Akash Network's Governance Proposal 346, opened October 26 with 98.42% approval and over 50% turnout, will migrate approximately 300 million AKT tokens from the Cosmos SDK to Solana as SPL tokens at a 1:1 ratio. The vote closes November 2. If ratified and audited, the migration portal opens withi...
"Akash stops paying for a chain it does not use. Network fees for the whole marketplace come to about 0.075 SOL a day, a few dollars." — Greg Osuri, Co-Founder, Akash Network
Akash Network's Governance Proposal 346, opened October 26 with 98.42% approval and over 50% turnout, will migrate approximately 300 million AKT tokens from the Cosmos SDK to Solana as SPL tokens at a 1:1 ratio. The vote closes November 2. If ratified and audited, the migration portal opens within 2-3 weeks — irreversible, with a two-year conversion window.
Akash is the third major protocol to abandon Cosmos in 2026 alone. Noble, the ecosystem's primary stablecoin issuance chain, departed in March for an EVM-based Layer 1. Sei Network completed its SIP-3 migration in April, stripping out Cosmos SDK, CosmWasm, and IBC functionality entirely to become EVM-only. ATOM trades at $1.64 as of October 27, down 95% from its 2021 all-time high and 63% year-over-year from $6.60 in January 2025.
The pattern is consistent: sovereign app-chains built on Cosmos SDK are concluding that maintaining independent validator sets and consensus infrastructure generates cost without proportional economic value. They are migrating to general-purpose Layer 1s — Solana, EVM chains — where liquidity, users, and ecosystem density already exist.
The migration proceeds in phases. The first step — a token-only migration — does not move Akash's compute marketplace. That requires a separate governance vote.
Phase 1: Token Migration
Phase 2: Marketplace Cutover (separate vote required)
The timeline from Proposal 346 ratification to full marketplace cutover spans at minimum 6+ months, assuming no delays in audits or provider onboarding. Akash's team has warned users to await the official portal domain publication, citing phishing risks targeting cross-chain approvals.
Akash's rationale is economic, not ideological. The network currently runs approximately 100 validators securing roughly 15,000 daily marketplace transactions. The cost-per-transaction ratio is, by the team's own assessment, disproportionate.
Transaction cost comparison (per Akash estimates):
| Environment | Cost per Transaction | |---|---| | Akash on Cosmos (current) | ~$0.43 under heavy load | | Akash on Solana (projected) | ~$0.0014 under heavy load | | Akash on Solana (calm) | ~$0.0008 | | Comparable EVM L2 | ~$0.007 |
Estimated daily network fees on Solana: 0.075 SOL, or approximately $11.25 at $150/SOL. That figure covers the protocol's entire daily settlement activity. By comparison, maintaining the current Cosmos validator set, CometBFT consensus layer, and Cosmos SDK dependency stack consumes substantially more in capital and engineering resources — costs the team describes as providing "no direct business value to the computing marketplace."
The RFP process to evaluate alternative homes ran from December 2025 through August 2026. Solana was selected. The decision factors: DePIN ecosystem concentration, stablecoin liquidity ($15 billion supply on Solana), active user density (2.2 million+ daily active paying users), and near-zero marginal cost for scaling transaction throughput.
Akash's departure follows a pattern that has accelerated in 2026.
Noble (departed March 2026): The stablecoin issuance chain that served as Cosmos's primary USDC conduit announced on January 20 that it would sunset its Cosmos app-chain for an EVM-based Layer 1. It launched the replacement in March. Circle will discontinue USDC and CCTP V1 on Noble by January 2027. New USDC minting via Circle Mint was disabled on October 13, 2026. CCTP V1 burn limits begin reducing to zero on October 31.
Sei Network (departed April 2026): Sei executed its SIP-3 migration on April 6-8, dropping Cosmos SDK, CosmWasm, and IBC functionality to become EVM-only. The stated goal: eliminating hundreds of thousands of lines of code to clear a path toward 200,000+ transactions per second. Exchanges and custodians were given until June 15 to migrate holdings to EVM addresses.
Akash Network (vote in progress, October 2026): 300 million tokens to Solana. Marketplace cutover pending second vote.
Additional departures include DATA Network's migration to Sui. Each project cites variations of the same thesis: the cost of maintaining a sovereign chain — validators, consensus, SDK maintenance — exceeds the benefit, particularly when general-purpose chains offer equivalent or superior infrastructure at lower cost.
