Hoskinson’s Decentralized Compute Vision Faces Scrutiny

Cardano founder Charles Hoskinson’s argument about decentralized compute infrastructure is facing technical scrutiny from industry analysts. According to critics, Hoskinson’s assertion that hyperscalers are necessary because no single Layer 1 blockchain can handle global computational demands overlooks the distributed trust models that cryptographic tools enable, even when operating on centralized infrastructure.

Hoskinson presented his perspective on decentralized compute at Consensus Hong Kong 2026, arguing that the computational scale required for global systems—representing trillions of dollars in data center infrastructure—necessitates reliance on hyperscalers. Critics, including analyst Fan, countered that this argument misses the point of distributed cryptographic solutions that can reduce, though not eliminate, infrastructure concentration risk.

The dispute centers on two primary cryptographic approaches: Multi-Party Computation (MPC) and Trusted Execution Environments (TEEs). MPC distributes cryptographic key material across multiple parties so no single participant can reconstruct secrets, meaningfully reducing single-node compromise risk. However, this approach expands the security surface across coordination layers, communication channels, and node governance. The single point of failure transforms into a distributed trust surface rather than disappearing entirely.

TEEs encrypt data during execution, limiting exposure to hosting providers. Yet according to security research, side-channel and architectural vulnerabilities continue emerging across enclave technologies. More critically, both MPC and TEEs typically operate atop hyperscaler infrastructure, meaning physical hardware, virtualization layers, and supply chains remain concentrated. If an infrastructure provider controls access to machines, bandwidth or geographic regions, it retains operational leverage, the analysis states. Cryptography prevents data inspection but cannot prevent throughput restrictions, shutdowns, or policy interventions.

Hoskinson’s position emphasizes that Layer 1 blockchains cannot feasibly match the computational throughput of centralized providers. However, proponents of decentralized solutions argue that zero-knowledge proving and distributed network architectures could handle critical computational components without requiring centralized rental from hyperscalers, thereby preserving decentralization principles.

This debate will likely shape how blockchain projects architect decentralized compute solutions. The outcome may influence whether future infrastructure relies on hybrid models combining cryptographic protections with distributed networks, or continues accepting hyperscaler dependency as inevitable.

Follow Hashlytics on Bluesky, LinkedIn , Telegram and X to Get Instant Updates