What Is Avalanche – and Do Its Blockchains Still Need AVAX?

September 2026 was AVAX's best month since November 2024: up 51 percent, against 6 percent for bitcoin. The headlines fit. At the end of September, cryptoticker and Stocktwits reported that Goldman Sachs was bringing a 100-billion-dollar fund to Avalanche. In July, Bridgetower announced it was tokenizing 11 billion dollars of Arizona mineral assets on Avalanche. FIFA runs its own blockchain there, and Nexon runs MapleStory.

The buying case behind it sounds logical. The more companies build their own blockchains on Avalanche, the more AVAX they need. Until December 2024, that is how it worked.

If that is why you buy AVAX, you should know one number: how much AVAX do Avalanche's chains pay today?

What Avalanche does: one main network and many custom chains

Avalanche has been live since 21 September 2020, built by Ava Labs under Emin Gün Sirer, a former computer science professor at Cornell. The network has two parts.

The first is the main network with three chains. The one that matters to you is the C-Chain. It works like Ethereum, the same wallets and apps run there, and it holds DeFi, stablecoins and most tokens. Anyone who does something on the C-Chain pays the fee in AVAX, and that fee is burned.

The second part is custom blockchains that companies and projects launch on Avalanche. They used to be called subnets; today they are L1s. An L1 has its own validators, its own rules and usually its own token for fees. The FIFA chain is one, as are Henesys for MapleStory and the game chain of Off The Grid.

Avalanche does not sell one blockchain. It sells a kit for building many.

What is technically different: fast finality and chains made to measure

Consensus. Avalanche has no block producer that decides what counts. Instead, every node repeatedly asks 20 randomly chosen validators which block they consider correct. If at least 15 give the same answer, that is one round. After 20 successful rounds in a row, the block is final. That takes about one to two seconds. For comparison, an Ethereum block is final after about 15 minutes, a Solana block after about 13 seconds.

L1s. Anyone who wants their own chain can set almost everything on Avalanche: who may validate, which token pays the fees, whether the chain is public or only for approved participants. That is why banks, game studios and FIFA ended up here. The chains talk to each other through Avalanche's own messaging system (Interchain Messaging). A message is only as secure as the validators of the chain it comes from. For a company chain with five validators, that means five.

Weaknesses. Just 24 validators together hold more than a third of the stake, 44 more than half. On 23 February 2024 finality stopped across the whole main network for about five to six hours because of a software bug. And of 569 registered subnets, only 66 have validators with a balance today; Avalanche's own statistics page counts 10 chains with more than 1,000 transactions a day.

Technically Avalanche is fast and flexible. The question is how much of that reaches the token.

What is verifiably running today

As of 7 October 2026, 30-day window 7 September to 6 October.

So real money and real gaming run on Avalanche. The money sits almost entirely on the C-Chain. The well-known L1s are either games with their own token or, like FIFA's, barely used.

On Avalanche, what makes headlines and what happens on-chain are two different lists.

Who uses Avalanche: the numbers halved in late September, on purpose

Look at C-Chain usage and you see a crash. In August it averaged 2.23 million transactions a day from about 400,000 active addresses. Since 27 September it is 303,000 transactions and about 62,000 addresses.

Nobody walked away. With the Helicon upgrade of 22 September, validators can vote on the minimum price of a transaction. Between 25 and 27 September they raised it in three steps to roughly 80 to 90 times its previous level. A transaction now costs about 6 cents instead of a fraction of that. The proposal behind it (ACP-283) names spam as the reason.

The result: transactions fell 86 percent, while fees paid rose from 509 to 1,660 AVAX a day. A large part of the old usage was bot load that only ran because it cost almost nothing.

Compared with other chains, same 30-day window, without system transactions:

7 September to 6 October 2026. Sources per chain in the evidence appendix. Solana counts differently from the Ethereum-style chains; only the EVM chains compare cleanly with each other.

By transactions the C-Chain sits near the bottom of the chains compared here. By fees it is close to Arbitrum, far behind Ethereum, Tron or Solana.

Since 27 September the C-Chain measures less, but more honestly.

The core: what Avalanche's chains pay in AVAX today

Until December 2024 there was a hard rule. Anyone who wanted to validate a subnet also had to validate the main network and stake at least 2,000 AVAX for it. A company chain with five validators tied up 10,000 AVAX. That was the mechanism behind "every new chain needs AVAX".

On 16 December 2024 the Etna upgrade went live. Two days later the research firm K33 summed it up: it "eliminates the requirement for validators to stake 2,000 AVAX (~$96,000), replacing it with a continuous fee system". Since then an L1 validator pays an ongoing fee of about 1.33 AVAX a month, as long as fewer than 10,000 L1 validators are active in total.

What that adds up to today is on the P-Chain:

P-Chain and Glacier API, 7 October 2026. 2,010 L1 validators are registered, 1,433 of them without a balance and therefore inactive.

At about 11 dollars, 9,200 AVAX is roughly 100,000 dollars a year. All of Avalanche's own chains together pay the network about as much as a single validator used to post as collateral. Whether this fee is burned is not specified in the underlying proposal, ACP-77.

For comparison, staking creates about 10.2 million new AVAX every year. For every AVAX the chains pay, more than 1,000 new ones are added.

On the main network, the number of validators has fallen from 1,567 to 587 since Etna. It fell fastest in mid-2025, when stakes started before Etna (one year at most) ran out: from 1,374 at the start of June to 906 at the start of September. That fits the new rule, but it does not prove that each of them was a subnet validator.

