Avalanche
KLEAR3 ResearchPublished Updated
AVAXAvalanche, a proof-of-stake network whose own documentation now calls every one of its chains an L1 rather than a subnet, has published code since March 2020 and run its mainnet since that September. Every transaction fee it collects is destroyed outright, paid to nobody at all.
What does Avalanche actually do?
Avalanche runs a proof-of-stake network secured by Snowman consensus, splitting work across three built-in chains plus any number of additional networks it calls L1s rather than subnets. AVAX, its native asset, pays every transaction fee, is staked to produce blocks, and is destroyed outright rather than handed to whoever processes the transaction.
Avalanche describes itself, in its own support documentation, as a network “designed to be customizable, scalable, and interoperable”, where “everything on Avalanche is an L1, and every chain is part of an L1”. That second sentence replaces a word the project used for years: what it called a subnet, it now calls an L1.
Source Avalanche Support, read 2026-10-18
Joining the network at all means running the Primary Network, which bundles three chains: the Platform Chain (P-Chain) coordinates validators and staking, the Exchange Chain (X-Chain) moves assets, and the Contract Chain (C-Chain) runs an Ethereum-compatible virtual machine for smart contracts. Every additional L1 is optional; the Primary Network is not.
Blocks are agreed on through Snowman, a protocol the project’s own documentation describes as combining classical and Nakamoto techniques for fast finality at low energy cost, a different design choice from the peer-reviewed Ouroboros that secures Cardano, and from the plain proof-of-stake Ethereum switched to in 2022.
Source build.avax.network, read 2026-10-18
Is the code actually open?
Published under BSD-3-Clause since 2020, with 163 versioned releases: a permissive licence that lets anyone fork the node and keep their own changes closed, unlike the copyleft terms Ethereum’s client carries. One independent review exists, and it found nothing above informational severity.
The node is published under BSD-3-Clause, a short, permissive licence: anyone can fork it, modify it, and ship a closed derivative without publishing a single line back. That is a lighter obligation than the LGPL-3.0 terms Ethereum’s best-known client carries, where touching the client itself forces the changes into the open.
Source GitHub, read 2026-10-18
163 numbered releases sit in the repository, each one a point a node operator can pin to rather than tracking a moving branch, the same discipline Ethereum’s client keeps, and a different one from protocols that publish no releases at all.
Source GitHub, read 2026-10-18
One independent review has been opened and read in full: Trail of Bits spent ten weeks on the node in August 2025 and reported five findings, every one of them informational, nothing critical, high, medium or low. That is a clean result, and it covers one review at one point in time, not every change made since.
Source Trail of Bits, read 2026-10-18
Is anyone still working on it?
Yes: 211 commits from 18 distinct Git identities in the 90 days to 18 October 2026, and a separate improvement-proposal repository that has merged 180 pull requests since October 2023, publishing 36 numbered proposals so far. Code and governance both move on public, dated records.
That commit count is a window, not a total, and it moves on its own: the fenced period ends the day we read it and begins exactly 90 days earlier. Eighteen Git identities are a floor on the number of humans behind them, not a headcount: one contributor can sign with several, and several can share one.
Source GitHub, read 2026-10-18
The repository that decides how the protocol itself changes is a separate, public one. avalanche-foundation/ACPs has merged 180 pull requests since it opened in October 2023, publishing 36 numbered proposals covering everything from dynamic fees to how subnets, now L1s, get renamed. Anyone can write one; what makes it real is that validators choose to run software that implements it.
Source avalanche-foundation/ACPs, read 2026-10-18
What shows it is running
- The chain responds — block 97,852,359
- The node repository is still being worked on — 211 commits over the 90 days to 2026-10-18, 18 distinct Git identities
Source api.avax.network · GitHub, read 2026-10-18
- The node repository goes public
- Avalanche mainnet goes live
- The public proposal repository (ACPs) opens
- This page, checked and published
Where does it actually run?
On its own network, verified directly: Avalanche’s official public endpoint answered with block 97 852 359 on 18 October 2026. The chain it runs on is one of three built into the Primary Network, alongside a growing number of additional L1s the project no longer calls subnets.
We asked Avalanche’s own public endpoint directly, rather than an aggregator: a request for the current block number on the Contract Chain returned 97 852 359 on 18 October 2026. That is the chain most smart contracts on Avalanche actually run on.
Source api.avax.network, read 2026-10-18
Unlike Ethereum, whose virtual machine the Contract Chain is compatible with, nothing here bridges value in from another network to be counted twice: the chains that make up the Primary Network, and every additional L1 beside them, are Avalanche’s own.
Network verified by reading its current block height on a public node.
What the token actually does
AVAX pays every transaction fee and is staked to produce blocks, but neither payment goes to anyone: validators earn new AVAX through staking rewards instead, and every fee collected is destroyed outright. Total supply is capped at 720 million; a little under two thirds already circulates.
