Who really used Avalanche in 2026? A wallet farm, gas burners and bots, until a validator vote

01 · The busy yearAvalanche's busiest year, on paper

On paper, this is Avalanche's busiest year yet. By the end of September it had carried more transactions than in all of last year. The research firm Nansen counted about 621,000 active addresses per day from April to June.

Avalanche is a blockchain that went live in 2020. Most of its apps and tokens sit on one part of it, the C-Chain, which works like Ethereum. Each transaction pays a small fee in the chain's own coin, AVAX, which traded near $11 at the end of September.

Bitquery indexes every C-Chain block, so we sorted that traffic by who sent it and what it did. Most of the boom came from a few programs. From April to June, 95% of the senders on a typical day were wallets working one contract, a program that lives on the chain. Two wallets that produced nothing anyone could see used 12% of all the gas, the network's measure of work, over the past 21 months, and paid about $7,000 to do it.

Then a new rule let the validators, the computers that run the network, raise the lowest price a transaction can pay. In the last days of September the higher price took effect, and most of that traffic stopped. People kept sending.

02 · The wallet farmOne contract, 1.19 million wallets

The address count has a short explanation. Over 4 weeks around the new year, one address sent each of 1.19 million new wallets a sliver of AVAX worth less than a cent. On average, that was more than each wallet would go on to pay in fees. Then the wallets went to work.

From early January, those wallets called the same contract again and again. Each sent about 90 transactions on average, spread over months, so each one looked like an ordinary user who came back. Together they made about 110 million calls before stopping in late August.

The function the wallets called is named setBetCount. Its code is not published, so we cannot say what it counts or who runs it. The address that funded the wallets got its own first 5 AVAX from a single wallet the day before it started. All 605 AVAX it ever received came from that one wallet.

In the second quarter these wallets were 95% of the addresses sending a transaction on a typical day, so everyone else came to about one address in twenty. Fresh wallets made in bulk are a familiar sight in airdrop farming. The funding trail here has that shape, with one sender, one small amount and more than a million receivers. The whole thing cost about $10,000.

03 · The roomHow Avalanche prices its space

Every transaction uses gas, a unit for the work the network does to process it. Avalanche sets a gas target, the amount of gas per second it aims to use. When demand runs above the target the price of gas rises, and when it runs below the price falls. Validators vote on the target. The network can sustain twice the target, so a chain that sits on its target is one where the price is doing the rationing.

For 535 days in a row, until 25 September, the C-Chain used its target every day. On days when the target stayed put, use never strayed more than 0.6% from it, and validators raised the target on 22 different days along the way, each rise used up within a day.

On Bitcoin we found blocks filling up with wallets that pay themselves to claim a free token. The question on Avalanche is the same. When a chain always uses all the room it is given, who is doing the filling?

04 · Who filled itFarming, burning and trading

We sorted each month's 50 biggest users of gas by what their contracts do. Trading took the biggest share. Token farming, apps that hand out a token to whoever sends enough transactions or burns enough gas, took the next. Two wallets that did nothing visible took 12% of it.

The two biggest farming apps on Avalanche used about 8% of all the gas each. XPower is a token minted by proof of work. A miner searches for a lucky number off the chain and sends it in to mint new tokens. XEN is a token anyone can claim by paying the gas, and the busiest users of it are tools that claim it for hundreds of wallets at once. In July, farming took 44% of the month's gas.

In the second week of September, a few hundred wallets were sending XPower's main contract hundreds of thousands of transactions per day. Of the 101 still at it on the 11th, all but one got their first AVAX from the same address. On the 12th the mining stopped. Transactions on the whole chain halved that week, yet gas use stayed at the target, because other users took the room.

The strangest users were the two gas burners. Each called its own contract, with no inputs, millions of gas at a time. Every one of their 3.2 million transactions failed with an "invalid opcode" error, an instruction that uses up all the gas the sender offered. So the transactions moved no tokens and no AVAX, and the contracts never recorded a single event. The first ran from February to April. The second ran from March to 25 September, and its wallet got its first AVAX from the first one's wallet. In April the second burner used 30% of all the gas on the chain. The two paid about $7,000 in fees between them. We do not know what they were for.

05 · The peopleThe people who come back

An address is not a person, and one program can control a million of them. So we counted the wallets that came back, those active on 5 or more days in a month. We call them regulars. Then we took out the farm's wallets and every wallet that sends like a bot, by three tests that range from 1,000 sends in one day down to 50 on a typical day. We used the same idea when we asked whether Algorand is dead.

