Global AI Compute Markets
Discover available AI compute capacity across regions, GPU architectures and contract durations.
| Product | Region | Term | Bid | Ask | Last executed | Available capacity | Delivery date | SLA | Verification | Status | Updated |
|---|---|---|---|---|---|---|---|---|---|---|---|
| Loading the market book… | |||||||||||
These are listed instruments — the configurations MACOMX quotes. A bid, ask or capacity figure appears against one only when a buyer or provider submits it. MACOMX does not generate indicative prices and does not publish a price it has not received.
Short-term capacity
For fragmented and underutilised GPU capacity, typically 1H to 30D. The purpose is to put idle infrastructure to work at short notice.
- Typical terms: 1H · 6H · 24H · 7D · 30D
- Fast provisioning against confirmed capacity
- Verified GPU-hour reporting on delivery
Medium and long-term procurement
For capacity reservation and compute offtake, typically 30D to 3Y. Used by AI companies, model developers, enterprises, telecom operators and cloud platforms.
- Typical terms: 30D · 90D · 180D · 1Y · 3Y
- RFQ → quote → reservation → contract → delivery
- Human verification at every step in V0.1
Index products
Reserved. MACOMX will publish an index only once it rests on executed transactions — no index value will be estimated or modelled before then.
Nothing here is priced yet. That is the point of V0.1.
MACOMX publishes a price when a real provider or a real buyer submits one. If you hold capacity or need it, you are what turns this book into a market.