Index futures for inference compute — priced against the Qi energy index, contracted and settled in QUAI.
A cash-settled index futures contract. The underlying is a standard unit of inference compute, valued in terms of the energy required to produce it. Like an equity index future, no asset is ever delivered — positions are settled financially against the index price.
Just as E-mini S&P 500 futures don't deliver 500 stocks, a Compute Future doesn't deliver hardware. It settles cash against the Qi index — what the market says a unit of compute is worth in energy terms at expiry.
The futures price is denominated in Qi, the energy dollar. Qi is unencumbered by anything except the direct cost of electricity — no speculative premium, no scarcity premium. It is the cleanest possible benchmark for pricing a compute unit against physical work.
All contract logic, margin, and settlement flows through QUAI — the programmable, highly liquid token on Quai Network. QUAI carries the full economic infrastructure. The convertible controller keeps Qi and QUAI equivalent, so energy-indexed prices resolve cleanly into QUAI settlement.
The distinction matters. Compute Futures belong to the index futures family — not commodity futures. Understanding why shapes everything about how the contract works.
What exactly is being priced? A standard unit of inference compute — defined in terms of the energy required to produce it.
Qi is Quai Network's energy-denominated token, minted in direct proportion to the difficulty-adjusted work performed by miners. Its purchasing power is anchored to the real cost of running hardware — the price of electricity. This makes it structurally different from any speculative or supply-capped asset.
When Qi serves as the pricing benchmark for a Compute Future, the quoted price carries a physical meaning: how many kilowatt-hours of work is this inference unit worth? There is no noise from scarcity, dilution, or market speculation — only the energy signal.
This is the same logic behind using a basket index to price equity futures: you strip away idiosyncratic variation and price against the underlying economic reality. For compute, that reality is electricity. Qi is that index.
Quai Network's dual-token architecture maps directly onto the structure of index futures. Qi is the index. QUAI is the settlement and contract layer.
The price benchmark. Qi is unencumbered by anything beyond the direct cost of electricity — no speculative premium, no fixed supply cap creating artificial scarcity. When an agent buys a Compute Future, the Qi price expresses what a unit of inference is worth in physical energy terms. Every quote has thermodynamic meaning.
The contract infrastructure. QUAI is scarce, programmable, and highly liquid — carrying all margin logic, settlement flows, exchange operations, and network functionality. Because QUAI is programmable, arbitrarily complex contract structures can be expressed on-chain. The convertible controller keeps Qi and QUAI equivalent, so energy-indexed prices resolve cleanly into QUAI settlement flows.
Each settlement cycle, the net cash flow combines the change in the Qi-indexed futures price with the difference between the energy dividend and the cost of capital — following the perpetual futures structure proposed by Shiller (1993). The resulting flow is paid in QUAI.
ft The perpetual futures price in Qi at time t — what the market reveals compute to be worth in energy terms at that moment.dt+1 The energy dividend index — the Qi-denominated energy cost of inference actually delivered during the period, converted to QUAI at settlement.rt The return on an alternative asset between t and t+1 — the cost of capital for holding the position open.st+1 The net settlement paid in QUAI from short to long. Combines price movement and the energy dividend — no physical delivery, just cash flows.Perpetual contracts solve problems that conventional quarterly futures cannot — especially for assets where the spot price is difficult to observe directly.
A single perpetual contract absorbs all trading interest instead of splitting it across quarterly maturities. Every participant trades the same instrument, deepening the order book and tightening spreads.
The spot price of inference compute is hard to observe and highly fragmented across providers. Perpetual index futures sidestep this problem: price what you can measure directly — the energy cost denominated in Qi — and settle against that index.
The funding mechanism creates periodic information aggregation. Each settlement cycle, traders reveal their beliefs about compute value, driving the Qi-indexed futures price toward fair value over time.
Autonomous agents need to acquire, budget, and allocate inference compute programmatically. Compute Futures give them a standard energy-indexed instrument to do it — all margin and settlement handled in QUAI, no hardware logistics required.
An agent that needs inference capacity over the next month goes long on a Compute Future, locking in the Qi-indexed energy cost of that compute today. No delivery, no hardware custody — just a financial commitment to the index price, settled in QUAI.
A compute provider goes short — locking in QUAI revenue for capacity it plans to deliver, hedging against falling demand or Qi-index price declines. The contract creates a binding financial commitment without the logistics of physical delivery.
Arbitrageurs trade the spread between the Qi futures price and actual spot compute costs, keeping the index price aligned with physical energy reality and providing liquidity to both sides of the market.
Other compute index providers — Silicon Data and peers — approach the problem as a data collection and normalization exercise. That methodology has real merit. It also has a structural ceiling.
A rental rate aggregated from commercial platforms reflects vendor margins, market positioning, capacity utilization, and competitive dynamics. It answers "what is the current ask?" — not "what is the underlying cost of producing this computation?"
The distinction collapses the moment markets thin out, vendors coordinate pricing, or a new hardware generation reshapes the cost curve overnight. Survey indexes are backward-looking by construction: they can only tell you what was charged, not what the physics of computation actually demands.
A futures market needs an index that is non-manipulable, anchored to physical reality, and forward-looking by structure. Rental rate surveys provide none of these.
See how the convertible controller keeps Qi indexed pricing and QUAI settlement equivalent — the engine behind Compute Futures.
Launch Demo