Gas Fees: the cost of using a Blockchain

Every transaction costs something. Not because anyone is greedy. Because computation on a decentralized network isn’t free, and never will be.

This is Article #9 in the Web3 with Axl series. ⬅️ Previous Articles: #1 | #2 | #3 | #4 | #5 | #6 | #7 | #8 |

Leia em português: [🇧🇷]

In May 2022, Yuga Labs (the company behind Bored Ape Yacht Club) launched a virtual land sale called Otherside. The demand was so massive that the Ethereum network became a war zone. Users competed for block space by offering higher and higher gas fees. In a single night, over $170 million was spent on gas alone. Some people paid hundreds of dollars in fees for transactions that failed and never completed. The gas was consumed anyway.

That night, thousands of people learned a painful lesson: on a blockchain, space is limited, and the price of using it is set by everyone who wants it at the same time.

If you’ve been following this series, you’ve seen gas mentioned in nearly every article since #5. Ethereum charges gas. Smart contracts consume gas. Token deploys cost gas. Stablecoins moved to Tron because of gas. This is the article where we finally explain what gas actually is, why it costs what it costs, and what’s being done to make it bearable.

Pay attention to one concept as you read: scarcity. Gas isn’t expensive because someone chose a high price. It’s expensive because block space is finite, and when demand exceeds supply, price is how the network decides who goes first.

What gas actually is

Gas is not a token. It’s not a fee set by a company. It’s a unit of measurement.

Every operation on Ethereum costs a specific amount of computational effort. Sending ETH from one wallet to another: 21,000 gas units. Approving a smart contract to spend your tokens: roughly 46,000 gas units. Swapping tokens on Uniswap: around 150,000 gas units. Deploying a new smart contract: potentially millions.

The gas itself doesn’t have a price. What has a price is each unit of gas. That price is measured in gwei, a tiny denomination of ETH (1 gwei = 0.000000001 ETH). If the gas price is 30 gwei and your transaction uses 21,000 gas units, you pay 21,000 × 30 = 630,000 gwei, which is 0.00063 ETH. At an ETH price of $3,800, that’s about $2.39.

Gas exists for two reasons: First, it pays the validators who process transactions. Without compensation, nobody would run the hardware that keeps the network alive; Second, it prevents abuse. If transactions were free, anyone could flood the network with infinite loops or spam, grinding everything to a halt. Gas makes every operation cost something, which means attackers would need to spend real money to disrupt the system.

Think of it as a toll road. The road is public. Anyone can use it. But every car pays a toll. The toll keeps the road maintained and prevents people from parking across all lanes just because they can.

Every transaction pays a toll. It keeps the network running and spam out.

Gas in one sentence: it’s the cost of computation, measured in units, priced by demand, and paid to keep the network secure.

How the price is set

Before August 2021, Ethereum gas pricing was a pure auction. Users chose how much to pay per unit of gas. Validators picked the highest bids first. Everyone else waited. The problem: nobody knew what to bid. Overpaying was common. Underpaying meant your transaction sat in limbo for hours. The experience was terrible.

EIP-1559 changed everything… Instead of a blind auction, the network now calculates a base fee automatically. The rule is simple: if the previous block was more than 50% full, the base fee goes up. If it was less than 50% full, the base fee goes down. The maximum adjustment is 12.5% per block, so the price can’t spike instantly but it tracks demand in near-real-time.

On top of the base fee, users can add a priority fee (also called a tip). This goes directly to the validator and incentivizes faster inclusion. During normal conditions, a small tip is enough. During high demand, larger tips push your transaction ahead of others.

Here’s the detail that connects gas to the economics of ETH itself: the base fee is burned. It doesn’t go to validators. It’s permanently destroyed. Every transaction on Ethereum removes a small amount of ETH from existence. When network activity is high enough, more ETH is burned than created through validator rewards, making the total supply of ETH shrink. Since EIP-1559 launched, over 4.6 million ETH have been burned.

This means gas fees do double duty: they regulate access to block space and they control the monetary supply of ETH. Every swap, every mint, every transfer slightly reduces the number of ETH that will ever exist.

Block space is limited. When everyone wants in at the same time, the price decides who goes first.

Gas pricing in one sentence: the network sets a base price that adjusts with demand, burns it to reduce supply, and lets users tip validators for priority.

When gas becomes a war

Most of the time, gas on Ethereum is manageable. A few dollars for a swap. Less than a dollar for a simple transfer. But during moments of extreme demand, the system reveals its constraint: block space is fixed, and when everyone wants it at once, prices spike.

December 2017: CryptoKitties, a game about breeding digital cats, became so popular that it congested the entire Ethereum network. Pending transactions piled up. Gas prices soared. It was the first time most users realized that Ethereum has a capacity problem.

Summer 2020: DeFi exploded. Yield farming, liquidity mining, and protocol launches drove average gas prices above 200 gwei for weeks. A simple token swap could cost $50 to $100. Using DeFi became a rich person’s game.

