Charles Hoskinson defended major cloud providers like Google Cloud and Microsoft Azure as essential to blockchain scalability at Consensus Hong Kong, claiming advanced cryptography ensures decentralization.
But critics argue that technologies like multi-party computation (MPC) and confidential computing do not eliminate systemic risk. While they protect data visibility, they rely on centralized infrastructure-physical hardware, supply chains, and network access-that remains vulnerable to throttling, shutdowns, or policy intervention.
The real issue isn’t computational capacity within Layer 1 blockchains, but who controls the off-chain infrastructure needed for verifiable computation. Rollups and zero-knowledge proofs depend on proof generation, which often runs on hyperscaler hardware.
Even if protocols are mathematically neutral, hardware ownership and distribution determine who can participate. A few dominant providers can still exert control through bandwidth limits, geographic restrictions, or compliance mandates.
Specialized networks built for high-throughput, deterministic workloads like ZK proving outperform general-purpose cloud platforms by design. Vertical integration reduces overhead and maximizes efficiency for sustained tasks.
The solution? Use hyperscalers only as optional accelerators-not core dependencies. Critical functions like proof storage and final verification must reside on decentralized infrastructure to ensure resilience and sovereignty.
True decentralization requires diversified hardware ownership and protocol-aligned incentives-not just cryptographic promises.