Gas Fees Explained: What They Are and How to Reduce Them

Most of what is written about gas fees was true in 2021 and has been quietly wrong ever since. Articles still warn about $50 Ethereum transfers and recommend Tron as the cheap alternative. As of September 2026, an Ethereum token transfer costs around six to fifteen cents and Tron costs over two dollars – the exact reverse.

This guide explains what gas fees actually are, why they move, what they cost right now across the major networks, and the practical ways to reduce them.

What Gas Fees Actually Pay For?

A gas fee is what you pay the network to include and execute your transaction. It is not a service charge collected by a company – it goes to the validators or miners who process and secure the chain.

The term comes from Ethereum, where every operation has a fixed cost in units of “gas.” Sending ETH costs 21,000 gas. Transferring a token costs more, typically 45,000 to 65,000, because it executes a smart contract rather than a simple value transfer. Complex operations still cost more.

Your fee is the product of two numbers:

Gas used × gas price = total fee

Gas used is determined by what your transaction does and does not change. Gas price is set by market demand and changes constantly – this is the part that makes fees unpredictable.

Gas serves two purposes beyond paying validators. It prevents spam, because flooding the network would cost the attacker real money. And it allocates scarce block space to whoever values it most at that moment.

• Gas pays validators for computation and security, not a company for a service
• Every operation type has a fixed gas cost
• Token transfers cost more gas than native transfers
• The variable part is gas price, set by demand
• Fees also prevent spam and ration limited block space

Why Do Fees Change Minute to Minute?

Networks handle this in fundamentally different ways, which is why “gas fees” means something different depending on the chain.

Auction-based pricing (Ethereum, Bitcoin). Block space is limited and users bid for it. When demand rises – a popular token launch, market volatility, heavy trading – the clearing price rises for everyone. When demand falls, fees collapse. Ethereum’s fee mechanism includes a base fee that adjusts automatically based on how full recent blocks were, plus an optional priority tip.

Near-fixed pricing (BNB Smart Chain, Solana). These networks set fees that barely respond to demand. Predictability is the design goal, and for anyone budgeting payment costs it is worth as much as a low headline number.

Resource-based pricing (Tron). Tron charges for bandwidth and energy rather than a simple gas price. It produces one quirk that catches merchants out: sending a token to an address that has never held it costs roughly double, because the network charges for creating the token account.

A practical consequence of auction pricing: your transaction is never rejected for paying too little. It simply waits. A transaction priced below the current market rate sits in the mempool until demand drops – sometimes minutes, sometimes days.

Gas Costs by Network (September 2026)

Here are current figures, measured rather than repeated.

Two conclusions from that table overturn the standard advice.

Ethereum is cheap now. Sustained low gas conditions – base fees under half a gwei – mean an ERC-20 transfer costs cents. The Dencun upgrade and subsequent improvements cut costs by roughly 95% from earlier peaks, and the network has stayed there. Advice to avoid Ethereum on cost grounds is several years out of date – if you want to accept Ethereum payments or ERC-20 stablecoins, fees are no longer the obstacle.

Tron is now the most expensive way to move stablecoins. At roughly $2.17 per USDT transfer, TRC-20 costs about thirty times more than ERC-20. Tron built its reputation on cheap transfers and retains enormous usage because exchanges default to it – but the cost advantage is gone.

This matters commercially. A merchant steering customers to Tron “because it’s cheaper” is now steering them to the most expensive option available.

Native Transfers vs Token Transfers

A distinction that explains a lot of confusion about why two transfers on the same network cost different amounts.

A native transfer moves the chain’s own asset – ETH on Ethereum, BNB on BNB Chain, SOL on Solana. It is a simple balance update and uses minimal gas. On Ethereum, 21,000 gas.

A token transfer moves an asset that exists as a smart contract on top of the chain – USDT, USDC and everything else. It executes contract code, which costs more gas. On Ethereum, typically 45,000–65,000, so roughly two to three times the cost of a native send.

On Tron, the gap is far wider. A TRX transfer costs well under a cent, while a TRC-20 USDT transfer costs around $2.17. The token layer, not the base network, is what is expensive.

This has a direct practical implication: you need the native asset to move tokens. A wallet holding only USDT cannot send USDT – it needs ETH, TRX, BNB or SOL for the fee. This is one of the most common reasons a transfer fails, and it affects anyone receiving payments who later wants to move them.

