Overview
Radix Labs is the research and development division of the Radix Foundation, and the successor development entity to RDX Works, continuing core protocol development for the Radix network. Its stated mission is to deliver the Xi'an protocol upgrade – the fully sharded mainnet that brings linear scalability to production without sacrificing atomic composability.
Xi'an's consensus design is the fully sharded Cerberus protocol, building on the earlier Cassandra research prototype. In parallel, the community-led hyperscale-rs initiative has emerged as the leading production candidate, diverging from Cerberus toward a HotStuff-2–derived per-shard consensus. Radix Labs coordinates with the Radix Accountability Council on protocol governance.
Roadmap
According to the Radix Labs roadmap, published on 30 November 2024 and not revised since, the Xi'an track targets an Alpha release in early 2027, a Beta later in 2027, and a full launch in the second half of 2027. A large part of the work is reimplementing the protocol entirely in Rust, moving away from the Java-based hybrid stack of Babylon, and hardening the consensus mechanism into a production network candidate.
The current live network is Babylon (2023–present), which delivered smart contracts and dApps via Scrypto. Xi'an is the next major milestone on the path to Hyperscale-grade linear scalability.
Development stopped in 2026
Radix Labs has published no plan since the roadmap of 30 November 2024, written by co-founder Dan Hughes. It set out three tracks running in parallel: Cassandra serving as a test platform until it was decommissioned late in 2025; the sharding upgrade to the Radix Engine completed in Q3 2025 and soft-audited in Q4; and Xi'an implemented in Rust across 2026, for an Alpha in early 2027, a Beta at mid-year and a launch in the second half. The Foundation closed the Hyperscale work it had been running on 20 February 2026 with Interim Hyperscale: Closing the Chapter, which handed the codebase to the community to carry forward.
The Radix Foundation then ended its own development. Its update of 28 April 2026 says that from May active development of new features and services becomes minimal
, and that by the end of June, the remaining full-time development team will have completed handover of active workstreams
, after which directors manage the financial and governance functions the transition needs. That leaves the Foundation's research and development arm without a development team, which is why this page records Radix Labs as dormant rather than active. Its predecessor RDX Works is recorded the same way.
Nothing has been withdrawn. The Radix Labs page, read on 9 September 2026, still calls the Labs the research and development arm of the Radix Foundation and still dates Hyperscale to 2027. That page is not maintained: its endorsement carousel repeats one quotation five times under the name Jamie Diamond
, a misspelling of the JPMorgan Chase chief executive Jamie Dimon, which is placeholder copy left in an unfilled content list rather than a customer endorsement.
The sharded consensus layer the roadmap describes is being written elsewhere. hyperscale-rs, a Rust implementation built outside the Foundation, is the leading production candidate for it, and funding and stewarding the Xi'an upgrade is an open item for the development working group of the DAO now being incorporated.