The Cosmos Hub itself retains over 100 live chains connected via IBC. But the departure of high-profile, revenue-generating protocols raises a structural question: if the most economically active chains leave, what remains to justify the ATOM token's economic role as a hub asset?
The migration is not occurring from a position of strength. Akash's compute marketplace revenue tells a nuanced story.
According to Messari's State of Akash Q1 2026 report, lease revenue compressed 45% quarter-over-quarter from $460,510 to $253,250. Annualized, that implies roughly $1 million in marketplace revenue.
Akash's own reporting claimed $5 million in quarterly compute spending for Q1 2026 — a roughly 5x discrepancy with Messari's independent tracking. By June, Akash reported cumulative compute spending of $6.4 million, with 291% year-over-year growth. These figures may reflect different measurement methodologies — gross compute spending versus net lease revenue captured on-chain — but the gap warrants scrutiny.
Q1 2026 usage metrics (per Messari):
The picture is mixed. Lease counts are growing. GPU utilization is low. Provider counts are shrinking. Revenue is either $1 million or $5 million annually depending on who is counting. Against this backdrop, the Solana migration is as much about reducing the denominator (infrastructure costs) as it is about growing the numerator (marketplace revenue).
Akash joins an established cluster. Solana-based DePIN projects collectively hold approximately $3.5 billion in market capitalization.
Solana DePIN ecosystem (selected projects):
| Project | Category | Approximate Market Cap | Key Metric | |---|---|---|---| | Helium | Wireless | ~$260M | 980K+ hotspots, ~$30M projected annual revenue | | Render | GPU compute | ~$1.65B | 12,000+ registered GPUs | | io.net | GPU compute | — | $7.6M annual revenue | | Nosana | GPU compute | — | Native Solana GPU network | | GEODNET | Geospatial | — | Revenue growth leader |
Helium and Render both migrated from their own chains to Solana in prior years, establishing the migration pattern Akash now follows. io.net and Nosana were built natively on Solana.
The clustering effect is self-reinforcing. Composability between DePIN protocols — compute, wireless, storage, geospatial — increases when they share a settlement layer. Solana's $15 billion stablecoin supply provides immediate payment rails. The 2.2 million daily active paying users represent a potential demand-side pool that Akash's Cosmos-based chain could not access.
Cumulative DePIN protocol revenue across Solana-tracked protocols crossed $22 million since January 2025, with $2.4 million in February 2026 alone.
ATOM trades at $1.64 as of October 27, 2026. For context:
The departures of Noble, Sei, and Akash do not remove IBC connections from the Cosmos Hub. Over 100 chains remain. But these departures remove economically significant protocols — the ones generating measurable transaction volume, fee revenue, or stablecoin flow. Noble's departure is particularly consequential: it eliminates Cosmos's primary pathway for native USDC issuance and cross-chain USDC transfers via CCTP.
The "app-chain thesis" — the proposition that specialized, sovereign chains connected via IBC represent a superior architecture — has not been invalidated in theory. It has been undermined in practice by the economics. Running 100 validators for 15,000 daily transactions (Akash), or maintaining a full consensus stack for stablecoin issuance when Circle will natively support dozens of chains (Noble), produces cost structures that the underlying revenue cannot justify.
The dissenting voice in Akash's vote — the cosmosrescue validator node, commanding 0.63% of voting power — flagged timing risks between the token migration and marketplace cutover. But 98.42% of participating stake voted yes. The economic argument, it appears, is settled for Akash's community.
Akash's migration vote is not an isolated event. It is the third data point in a seven-month trend of economically significant protocols concluding that sovereign Cosmos chains cost more than they deliver. The app-chain thesis proposed that sovereignty — control over consensus, governance, and fee structures — was worth the overhead. For Akash, Noble, and Sei, the answer in 2026 is no.
The implications extend beyond Cosmos. Any protocol operating its own validator set should be running the same cost-benefit calculation Akash performed: does maintaining independent consensus infrastructure generate economic value proportional to its cost? For a compute marketplace processing 15,000 daily transactions, the answer was $11.25/day on Solana versus substantially more on a sovereign chain.
What remains unclear is whether the departures represent a verdict on Cosmos specifically or on the app-chain model generally. The answer likely depends on scale. A chain processing millions of daily transactions may justify sovereign infrastructure. One processing 15,000 cannot. The market, at 98.42% approval, has spoken on where Akash falls on that spectrum.