How to Reduce Gas Fees?

Seven approaches, roughly in order of effectiveness.

1. Choose the right network. By far the largest lever. Moving a payment from Tron to Solana or BNB Chain cuts the cost by a factor of hundreds. Nothing else comes close.
2. Send tokens rather than converting. Every swap or bridge is an additional transaction with its own fee. Receiving and holding the same asset avoids them.
3. Time it on auction-priced chains. Ethereum and Bitcoin fees fall during quiet periods. For non-urgent transfers, waiting for lower demand genuinely works. Gas trackers show current rates.
4. Batch outgoing payments. Instead of fifty individual transfers, use a batching tool. Fewer transactions, fewer fees, fewer manual steps.
5. Pre-fund destination addresses on Tron. Since a first-time token transfer costs double, sending once to an address removes that premium from every later payment.
6. Set fees deliberately rather than accepting wallet defaults. Many wallets pad their estimates. Where the wallet allows a custom fee, a slightly lower setting often confirms almost as fast.
7. Use layer-2 networks for high-frequency small payments. Rollups on Ethereum and the Bitcoin Lightning Network reduce per-transaction cost dramatically, at the cost of additional setup.

The ranking matters. Merchants often spend effort on fee-timing tactics that save fractions of a cent while continuing to route payments through the most expensive network available. Fix the network choice first; everything else is optimisation around the margins.

The network cost comparison has a fuller breakdown by payment size, and the TRC-20 vs ERC-20 vs BEP-20 comparison covers the stablecoin-specific trade-offs.

Who Pays Gas in a Merchant Payment?

A point that causes budgeting errors in both directions.

In a standard direct-to-wallet payment, the customer pays the gas fee when they send the transaction. It does not come out of your revenue. You receive the invoiced amount.

You pay gas in three situations:

• Refunds – a new outbound transaction you initiate
• Moving your own funds – sweeping to cold storage, or converting
• Payouts – paying staff, contractors or suppliers

So gas is not a merchant cost on the receiving side. But it is absolutely a conversion factor. A customer facing a $4 network fee on a $30 order frequently abandons. That is not a line on your P&L, but it is a real loss.

The practical response is to present the cheapest network first at checkout. Simply reordering the options shifts a meaningful share of volume onto low-cost rails without any customer education.

This is also where the gateway’s own fee sits, separately from gas. The two should not be conflated – the gateway fee comparison covers the service-fee layer. Bcon Global charges a flat 1% with no withdrawal fee and no conversion spread, because payments settle directly from customer to merchant wallet – meaning the only network fee in the flow is the one the customer pays to send.

How a Gas Fee Is Calculated (Worked Example)

Seeing the arithmetic once removes most of the mystery, and lets you sanity-check any figure you are quoted.

The formula is gas units × gas price × asset price.

Example: an ERC-20 USDT transfer on Ethereum.

1. Gas units for a token transfer: approximately 63,000
2. Gas price at the time of writing: approximately 0.4 gwei (a gwei is one billionth of an ETH)
3. Fee in ETH: 63,000 × 0.0000000004 = 0.0000252 ETH
4. ETH at approximately $2,468
5. Total: about $0.062

The same calculation for a native ETH send:

1. Gas units: 21,000
2. 21,000 × 0.0000000004 = 0.0000084 ETH
3. Total: about $0.021

The difference between the two is entirely the gas units – the token transfer executes contract code, the native send does not.

Why estimates vary. Wallets pad their gas price estimates to ensure fast confirmation, sometimes substantially. The gas units are predictable; the gas price is the wallet’s guess about what the market requires right now. Where your wallet allows a custom setting, a slightly lower gas price often confirms almost as quickly.

Run this calculation yourself when a figure looks wrong. Most outdated fee claims fall apart immediately once you multiply current gas price by the actual gas units rather than repeating a number from an old article.

Gas fees are the most misunderstood number in crypto payments, largely because the widely repeated figures stopped being accurate several years ago and nobody updated them.

The current reality is straightforward: Solana and BNB Smart Chain cost fractions of a cent, Ethereum costs cents, Bitcoin is cheap while the mempool is quiet, and Tron – long recommended as the budget option – is now the most expensive way to move a stablecoin by a wide margin.

Check live figures rather than inheriting assumptions, choose the network before optimising anything else, and remember that on the receiving side gas is your customer’s cost and therefore your conversion problem rather than your expense.