The decentralized compute sector attracted between $2 billion and $3 billion in token investment from 2023 to 2025, building infrastructure that solved one problem while leaving the defining problem untouched.
That is the argument put forward by Leo Fan, founder of Cysic, who contends that every major decentralized compute network today decentralizes GPU supply and crypto payments while keeping trust firmly centralized. According to the piece, no leading network can give a smart contract mathematical certainty that outsourced work was executed correctly — making the entire category, in Fan’s framing, “just Airbnb for GPUs.”
The Verification Gap
The numbers illustrate the structural ceiling. Akash generated approximately $11 million in Q3 2025 revenue; Render managed roughly $18 million. Both figures are framed as coordination-layer achievements, not infrastructure milestones, measured against AWS‘s annualized run rate exceeding $100 billion. The gap is not simply one of scale — it reflects which workloads these networks can credibly serve.
Real-world failures in 2025 put the verification gap in concrete terms. Bad actors returned corrupted renders through Render’s network with no onchain mechanism to detect the manipulation. Io.net identified a Sybil cluster gaming reputation scores in May, then faced a second incident in November when a cluster attributed to aPriori claimed 60% of an airdrop across 14,000 wallets. Gensyn‘s own whitepaper, the analysis notes, acknowledges that its “learning game” tolerates less than a 49% malicious node threshold in practice. These are not edge cases — they are the predicted output of social enforcement substituting for mathematical proof.
The consequences extend directly to Web3’s core use cases. A Layer 2 rollup outsourcing STARK proof generation to any current decentralized cloud still requires a trusted multisig or a single honest prover, leaving the centralization risk unchanged. An autonomous agent running inference on io.net operates on a contract that has no mechanism to verify whether the output was correct or compromised. As Fan frames it, the sector has recreated the oracle problem with additional complexity.
What Cryptographic Verification Would Change
The addressable market argument follows directly. Without verifiable execution, regulated and sensitive workloads — financial institutions requiring provable compliance, healthcare systems needing auditable inference, rollups demanding trustless proof generation, AI agents executing high-value transactions — remain inaccessible. The competitive field shrinks to hobbyist rendering and Stable Diffusion workloads, a foundation Fan argues cannot support a trillion-dollar market thesis.
More Read
The piece invokes Vitalik Buterin‘s observation at Devcon 2024: “If your scaling solution reintroduces trusted parties, you haven’t scaled. You’ve just outsourced.” Fan applies that directly — replacing AWS with a thousand smaller AWS-equivalent nodes does not change the trust architecture, it distributes it.
The proposed path requires cryptographic proof accompanying every compute result, specifically zkSNARKs, STARKs, or optimistic fraud proofs, verifiable in under one second by any smart contract. Fan argues this is no longer theoretical. Hardware-accelerated proving using FPGAs and custom ASICs makes the economics viable at GPU-scale bandwidth, with 2024–2025 ZPrize results showing STARKs over cycle-accurate circuits executing in under eight seconds on current FPGA clusters, with sub-second performance as the stated trajectory.
The sector’s current state, the analysis concludes, has decentralized the supply chain while leaving the trust layer intact — a distinction that determines which markets remain permanently out of reach.
Disclaimer: The information provided in this article is for educational and informational purposes only and does not constitute financial or investment advice.
Photo by Pixabay
This article is a curated summary based on third-party sources. Source: Read the original article
