Radix Week in Review: The Protocol Outgrows Its Company

The great chartered companies of the seventeenth century did a strange thing when their settlements succeeded: they wound themselves down. Having fronted the capital, drawn the maps, and proven the harbour, the company would pre-pay the winter's stores and step back, leaving the colonists to invent a government, run the docks, and decide what to build next. The enterprise was measured by a single test: whether the settlement it founded could outlive it.

Radix reached that moment some time ago, and this was the week you could watch it happen in real time. The Foundation is in maintenance mode, its critical infrastructure pre-funded through late 2026 while it waits for a community entity to take over operations. The week's Radix news was written entirely by the settlers – the Rust engineers rebuilding the protocol's scaling layer, the toolsmiths filling gaps the core team never got to, and the councils arguing, at length, over who holds the keys.

The Machine Changes Hands

The largest single thing the Foundation built and then let go of is Hyperscale, the sharded architecture meant to carry Radix to the throughput a global financial system actually needs. When the Foundation completed the interim phase, it shipped a baseline: working software, the test scripts, the network configurations, and a proven number – a 500,000-transaction-per-second public test of real swaps across shards, with cross-shard atomicity intact. The point of handing over a baseline rather than a promise is that progress stops depending on a Foundation mandate. Anyone can pick up the survey and keep building.

Someone did. hyperscale-rs, the community's ground-up Rust implementation of that architecture, is now the place the R&D actually lives, and this week it was visibly closing on its first milestone. Governance of the network's own parameters has landed on-ledger, which leaves one hard problem between the team and Milestone 1: state contention. When two transactions want the same substate – the price of a single DEX pool, say – you cannot simply shard your way past them, and the team's current work is a batching scheme that lets one shard sequence the contended writes under a single amortized lock. It is the least glamorous and most load-bearing kind of engineering, the sort a closed research team used to do behind a wall and a community now does in a Telegram channel with the Foundation supplying the compute for scaled tests. The performance baseline is inherited; the road past it is being paved by hand.

The Toolsmiths

A protocol is only as usable as the primitives sitting one layer above it, and for years several of Radix's most basic ones lived only in JavaScript. This week a community developer, genkipool, published a native Rust SDK that brings the off-ledger building blocks – ROLA login verification, transaction building and signing, key management, address derivation – into a language the backend and infrastructure world actually deploys. It ships dual-licensed under MIT and Apache-2.0, and it carries a detail that reads like a postcard from 2026: a built-in MCP server, so an AI agent can hold a conversation with a Radix wallet. It has one star and eight commits. That is exactly what day-one community infrastructure looks like, and it is worth watching rather than crowning.

Alongside it, Hookah went fully open. Once a hosted service, it is now an MIT-licensed, self-hostable platform that lets a dApp authenticate with ROLA, register event triggers, and receive webhook deliveries the moment matching on-ledger events cross the transaction stream – the connective tissue that turns a ledger into something an application can react to. The pattern in both cases is the same one the whole ecosystem is living through: capability that used to be somebody's product becoming everybody's commons.

Learning to Govern

The hardest thing a settlement inherits is the question of who runs it. That question dominated the Radix Developers and main community channels all week, and for once the argument was the point. The Radix Accountability Council, the emerging DAO, and its working groups are being stood up in public on RadixTalk, and the debates were the real ones a young republic has: how much authority a council should hold, where legitimate moderation ends and censorship begins, how to keep the surface area of any single body small enough that no one can capture it. A community that intends to be worth a great deal one day is teaching itself, now, to be paranoid about concentrated power. For the fuller frame, the wiki's Radix governance page tracks how the pieces fit.

Stewardship showed up in the plumbing too. A community operator running the bulk of the Stokenet test network laid out a plan to reset it in Q3 – a fresh genesis, no migration, a clean start – specifically to bring hosting costs down to something the community can sustain without a corporate budget behind it. It is an unglamorous line item, and it is precisely the kind of decision that used to be made inside a company and is now made in the open by the person paying the bill. The same current runs through the wallet-stewardship proposals circulating this week: keep the mobile wallet maintained and lightly improved through the horizon of the next major upgrade, and decide as a community what it should become.

None of this is a launch. That is the story. A week with no Foundation announcement, no headline number, no company at the podium, and yet the protocol moved – its scaling layer advanced, its tooling deepened, its governance took shape – because the people who use it have started to act like the people who own it. The colony is learning to run the docks. Next week it keeps building.

Sources

• Radix Foundation – Moving to Maintenance Mode (Apr 28, 2026)
• Radix Foundation – Hyperscale: Completing the Interim Phase
• Hyperscale Update: 500k+ Public Test Done
• genkipool – radixdlt-rust-sdk (GitHub)
• xstelea – Hookah (GitHub)
• RadixTalk – Governance
• radix.wiki: hyperscale-rs, Hyperscale 500K TPS Public Test, Hookah, Radix governance
• Radix Developers and main community Telegram (public discussion, July 6–12, 2026)