Radix Week in Review: The Consensus Work Is Nearly Done

The week in brief

• The lead developer on hyperscale-rs said the consensus implementation is close to finished; what remains is verification and tightening the security maths.
• A formal model now proves the network can resize its shard groups while running, without breaking.
• Post-quantum cryptography is the near-term item on the roadmap; confidential transfers were considered and deliberately deferred.
• An outside developer opened a pull request on the official Android wallet, the first from outside the core team in over a month.
• Six demonstrations on the test network drove Radix accounts through two agent-payment protocols.
• Candidates began putting themselves forward for the council seats that had been going unfilled.

The consensus layer is the part of a blockchain that decides which transactions happened and in what order, and it is the part that took the field years and a peer-reviewed paper to get right. On hyperscale-rs, the community's Rust rebuild of Radix's scaling design, that part is now essentially built.

Proving it rather than testing it

The remaining work moves to the Radix Engine, the gateway and the desktop validator: awkward jobs, but far easier for other people to pick up than the consensus core ever was.

The week's proof arrived on schedule. The project's eleventh week on its first milestone produced 55 commits and a formal model, written in Quint, that mechanically proves the reshape lifecycle safe. Reshaping is how the network resizes its shard groups while running, by splitting them, merging them and moving validators between them, and it is the sort of thing that is very hard to test by example because the failures hide in orderings nobody thought to try. Pairing deterministic simulation with formal verification checks the protocol against every case at once rather than the ones someone wrote down.

Past that, the roadmap is visible. Post-quantum cryptography is near-term. Confidential transfers, of the kind Solana and Sui offer, were sketched and set aside: a shieldable resource would be straightforward to build, but the cryptography those systems rely on breaks against a quantum computer, and no chain has a durable plan for keeping the feature working afterwards. The reasoning is that it is better to wait for post-quantum privacy to settle and adopt what survives than to ship something with a two-year shelf life.

Hands on the tools

On 17 July a community developer opened pull request #1445 against the official Android wallet, adding view modes for NFT collections. It is the first contribution there from outside the core team in over a month, and it lands while the Radix Foundation is in maintenance mode, so the official wallet is now something the community co-authors.

The same pattern showed up elsewhere: a longtime contributor is building a fresh Babylon wallet; a project called Vulper is readying a beta that shows a holder's lending positions inside the wallet; and six demonstrations went up on Stokenet driving Radix accounts through x402 and AP2, two protocols for letting software agents pay each other. An essay on self-tokenization circulated the same week. All of it is early and none of it is proven.

Filling the council seats

Under the Foundation the Radix Accountability Council advised. Under the DAO that replaces the Foundation it becomes a body of elected delegates who hold the assets and steer the treasury, which is a different job and needs enough people willing to do it.

For most of the week the worry running through the developer and Hyperscale channels was that not enough had come forward, and that the DAO could stall for want of delegates. Then people began standing. Giles Morris filed a formal proposal on RadixTalk for a business and finance seat, unpaid, for three to six months. Others said in the same days that they would serve if wanted, among them a longtime Scrypto contributor and a construction-business owner moving full-time into crypto. RadixTalk has become the governance forum in practice, and the arguments there are sometimes heated, which is what a real handover sounds like.

The week on the ledger

Read from the Radix Gateway as the ledger stood at midday on 19 July 2026, epoch 327,401, against the same reading a week earlier.

Stake rose by 34 million XRD and everything else held still.

Concentration watch

Eight validators held a third of all staked XRD, unchanged, and 15 held half.

This week on the wiki

Read hyperscale-rs for what the milestone actually covers, and Radix governance for how the council fits into what replaces the Foundation.

What to watch

• Whether enough candidates come forward to seat the council, which decides whether the DAO convenes on time.
• The verification passes on consensus, and the work moving into the Radix Engine where more people can join it.
• Whether the Android wallet takes a second outside pull request, which is what turns one contribution into a pattern.

Corrections

This issue originally called the body the council becomes "the MiDAO". MIDAO is the firm that handles Marshall Islands incorporation, not the DAO itself, and the text now says the DAO.

Next: whether the seats fill, and whether consensus clears verification.

Radix Week in Review series, Issue #4: ← Previous: Issue #3, Jul 6–12, 2026 · Next: Issue #5, Jul 20–26, 2026 → · All recaps · Subscribe

Previously: Everything that moved this week was built by the community: the Rust rebuild of Radix's scaling design, a Rust SDK, and Hookah going open source.