Akash Network runs a decentralized marketplace for compute: providers list capacity, customers deploy workloads, and a blockchain coordinates settlement. Its newest economic change, the Burn-Mint Equilibrium, targets a question every decentralized infrastructure project eventually hits — can a marketplace keep prices stable in dollars while still giving its native token a direct role in every transaction?
The problem BME addresses
If customers pay in USDC, usage can grow while the native token becomes optional to the network’s economics. If customers must pay in a volatile token, enterprises walk away, because a month-long training job should not swing in cost on a liquidation cascade. Akash’s earlier AEP-23 settlement let tenants pay in USDC, which helped adoption but weakened AKT’s connection to real demand.
How the cycle works
Burn-Mint Equilibrium inserts AKT into the payment path without exposing customers to its volatility. A tenant burns AKT to mint ACT, a non-transferable credit pegged to the US dollar, then spends that credit on compute. At settlement, the protocol burns the consumed ACT and mints fresh AKT for the provider at the current oracle rate. Customers see a stable dollar cost, providers receive AKT, and the token becomes a settlement rail rather than a speculative sideshow.
Why net burns are not automatic
The deflationary story is conditional, because both legs of the cycle matter. When a tenant burns AKT to mint ACT, supply contracts. But at settlement the protocol can mint new AKT for the provider. Akash’s own explanation walks through the arithmetic: if AKT trades at $1.14 and a tenant needs $1,000 in credits, roughly 877 AKT is burned to mint 1,000 ACT. If AKT later rises to $1.50, the protocol needs only about 667 AKT to cover the same $1,000 settlement, so roughly 210 AKT is permanently removed. If AKT falls instead, the adjustment runs the other way. Treating every dollar of gross compute spend as a guaranteed buyback or permanent burn misreads the mechanism.
The oracle and vault dependency
BME settles leases against a 30-minute time-weighted average of AKT/USD. Akash’s AEP-81 integration specifies a Pyth-based feed with Wormhole-verified price messages and health checks that reject stale or outlier data. A vault holding burned AKT tracks a collateral ratio, and circuit breakers throttle new minting if that ratio degrades — enhanced monitoring near a 0.95 ratio, with new minting slowed at 0.90 while existing settlements continue. This is deliberate engineering, but it also means settlement quality depends on accurate, timely price data and on the governance managing those parameters.
What to actually watch
Track the net change in circulating supply across a full cycle, quarterly dollar-denominated compute spend, and whether paying workloads return. Akash’s Q1 2026 report cited more than $5 million in cumulative compute spend and a March 23 mainnet activation of BME following Proposal 318 — usage milestones, not proof of future token appreciation. The mechanism links usage to token flows, which changes how AKT derives value, but it does not guarantee demand, higher prices, or any particular target.
Adapted from Akash Network Price Prediction: AKT Targets to 2030.