Akash Network, the decentralized GPU and cloud-compute marketplace, has formally set its migration to Solana in motion. On October 26, the project’s core team opened governance Proposal #346 on its Cosmos-based akashnet-2 chain, asking token holders to approve a move that would abandon Akash’s own sovereign chain and rebuild the marketplace as native programs on Solana mainnet. Co-founder Greg Osuri laid out the decision in a migration report published alongside the vote. AKT, the network’s native token, rose about 4% over the following 24 hours to roughly $0.68, a positive but hardly euphoric read on a decision that reshapes where one of the largest decentralized compute networks settles its transactions.
The vote opened at 23:02 UTC on October 26 and runs through November 2. The bar to clear is modest relative to the enthusiasm: the proposal needs a 33.4% quorum and a simple majority to pass. If it does, AKT would relaunch as a Solana SPL token at a one-to-one ratio, keeping the same ticker, six decimals, and total supply, with a one-way migration portal open for two years. Nothing about the decision is instant, which is precisely the point of a network that has spent more than a year weighing its options.
Why Akash Is Walking Away From Its Own Chain
The core argument is cost. Akash runs roughly 100 validators that bond AKT purely to secure consensus, and its core team maintains its own forks of cosmos-sdk, CometBFT, and gogoproto, plus an entire CosmWasm subsystem whose only production job is importing Pyth price feeds. On Solana, that maintenance burden disappears because every program inherits security from the validator set with no bonding, no slot auction, and no SOL stake required of Akash itself. “Akash stops paying for a chain it does not use,” Osuri wrote, before putting a number on the alternative: “Network fees for the whole marketplace come to about 0.075 SOL a day, a few dollars.”
That figure is worth unpacking. At Solana’s price near $100.77 at the time of writing, 0.075 SOL works out to roughly $7.60 a day, or about $2,750 a year, to run the entire marketplace’s settlement layer. The comparison is even starker on a per-transaction basis: a median Solana transaction costs about $0.001, and Akash’s roughly 15,000 daily marketplace transactions amount to a rounding error next to Solana’s 112 million daily user transactions. Even at 100 times today’s load, Akash’s activity would not meaningfully dent Solana’s throughput or its fee economics.
The report lists three further reasons. Decentralized physical infrastructure networks, or DePIN, already cluster on Solana: Helium, Render, io.net, Hivemapper, and Nosana all settle there, and Helium’s 2023 retirement of its own chain into another L1 remains the closest precedent for what Akash is attempting. Liquidity and users follow the same logic. “Solana ranks first in weekly DEX volume, carries $15B in stablecoins and sees about 2.2M daily fee payers,” Osuri wrote, while the Cosmos ecosystem that Akash is leaving has been consolidating, with Noble, its stablecoin chain, departing in March 2026.
Token First, Marketplace Second
The migration is deliberately sequenced, and understanding that sequencing matters for anyone holding AKT. The token moves first. Two weeks after the migration proposal passes, and only once two independent audits have cleared the token and portal code, the core team launches AKT as an SPL token on Solana and opens a one-way portal that stays available for two years. Staked AKT has to be unstaked first, which carries a 21-day unbonding period, before it can cross.
The marketplace itself moves later and separately. A second governance vote, the cutover proposal, sets that date only after additional audits and migration rehearsals pass. “nothing changes for live workloads until cutover,” the report states, and tenants keep their existing SDL files and live leases while providers upgrade one daemon and re-register. Once cutover happens, existing workloads keep running for 90 days while tenants redeploy, and only then does the old chain halt. Staking rewards end at cutover, and validator roles and income end when akashnet-2 finally stops.
That split is the honest center of the story: token holders are being asked to approve a directional decision now, while the operational move of the actual compute marketplace remains gated behind audits, rehearsals, and at least two more on-chain votes. The full plan is documented in the project’s public migration program, AEP-79, which specifies both a Solana path and an Ethereum layer-2 path to the same depth.
Solana vs. an Ethereum Layer 2: What Decided It
The choice was not made on branding. Akash ran an open request for proposals that opened in December 2025, and by August 2026 the field had narrowed to two finalists that cleared every hard requirement: Solana mainnet and an existing EVM layer 2. The project’s own cost modeling found that a marketplace operation is about nine times cheaper on Solana on a calm day and roughly 300 times cheaper under load, with a bid costing about $0.0008 on Solana versus $0.007 on an EVM L2, and closing an elevated-load deployment about $0.0014 versus $0.43.
Ethereum’s path won on some axes, most notably its security budget, live slashing, and the deepest engineering and audit market, but lost on the points Akash said it could not compromise: cost under load, the absence of a single commercial sequencer, and a runtime that matches how the marketplace already behaves. Solana’s blocks now hold 100 million compute units, and because each lease lives in its own account, Akash’s throughput scales with the whole block rather than contending for a single sequencer’s ordering.
The reliability question was anticipated rather than dodged. Solana suffered several full outages between 2021 and February 2024, each with a public post-mortem, and the report addresses the concern directly: every candidate had to demonstrate 99.5% liveness over the trailing 24 months, and Solana met that bar. Akash’s escrow settles by elapsed time, so an outage would delay settlement but not lose funds.
What the Market Is Saying
AKT’s response has been constructive but far from a breakout. The token climbed roughly 4% to about $0.68 in the 24 hours after the proposal opened, against a market capitalization near $203 million and a 24-hour trading volume of about $6.7 million, which implies a turnover of roughly 3.3% of circulating supply in a day. For context, AKT remains about 92% below its April 2021 all-time high of $8.07, and it has spent the past month nearly flat, down about 2% over 30 days.
The muted reaction makes sense when the mechanics are read carefully. A migration that keeps the supply, ticker, and decimals unchanged, spreads the token move over a two-year portal, and defers the actual marketplace cutover to a future vote does not deliver an immediate supply shock or an instant liquidity event. What it does deliver is optionality: access to Solana’s $15 billion stablecoin base and its roughly 2.2 million daily fee payers, which is the distribution channel a Cosmos appchain never had.
The precedent set matters more than any single-day price move. ZetaChain’s own vote to migrate ZETA to Solana drew 99.4% support earlier this cycle, and Helium’s migration three years ago is still cited as the proof that a DePIN network can retire its own chain and survive. Akash is now the second sizable Cosmos project in a matter of weeks to conclude that Solana, not an appchain and not an EVM rollup, is the cheaper and better-distributed home for a busy marketplace.
Solana itself has spent the past year hardening the pitch that made this possible. Its Alpenglow upgrade cut finality from 12.8 seconds to 150 milliseconds, and the network’s stablecoin and DePIN cluster has only thickened. For institutional observers, the Akash decision is a data point in a broader re-sorting: infrastructure teams are concluding that running a chain is a liability unless the application genuinely needs custom block processing, which Akash’s marketplace, by its own audit, does not.
What is not yet decided is at least as important as what is. The migration proposal itself must clear quorum and complete its audits, a process the core team has framed as taking roughly two to three weeks for the token and portal alone. The marketplace cutover has no date, staking rewards do not end until that cutover, and validator income continues until the old chain halts. A governance rejection would stop the entire program and send it back to option analysis, which is the one risk the report labels with its sharpest framing. For now, AKT holders have a direction, not a delivery date, and the market’s 4% shrug is, if anything, the appropriate-sized reaction.