In January about 33,000 regulars passed all the tests. By August the count was the lowest in our data, which goes back to 2023, and a quarter of the 2023 high. That month bots sent nine in ten of everything sent on the C-Chain. Avalanche is not alone in this. On Solana, most of the DEX trading we index was not real by our tests.

06 · The voteValidators set a price floor

On 22 September, Avalanche switched on an upgrade called Helicon. Among other changes, it let validators vote on the minimum gas price, each one setting the floor it would like. Gas prices are quoted in gwei, a billionth of an AVAX. Until Helicon the minimum was 1 wei, a billionth of a gwei, which is as close to free as the rules allow.

The proposal behind the change, ACP-283, gives its reason plainly. At the lowest prices seen in practice, about 0.0001 gwei, it says, "spam protocols like XEN remain profitable".

Each new block nudges the floor up, down or not at all, as the validator that made it votes. From the first day, a third or more of the blocks pushed the floor up as fast as the rules allow. None of the blocks we checked pushed it down. The rest left it alone.

For 3.5 days the floor sat below the market price and changed nothing. Then it bit. Early on the 26th it reached the going price, and from then on the floor was the price. It stopped at 5 gwei the next morning, about 29 times the median price paid in the two weeks before.

07 · What changedBots cut back, people kept sending

In the week after the floor, gas use fell to under 40% of the target, and the number of sends fell by more than half.

The fall was almost all bots. People, meaning every wallet that is neither a bot nor a farm wallet, sent about as many per day after the floor as before. Bots sent less than half as many. The second gas burner stopped the evening before the floor took hold, and a batch sender that moved XPower tokens stopped the next morning. XEN farming all but vanished.

People pay more now, and still very little. Their median fee rose about 9 times, to less than half a cent. Avalanche burns every fee it collects. On a median day the AVAX it burned rose 87%, and 92% in dollars as the AVAX price rose. The bots that stayed are the ones that can afford it. One trading bot paid about $25,000 in fees in its first 8 days under the floor. When Robinhood Chain's gas price went up 25 times, 31 wallets with one funder ended up paying 15% of everything the chain earned.

08 · So who used Avalanche?Mostly software, and a steady few people

Software, mostly. In Avalanche's busiest year one farm of 1.19 million wallets made the address count, and farming and burning used a third of the gas. By August, bots sent nine in ten transactions, and the wallets that look like people coming back were a quarter of their 2023 high.

The floor did what ACP-283 set out to do. Bots sent less than half as many transactions after it, and people kept sending at the same pace for less than half a cent each. The chain is quieter now, and its numbers are closer to the people using it.

If you want to know how many people use a chain, the active address count will not tell you. Count the wallets that come back, and leave out the ones a single funder created.

09 · How we measured itMethod

We read every C-Chain transaction from January 2023 to early October 2026, and monthly totals back to 2021 for the long view. A sender is the wallet that signed a transaction. A regular is a wallet active on 5 or more days in a month. A bot, by our loosest test, sent 50 or more transactions on a typical active day, or 1,000 on any day. The farm's wallets are those that sent at least half their transactions to its contract, which in practice means the wallets the funding address paid. The gas target and the price floor come straight from block headers, which we read from Avalanche's public node. Dollar figures use the daily AVAX price.

The transactions, contract calls and blocks docs describe the fields we used. We asked the same question of Optimism and Dogecoin.

10 · The recordAddresses behind this story

FAQ

How many people use Avalanche?

Nobody can count people, only wallets. In August 2026, about 21,500 wallets were active on 5 or more days and were neither bots nor part of one large wallet farm. In the second quarter, 95% of the addresses active on a typical day belonged to that farm.

Why did Avalanche's activity drop in September 2026?

Two things happened. On 12 September a group of XPower mining wallets stopped, and daily transactions halved. From 26 September a price floor set by validator vote, under the Helicon upgrade and ACP-283, rose above the market price and reached 5 gwei. Gas use fell by more than two thirds, and bots sent less than half as much.

What is ACP-283?

It is an Avalanche proposal, switched on with the Helicon upgrade on 22 September 2026, that lets each validator vote for the C-Chain's price floor. The floor moves toward what the validators making blocks want. It started at 1 wei and reached 5 gwei by 27 September.

How much does an Avalanche transaction cost now?

After the floor reached 5 gwei, the median fee people paid was less than half a cent, about 9 times what it was in the two weeks before.

Is AVAX dead?

No. People kept sending about as much each day after the floor as before it. What fell away was software, such as a wallet farm, gas burners, token mining and bots.