CME Group and Silicon Data announced plans to list futures on GPU rental indexes, starting with H100 and B200. The stated launch date is October 5, 2026.

No. The launch remains subject to regulatory review. Nothing on GPUQuant asserts that the contracts are trading, and GPUQuant carries no exchange feed that would show it.

What one contract represents

Approximately one month of renting one GPU: the same 730-hour unit GPUQuant quotes as a GPU-month. The contracts track an hourly rental index rather than the price of hardware.

What is actually settled

These are cash-market benchmark derivatives. Nobody takes delivery of a GPU. They settle against an index of rental prices, which is why the index definition matters more than any single provider quote.

A buyer of compute wanting to fix a cost months ahead, a supplier wanting to lock in revenue, and traders taking a view on whether AI compute gets cheaper or scarcer.

None in the market itself. GPUQuant publishes the cash-market context: what the three largest cloud providers publicly charge for the same GPU, month by month and region by region.

NVIDIA GPU compute futures

Renting an H100 for an hour has become a standardized unit, bought and sold at a price that moves. That is the precondition for a futures market, and one is being built: CME Group and Silicon Data plan to list H100 and B200 rental-index futures on October 5, 2026, subject to regulatory review.

This page explains what those contracts are, what they are not, and how GPUQuant's published benchmark relates to them.

Read the CME Group announcement ↗

Planned launch

Pending regulatory review

—days

CME Group and Silicon Data plan to launch H100 and B200 rental-index futures on October 5, 2026. The launch remains subject to regulatory review and the contracts are not trading today.

Cloud list-price benchmark

What compute costs today

The price the cloud providers currently publish to rent one GPU for one hour, on demand, with no commitment. It is observable, it is verifiable against the provider, and it is what GPUQuant publishes.

It is a list price, not a transacted price. Large buyers negotiate below it and committed-use discounts sit well below it.

Futures price

What the market expects compute to cost

The price agreed today for a contract month in the future. It embeds expectations about supply, demand and the pace of new hardware, and it can sit far above or below today's cash price.

GPUQuant does not publish a forward curve. When futures data becomes available through an approved source, it would appear as a separate, clearly labelled series. Nothing on this site estimates one.

Contract context

The two GPUs with planned contracts

GPUQuant's reference price is independently calculated from AWS, Azure and Google Cloud list prices. It is not the CME settlement index and not a Silicon Data figure.

H100

80GB HBM3 · Hopper

Planned

Current GPUQuant cloud list-price reference: $13.19per GPU-hour
730-hour rental equivalent: $9,626per GPU-month
Planned CME-linked product: Silicon Data H100 Rental Index Futures

Median across 3 providers and 79 regions, Aug 2026. Regional range $6.88 to $23.33.

H100 price history →

B200

180GB HBM3e · Blackwell

Planned

Current GPUQuant cloud list-price reference: $15.18per GPU-hour
730-hour rental equivalent: $11,078per GPU-month
Planned CME-linked product: Silicon Data B200 Rental Index Futures

Median across 2 providers and 9 regions, Aug 2026. Regional range $14.24 to $20.08.

B200 price history →

H200 and A100 80GB are tracked by GPUQuant but have no announced CME contract. Nothing on this site should be read as suggesting otherwise.

The short version

What was announced: CME Group and Silicon Data announced plans to list futures on GPU rental indexes, starting with H100 and B200. The stated launch date is October 5, 2026.
Is it live: No. The launch remains subject to regulatory review. Nothing on GPUQuant asserts that the contracts are trading, and GPUQuant carries no exchange feed that would show it.
What one contract represents: Approximately one month of renting one GPU: the same 730-hour unit GPUQuant quotes as a GPU-month. The contracts track an hourly rental index rather than the price of hardware.
What is actually settled: These are cash-market benchmark derivatives. Nobody takes delivery of a GPU. They settle against an index of rental prices, which is why the index definition matters more than any single provider quote.
Who would use them: A buyer of compute wanting to fix a cost months ahead, a supplier wanting to lock in revenue, and traders taking a view on whether AI compute gets cheaper or scarcer.
GPUQuant's role: None in the market itself. GPUQuant publishes the cash-market context: what the three largest cloud providers publicly charge for the same GPU, month by month and region by region.

Why a rental index and not a hardware price

You cannot deliver an H100 into a futures contract in any useful way. Chips are allocated, not traded on an open spot market, and a physical unit would be unsettleable.

What is tradeable is access: an hour of a GPU, which thousands of providers sell at published prices in a form that is close to identical. Index that hour and you have a settleable reference, which is exactly the structure used for electricity, freight and bandwidth before it.

The GPU-month exists for the same reason a barrel or a megawatt-hour does: a contract needs a unit large enough to be worth trading and standard enough that two parties mean the same thing by it. 730 hours is one month of continuous access to one GPU.

What GPUQuant is not

Not a trading venue. GPUQuant does not provide, facilitate or route access to futures trading of any kind.
Not the CME settlement index. The GPUQuant Reference Price is calculated independently from published hyperscaler list prices using the method on the methodology page.
Not a Silicon Data product and not affiliated with CME Group.
Not investment advice, and not a record of completed transactions. List prices are what providers publish, not what every customer pays.

Track the cash market from here

A concise monthly update when official GPU rental prices change, which is the number any compute contract ultimately references.