May 2022: the Otherside mint. Yuga Labs’ virtual land sale generated so much demand that gas prices hit over 8,000 gwei. Transactions that failed still consumed gas. One user reportedly spent $44,000 in gas fees across multiple failed transactions, getting nothing in return.

These events exposed a hard truth: Ethereum’s security comes at the cost of capacity. The network processes roughly 15 to 30 transactions per second. When demand exceeds that, the only mechanism is price. And price doesn’t care whether you’re a whale moving millions or someone trying to send $20 to a friend.

The most frustrating part for new users: failed transactions still cost gas. When a transaction reverts (because slippage was too high, the NFT sold out, or the contract conditions weren’t met), the network still consumed computational resources trying to execute it. You pay for the attempt, not the result.

Not all blockchains cost the same

Ethereum’s high gas fees aren’t a universal blockchain problem, they’re a specific consequence of Ethereum’s design choices: maximum decentralization and security, with limited throughput. Other blockchains make different tradeoffs.

Solana processes roughly 4,000 transactions per second and charges an average of $0.003 per transaction. It achieves this through a different consensus mechanism (Proof of History combined with Proof of Stake) and a more centralized validator set. The tradeoff: Solana has experienced multiple full network outages, something that has never happened to Ethereum.

Tron charges near-zero fees for USDT transfers, which is why it dominates stablecoin volume in emerging markets. A USDT transfer on Tron costs fractions of a cent and the same transfer on Ethereum L1 can cost several dollars. For someone in Nigeria sending $50 home, that difference isn’t technical, it’s the difference between usable and not.

BNB Chain, Polygon, Avalanche: all offer lower fees than Ethereum with varying degrees of decentralization. The pattern is consistent: cheaper usually means more centralized, which means more trust in fewer validators.

The blockchain trilemma (decentralization, security, scalability: pick two) isn’t theoretical. It’s the reason your transaction costs what it costs.

Same destination. Three doors, three different prices. Every blockchain makes a tradeoff.

The cost difference in one sentence: cheap gas usually means fewer validators, which means more trust, which means less of what made blockchain revolutionary in the first place.

The solution that’s already here

Ethereum didn’t try to make its main network cheaper. Instead, it became the security layer for a new generation of networks built on top of it (these are called Layer 2s).

The concept: L2s process thousands of transactions on their own network, then publish a compressed summary (or cryptographic proof) back to Ethereum. The L1 verifies the proof. The security comes from Ethereum. The speed and low cost come from the L2. Users get both.

In March 2024, the Dencun upgrade introduced a feature called blobs (EIP-4844), which created a dedicated, cheaper data channel for L2s to publish their summaries. The impact was dramatic. Transaction costs on Arbitrum dropped from approximately $0.37 to under $0.01. Base, Optimism, and other L2s saw similar reductions.

Before Dencun, using DeFi on Ethereum was prohibitive for most people. After Dencun, swapping tokens on Base or Arbitrum costs a few cents. The technology that was supposed to bank the unbanked finally became cheap enough to actually do it.

The two main types of L2s: Optimistic Rollups (Arbitrum, Optimism, Base) assume transactions are valid unless challenged within a dispute window. ZK-Rollups (zkSync, StarkNet, Scroll) use mathematical proofs to verify every batch of transactions. Both achieve the same goal: Ethereum-level security at a fraction of the cost. Article #10 will go deeper into how they work.

The future: gas you don’t even notice

The next evolution isn’t cheaper gas. It’s invisible gas.

Account abstraction (ERC-4337) allows smart contract wallets to define custom rules for transaction execution. One of the most powerful: a third party can pay your gas. This concept, called a paymaster, means applications can sponsor their users’ transactions. You swap tokens, mint an NFT, or cast a vote, and the app pays the fee. You don’t need ETH in your wallet. You don’t even need to know gas exists.

Several applications on Base and other L2s already implement this in 2026. New users create a wallet, interact with DeFi, and never see a gas prompt. The friction that made Web3 unusable for non-technical people is disappearing.

The broader trend is called chain abstraction: the idea that users shouldn’t need to know which blockchain they’re on, which token pays for gas, or how many confirmations their transaction needs. They just click a button and the system handles the rest. We’re not fully there yet, but the direction is clear.

There’s an ironic tension in this progress: as L2s get cheaper and gas becomes invisible, less ETH gets burned on the mainnet. The same mechanism that made ETH deflationary depends on gas being expensive enough to generate significant burn. Ethereum’s challenge going forward is balancing user accessibility with the economic model that gives ETH its value.

What comes next

Now you understand what gas is, why it costs what it costs, and how the ecosystem is solving it. You’ve seen the mechanism (EIP-1559), the wars (Otherside), the alternatives (Solana, Tron, L2s), and the future (account abstraction).

But the L2 solution raises a new question: if these networks process transactions outside of Ethereum and submit summaries back, how does that actually work? What are rollups? What are blobs? And what does it mean for a blockchain to scale without sacrificing what makes it decentralized?

That’s the next article. Layer 2s.

→ Web3 with Axl · Next: Layer 2s and Other Blockchains: The Scaling Problem.

Follow AxlMint ➡️ we explain Web3 the way it should have been explained from the start.