Most L1 token designs still assume one thing: security must be bought with a fixed stake. Avalanche changed that with the Etna upgrade (ACP 77). L1 validators no longer lock 2,000 AVAX each. They pay a continuous fee instead, 512 nAVAX per second, about 1.33 AVAX per validator per month. Our L1 today: 13 validators, total weight 13,670. Four nodes carry 12,800 of it, 94% of the weight. Prepaid fee balance is 71.5 AVAX, burning 0.58 AVAX per day, roughly 4 months of runway. That shift moves capital. Nothing idle is locked at the infrastructure layer, so capital moves up to the application layer where $CDK lives: governance, delegation, incentive design. Security is rented from the protocol. The token is free to do real work. Question for L1 builders: if block production is a subscription, what should your token actually secure?

CodeNekt Ecosystem, week of 20 to 26 October. Nine posts. One production L1. And one number most L1 accounts never publish. Bookmark this if you track real world data going on chain.

The uncomfortable part. Four nodes hold 12,800 of the 13,670 weight. That is 93.6%. No new validator joined this week. Decentralization is easy to claim and hard to measure. We publish the concentration while the set is still small enough for the number to be honest.

Most vehicle data on blockchain pitches get the core mechanic wrong. They promise to put the data on chain. You cannot. A single connected car generates tens of gigabytes per hour. No L1 stores that, and none should. What goes on chain is small: a 32 byte hash of each dataset, the issuer attestation, and the access policy. The mechanics, step by step. The raw telemetry stays in the owner store. A cryptographic digest gets anchored per batch, with a timestamp. An access contract holds who may read what, and under which condition. Any third party recomputes the hash and verifies integrity without ever touching the raw record. The outcome is provenance without exposure. You can prove a mileage record was not altered after the fact, without publishing the record itself. That is the entire design of an on chain Digital Product Passport. Not a database migration. A verification layer on top of one.

CodeNekt L1, state of the network, Sunday 25 October 2026. 13 active validators. 13,670 total weight. 72.14 AVAX of prepaid validator balance. Here is what is live, what is not, and what ships next.

Next, in this order. 1. Restore public RPC access so builders read the chain without asking us. 2. Top up validator prepaid balances before runway gets short, and widen the top of the set. 3. Put the on chain Digital Product Passport path in front of real vehicle data. No dates until each one is verifiable.

Most real world asset projects make the same architectural mistake. They mint a token for the asset, then keep the records in the owner's wallet. Sell the car and the history stays behind with the seller. Three owners later, the vehicle's past is split across three addresses nobody can reconcile. ERC-6551 fixes this at the standard level. The asset token gets its own smart account, derived deterministically from the chain id, the token contract and the token id, then deployed with CREATE2. Records live in that account. Transfer the token and the account goes with it. The full history follows the vehicle across every owner, no migration script, no database reconciliation. That is the difference between tokenizing an asset and tokenizing its identity. Ownership changes. The record does not.