H100 SXM
Standard contract
AI compute exchange
AI compute,
made liquid.
Arken Compute transforms fragmented GPU capacity into standardized, tradable contracts for spot capacity, forwards, secondary liquidity, and clearing.
Standard GPU contracts. Public prices. API delivery.
A contract is defined by a small set of objective parameters: Nvidia GPU SKU, region, duration, delivery window, quantity, and SLA.
From idle capacity to a liquid market.
Neoclouds can monetize underused inventory. AI labs can secure capacity without overcommitting to rigid 12-36 month contracts.
Reference market
The Brent of GPU compute.
Public curves for standardized GPU capacity, beginning with the most demanded Nvidia SKUs.
H100 SXM · EU-WEST · M3 $2.82/hr
H200 SXM · US-EAST · M6 $3.18/hr
B200 · US-WEST · M12 $4.74/hr
H200 SXM
Forward contract
B200
Cleared capacity
GB200 NVL72
Cluster contract
Contract primitive
Package capacity into instruments traders can price.
Arken turns a compute quote into a fungible market unit that can be matched, cleared, resold, financed, or used as collateral.
Clearing layer
The exchange prepares. The market decides.
Index
A public reference price for GPU compute, built from standardized contract activity.
Spot + forward
Orderbook matching for immediate capacity and future delivery windows.
Secondary + swaps
Resale, hedging, and debt coverage for GPU-backed infrastructure.
Market data
Meet ACX.
The Arken Compute Index tracks the clearing price of standardized AI infrastructure capacity.
Reference index
ACX
Contract base
SKU + region + duration + SLA
Market phase
Private beta