Every transaction pays a fee in AVAX, and on the Contract Chain that fee splits the way Ethereum’s does, into a base amount and a priority tip. The difference is what happens next: both halves are destroyed, not just the base. Nobody, validator or otherwise, receives any part of it.
Source build.avax.network, read 2026-10-18
Supply is capped at 720,000,000 AVAX, of which 443,574,503 already circulates, a figure read from an aggregator rather than from the chain itself. What we did not measure is the rate at which staking rewards mint the rest, the schedule that eventually decides whether burning catches up with issuance.
Source CoinGecko, read 2026-10-18
- In circulation443,574,502.8
- Issued, not circulating26,668,618.08
- Not yet issued, up to the cap249,756,879.12
- What the token is forDirectly verified
- How many existThird-party supported
- How new ones appearNot verified
- Who can change the rulesNot verified
The KLEAR3 rating
3.7out of 5 in the KLEAR3 rating
Each axis gets 0 to 5 from the facts in the register, against thresholds written in advance. The overall score is their weighted average, and it only appears once at least 5 of the 6 axes rest on a measured fact.
- Audits ×1.3
- Open code ×1.2
- Business model ×1.1
- Token alignment ×1
- Governance ×1
- Maturity ×0.7
- Open code5
public repository · BSD-3-Clause licence · 163 published releases · 211 commits in 90 days
- Audits3
1 firm on the core · 1 audited scope · report opened: 0 critical, 0 high, 0 medium · no critical or high finding · code reviewed in the open: 211 commits in 90 days, 6 years in production
What holds it back audit stops at 2025-08: the code has moved since
- Business model5
business model published by the protocol · paid by whoever sends the transaction · collected by nobody: no validator and no miner ever receives it · destroyed outright, which shrinks total supply for the benefit of every holder
- Governance3
decision mechanism published · 36 improvement proposals published
- Token alignment2
supply capped at 720 million · 62% already in circulation
What holds it back emission schedule unknown
- Maturity4
6 years of public code · 1 network verified by us
We rate the protocol: how it is built, how much of it can be checked, how its token lines up. Never an investment. A high score is not advice to buy, and a low one is not advice to sell.
What we could not establish
How new AVAX actually enters circulation through staking rewards, whether the five informational findings in the one audit we opened were ever fixed, and how much real influence Ava Labs retains over which community proposals a validator ends up running. We read the repositories, not every arrangement behind them.
Avalanche is not built by an anonymous team: Ava Labs names itself, on its own site, as the company behind the core software, founded in 2018. What it does not spell out, in anything we read, is how much practical weight it still carries once a community proposal reaches the point where validators decide whether to adopt it.
Source avalanche.com, read 2026-10-18
We also did not establish the exact staking-reward schedule that mints new AVAX, or whether the five informational findings in the one audit we opened have since been addressed. An audit is a photograph of one moment; the repository has moved since.
Frequently asked questions
Who actually receives Avalanche’s transaction fees?
Nobody. The project’s own documentation states the fee is burned, destroyed forever, and on C-Chain that covers both the base fee and the priority tip. Neither a validator nor a miner collects any part of what you pay to send a transaction.
Ethereum burns only the base fee and still pays the tip to whoever produces the block; Bitcoin pays a miner the whole amount. Avalanche destroys both halves on every one of its three built-in chains, which means nobody’s incentive to include your transaction depends on what you paid.
What still pays validators is staking: new AVAX minted as a reward, separate from anything a sender hands over. A fee that nobody can keep is also a fee nobody can be bribed to prioritise, which is a different kind of guarantee than a low gas price.
Is “subnet” still the right word for part of Avalanche?
No, and the project changed its own vocabulary to say so. Its support site now states that everything on Avalanche is an L1, and every chain is part of an L1, replacing the term it used for years to describe exactly the same thing.
Only one grouping keeps the old name in practice: the Primary Network, the three chains every validator must run together. Everything else the project now calls an L1, whether it is permissioned, permissionless, public or private, and whether it existed before the renaming or launched after it.
A directory still calling these subnets is not simply using an older synonym: it is describing a structure the project itself no longer draws that way, which matters the moment you are deciding where a validator’s real obligations actually start and end, and where they stop.
Does AVAX carry a governance vote the way ADA or PENDLE do?
No formal weighted vote was found anywhere we looked. Protocol changes advance through a public repository of numbered proposals that validators adopt simply by running compatible software, not by casting AVAX in a ballot the way staked ADA or locked PENDLE does.
The repository has merged 180 pull requests since it opened in October 2023, publishing 36 numbered proposals, and anyone can submit one. What decides whether an accepted idea actually ships is whether enough validators choose to run the client that implements it.
We did not measure how concentrated that choice really is in practice, or how much weight Ava Labs itself still carries once a proposal reaches that final stage of client adoption. A process that is open to read is not automatically a process where every participant holds equal sway over what ships.
Where to read on, at the source — every address answered 200 on 2026-10-18.
Next: our Solana profile covers another high-throughput proof-of-stake network, and our methodology sets out what counts as a source here.
Ask an assistant to summarise it and check what it claims.