The L1s use Avalanche's technology. They hardly need its token for it anymore.

Which leaves the question whether the chains need AVAX some other way.

The obvious objection

"The fee doesn't matter. The L1s bring companies, money and users into the ecosystem, and that drives AVAX through the C-Chain."

That is the right place to look. The real money sits on the C-Chain: 483 million dollars in BUIDL, 637 million dollars in DeFi, 1.70 billion dollars in tokenized assets that can actually be transferred. Anyone who uses them pays in AVAX, and the fee is burned.

But those fees are small. Over the last 365 days the C-Chain collected at most 296,000 AVAX. Even at the new pace after the spam brake, 1,660 AVAX a day, it would be about 610,000 AVAX a year. Against that stand about 10.2 million new AVAX from staking, 17 to 35 times as much.

And the big numbers from the headlines do not land there. Of the 11.40 billion dollars rwa.xyz lists as "represented on Avalanche", 11.06 billion is a single entry: Bridgetower's mining assets in Arizona. They can be neither transferred nor traded. Goldman Sachs offers its fund through the Lynq platform, whose shares do not sit on Avalanche as tokens.

Demand for AVAX comes from the C-Chain, and the C-Chain burns less than a tenth of what is newly created each year.

Tokenomics: who holds AVAX and what gets added

Supply. At most 720 million AVAX. 475 million have been issued, 443 million are in circulation according to CoinGecko.

New AVAX. About 10.2 million a year from staking, around 2.2 percent of supply. On top, the Foundation releases 1.667 million AVAX every quarter, the next tranche on 24 October 2026, the last in July 2030. Since 22 September Helicon has been lowering the staking yield step by step, but only for new stakes; the full effect arrives at the end of 2027.

Burned. 5.10 million AVAX in total since launch, almost all from C-Chain fees. Most of it dates from 2021 to 2023, when DeFi on Avalanche boomed.

Staked. 180.9 million AVAX across 587 validators, about 41 percent of circulating supply. The yield is about 5.7 percent a year, and it comes from new AVAX, not from fees.

Companies and funds. Two listed treasury companies together hold about 27 million AVAX, a little over 6 percent of circulating supply. The three US ETFs from VanEck, Grayscale and Bitwise hold about 5 million combined.

The treasury companies deserve a second look. According to its filing, Avalanche Treasury Co. (AVAT) bought its AVAX from the Foundation in October 2025 at a 60 percent discount. On 23 September 2026 it sold 2.08 million locked AVAX back to the Foundation for 15 million dollars, about 7.20 dollars each, while AVAX closed at 10.26 dollars that day. It is using the money to pay down debt. On the stock market AVAT is worth only about 0.6 times its AVAX, AVAX One (AVX) about 0.3 times.

Holding AVAX means being diluted by issuance and barely compensated by the C-Chain.

How that shows in the price is answered by the comparison with bitcoin.

The price against bitcoin

Binance monthly candles AVAXUSDT and BTCUSDT, closing prices, through 30 Sep 2026.

Since the start of 2023 AVAX has earned nothing, while bitcoin has grown fivefold. When bitcoin falls, AVAX almost always falls with it, usually harder. When bitcoin rises, AVAX keeps up in only one month out of three. September was one of those rare months, and it only just lifted the balance since 2023 from minus 34 percent to zero.

Since 2023 AVAX has been a bet that rarely joins bitcoin's good months and almost always joins its bad ones.

What was announced and what can be checked

The verified entries sit on the C-Chain. The big numbers hang on L1s or on platforms that do not need AVAX.

What this means for you

If you hold AVAX or want to buy it:

• Don't buy the headline about the next chain. Since Etna, a new L1 pays the network about 16 AVAX per validator per year. A FIFA, bank or game chain changes almost nothing about demand for AVAX.
• Watch the C-Chain fees. They are the one place where usage burns AVAX. Today that is about 1,660 AVAX a day. Only when it gets close to new issuance, about 28,000 AVAX a day, does usage carry the token.
• Expect dilution. About 10.2 million new AVAX a year from staking, plus 6.7 million from Foundation releases. If you don't stake, you lose share every year.

What is not claimed here: that Avalanche is technically weak or that the price has to fall. The technology works, and the real money on the C-Chain is growing. The open question is whether the new fee level after 27 September holds once the bot load is gone and only paid usage remains. That will show from the end of October, when 30 days of data are in.

FAQ

What is the difference between a subnet and an L1? Same principle: a chain with its own validators. Since Etna they are called L1s, and their validators no longer have to validate the main network or stake 2,000 AVAX.

Don't the L1s pay transaction fees in AVAX? Mostly not. An L1 sets its own fee token, for games often their own game token. Only the validator pays AVAX, for its seat on the P-Chain.

Why do Avalanche's transaction numbers differ so much between sites? Because the bot load disappeared at the end of September, and many sites show 30-day averages that mix the old and new periods. Here both are shown separately.

Is there an Avalanche ETF? Yes, three in the US: VanEck VAVX since January 2026, Grayscale GAVA since March 2026, plus Bitwise BAVA. Together they hold about 5 million AVAX, a little over 1 percent of circulating supply.

All AVAX monthly returns in US dollars

Binance AVAXUSDT, monthly return close to close in percent; September 2020 from trading start on 22 Sep.

Not investment advice, not a recommendation, not a forecast — historical patterns are no promise.

Study the Past — Improve your Future 🥋