Two parallel critiques of Web3's foundational promises are emerging, highlighting structural limitations in decentralized compute networks and governance models. In a detailed opinion piece, Leo Fan, founder of ZK hardware acceleration firm Cysic, argues that the "decentralized compute" sector has fundamentally failed by not solving the trust problem. Concurrently, the crypto industry is witnessing a potential trend as decentralized autonomous organizations (DAOs) consider abandoning decentralization to court institutional capital.
Fan asserts that despite $2-3 billion invested in "decentralized cloud" tokens from 2023-2025, no major network provides cryptographic certainty that computational work was performed correctly. He describes current leaders like Akash and Render as merely "sophisticated spot markets" that have decentralized supply and payments but kept trust centralized. Akash generated approximately $11 million in Q3 2025 revenue, while Render managed about $18 million—figures Fan calls "trivial" compared to AWS's $100+ billion annual run rate.
The article cites multiple real-world failures: In 2025, bad actors returned corrupted Blender renders through Render's network with no onchain detection method. Io.net encountered Sybil attacks in May and November 2025, with one cluster claiming 60% of an airdrop across 14,000 wallets. Gensyn's whitepaper reportedly admits their "learning game" has less than 49% malicious tolerance in practice. "These are the predictable outcomes when you replace mathematical proofs with social enforcement," Fan writes.
This trust gap, according to Fan, caps the Total Addressable Market (TAM) for decentralized GPU networks at rendering and basic training, excluding sensitive workloads like DeFi bots, medical inference, and proprietary models. He quotes Vitalik Buterin's Devcon 2024 statement: "If your scaling solution reintroduces trusted parties, you haven't scaled. You've just outsourced."
The solution, Fan argues, lies in cryptographic proof accompanying every result—using zkSNARKs, STARKs, or optimistic fraud proofs verifiable in under one second by any smart contract. He points to 2024-2025 ZPrize winners demonstrating STARKs over cycle-accurate circuits running in under eight seconds on FPGA clusters, moving toward sub-second verification on next-generation silicon.
Meanwhile, the DAO model faces its own existential challenge. On March 11, 2024, Across Protocol (ACX) proposed transitioning from a DAO to a private U.S. C-corporation through a token-to-equity exchange buyout. Risk Labs, the team behind Across, stated the token and DAO structure "materially" impacted its ability to close deals with enterprises and institutions.
Across co-founder Hart Lambur explained on X that having a token "generally hurts more than it helps" in the current macro environment, noting ACX is down 97.5% from its all-time high. The protocol focuses on stablecoin infrastructure, requiring contracts and offchain payment arrangements ill-suited to DAO structures. The proposal remains a "temperature check," with a governance vote expected in early April.
Industry reaction has been divided. DeFi researcher Ignas called it a "huge failure of crypto" and "a betrayal of the crypto spirit." However, Matthew Pinnock, founder of DeFi project Altura, told Cointelegraph that as the industry moves toward real-world assets and institutional capital, "protocols are running into structural limitations. Institutions typically need a clear legal counterparty that can sign contracts and undergo due diligence, something a decentralized collective cannot easily provide."
Tim Black, product lead at ShapeShift DAO—which dissolved its corporate entity in 2021—acknowledged many teams adopted DAO structures during the last cycle without accounting for operational complexity. "What Across is proposing is essentially admitting that," Black said. "They're saying the DAO experiment helped bootstrap the network, but a company structure is better suited to the next phase." He suggested the industry may split into "corporate crypto" protocols run like fintech companies and genuinely decentralized protocols that accept operational friction.
