ranging from $1.40 to $4.27 per hour across 13 marketplaces, while H200 GPUs average between $2.34 and $5 per hour across 10 marketplaces. Over the past seven days alone, average H100 prices have fluctuated between $2.79 and $3.33.
Though GPU markets are well established, token infrastructure remains underdeveloped. Tokens are the fundamental units of modern AI models, and major AI companies often price enterprise plans based on token usage. For instance, OpenAI charges $5 per million input tokens and $30 per million output tokens for its latest GPT‑5.5 API. Cloud providers like Amazon’s Bedrock are also adopting per‑token pricing.
This development coincides with an unprecedented expansion of AI infrastructure. Cloud service providers, private equity firms, and infrastructure companies have invested hundreds of billions in data centers, anticipating continued growth in demand for GPUs and compute. A new wave of global neocloud companies is also competing for a share of this demand, with some specializing in inference and others challenging giants like Oracle, AWS, and Google Cloud for AI company contracts.
By targeting AI tokens, the Shanghai exchange’s derivative product would be tied directly to how AI companies price their services, offering businesses, investors, and data center operators a way to hedge against compute costs.
Related article
Clawdmeter transforms Claude Code usage data into a compact desktop dashboard
The tokenmaxxing trend in Silicon Valley now has its own physical device. A new open source project converts your Claude Code usage statistics into a compact desktop dashboard, enabling AI power users to monitor their consumption at a glance.Of cours
Reid Hoffman Addresses the 'Tokenmaxxing' Debate
Days after Meta shut down its internal "tokenmaxxing" dashboard following news of the AI leaderboard leaking to the press, LinkedIn co-founder and venture capitalist Reid Hoffman voiced support for the concept that has recently taken Silicon Valley b
DeepMind CEO Hassabis: I sleep six hours a day, usually feel energetic around 1 a.m.
Fortune recently featured an interview with Demis Hassabis, CEO of Google DeepMind, revealing his unconventional approach to rest and productivity. Hassabis disclosed that he sleeps very little, structuring his waking hours into two distinct work blo
Related Special Topic Recommendations
Comments (0)
0/500

The next major market could revolve around LLM tokens, with financial institutions racing to build supporting infrastructure.
According to Reuters, the Shanghai Futures Exchange is developing a derivatives market for AI tokens. This follows separate announcements from CME Group and the Intercontinental Exchange, owner of the NYSE, that they are working on futures contracts for GPU rentals.
While the GPU market is still maturing, the wide range of companies using, selling, and renting GPUs has created a robust spot market for GPU rental, typically billed by the hour. AI Mining Co., which tracks daily GPU rental pricing across 28 marketplaces and cloud providers, reports median prices for Nvidia H100 GPUs ranging from $1.40 to $4.27 per hour across 13 marketplaces, while H200 GPUs average between $2.34 and $5 per hour across 10 marketplaces. Over the past seven days alone, average H100 prices have fluctuated between $2.79 and $3.33.
Though GPU markets are well established, token infrastructure remains underdeveloped. Tokens are the fundamental units of modern AI models, and major AI companies often price enterprise plans based on token usage. For instance, OpenAI charges $5 per million input tokens and $30 per million output tokens for its latest GPT‑5.5 API. Cloud providers like Amazon’s Bedrock are also adopting per‑token pricing.
This development coincides with an unprecedented expansion of AI infrastructure. Cloud service providers, private equity firms, and infrastructure companies have invested hundreds of billions in data centers, anticipating continued growth in demand for GPUs and compute. A new wave of global neocloud companies is also competing for a share of this demand, with some specializing in inference and others challenging giants like Oracle, AWS, and Google Cloud for AI company contracts.
By targeting AI tokens, the Shanghai exchange’s derivative product would be tied directly to how AI companies price their services, offering businesses, investors, and data center operators a way to hedge against compute costs.
Clawdmeter transforms Claude Code usage data into a compact desktop dashboard
The tokenmaxxing trend in Silicon Valley now has its own physical device. A new open source project converts your Claude Code usage statistics into a compact desktop dashboard, enabling AI power users to monitor their consumption at a glance.Of cours
Reid Hoffman Addresses the 'Tokenmaxxing' Debate
Days after Meta shut down its internal "tokenmaxxing" dashboard following news of the AI leaderboard leaking to the press, LinkedIn co-founder and venture capitalist Reid Hoffman voiced support for the concept that has recently taken Silicon Valley b
DeepMind CEO Hassabis: I sleep six hours a day, usually feel energetic around 1 a.m.
Fortune recently featured an interview with Demis Hassabis, CEO of Google DeepMind, revealing his unconventional approach to rest and productivity. Hassabis disclosed that he sleeps very little, structuring his waking hours into two distinct work blo





Home






