Headline: Cardano vs. Cysic at Consensus Hong Kong — a fight over whether cloud giants undermine decentralization
At Consensus Hong Kong 2026, a debate over the role of hyperscalers in blockchain infrastructure took center stage. Cysic founder Leo Fan warned that leaning on cloud giants like Google Cloud and Microsoft Azure risks hollowing out crypto’s decentralization ideals — comments that followed Cardano founder Charles Hoskinson’s presentation of Midnight, Cardano’s privacy-focused project.
What Hoskinson announced
Hoskinson used the stage to outline Midnight, a privacy-first layer built to handle heavy cryptographic workloads such as zero-knowledge proofs. He confirmed partnerships with companies including Google and Telegram and said Midnight’s mainnet is slated to launch at the end of March. Fahmi Syed, CEO of the Midnight Foundation, added that the network will debut with 10 federated nodes as part of a “responsible” approach to decentralization, with Google Cloud among the early infrastructure collaborators.
Hoskinson defended the decision to work with hyperscalers by arguing that large-scale, privacy-preserving systems require vast compute capacity. “When people spend a trillion dollars building data centers, we should probably use what they spent the trillion dollars on instead of trying to build a completely different network,” he said. He described Midnight as a neutral coordination layer that routes heavy workloads — particularly those tied to privacy and zero-knowledge cryptography — to cloud providers. Technologies such as multi-party computation and confidential computing, he argued, let providers supply hardware without accessing the underlying data.
During a live demo, Hoskinson said Midnight processed thousands of transactions per second with Microsoft Azure powering the backend compute layer.
Cysic’s counterpoint
Fan pushed back, arguing that relying on a small set of global data-center operators creates a structural centralization risk even if nodes and data are technically insulated. “If your validators look decentralized but all run in the same data center, that’s still a single point of failure,” Fan told CoinDesk. “Blockchain is supposed to remove single points of failure. If the infrastructure is centralized, that’s a contradiction.”
Cysic runs a decentralized compute network focused on zero-knowledge proof generation and says its distributed approach can outperform cloud providers in some tasks. Fan noted one customer cut proof-generation times from as much as 90 minutes on AWS to roughly 15 minutes using Cysic’s hardware network. “In some scenarios, we can deliver better performance,” he said. “We don’t need to defeat them immediately, but we can compete.”
Where the disagreement really lies
The clash isn’t only technical — it’s philosophical. Hoskinson is prioritizing cryptographic neutrality and pragmatic use of existing cloud capacity to achieve scale and enterprise readiness. Fan insists decentralization should extend to the compute layer itself, not just to protocol governance or node count. Even if computations are encrypted and workloads can be shifted between providers, concentrating demand for GPUs and data-center capacity among a few hyperscalers still concentrates power.
A middle path
Fan stopped short of calling for an outright rejection of hyperscalers. Instead he advocated a hybrid model: “Use big vendors in a limited way. Combine them with decentralized networks to make the system more robust. Do not give up decentralization because that’s the nature of our community.”
Why this matters
As blockchains push toward enterprise adoption and global scale — and as privacy-focused apps demand heavy GPU and compute resources — the industry faces a choice. Build parallel, decentralized compute infrastructure that challenges Big Tech’s dominance, or integrate with existing hyperscale providers and trade some infrastructure independence for immediate capacity and speed. The resolution of that choice could shape how decentralized “decentralization” actually is in practice.
Read more AI-generated news on: undefined/news
