Chains & Infrastructure
The Cost of Sending a Stablecoin, Chain by Chain
Across eight major chains, the median network cost of a simple stablecoin transfer at 11 to 12 September 2026 conditions ran from five ten-thousandths of a dollar to about half a cent. Tron's documented ceiling sits far higher. Every figure names its method, provider, and date.
| Fact | Value | As of | Source |
|---|---|---|---|
| Ethereum, USDC, conditions fee | 0.0056 then 0.0051 dollars across two same-day sessions, 0.0547 to 0.0507 gwei median base times 40,348 gas | C-07 | |
| Ethereum, USDC, observed median | 0.21 dollars over 40 sampled transfers, batch-inclusive, exchange 2-gwei pricers dominate the sample | C-07 | |
| Tron, USDT TRC-20, documented rate ceiling | 2.18 dollars full Energy burn at the docs rate, a ceiling, not an observed median; no conditions leg exists | C-13 | |
| Tron, USDT TRC-20, observed median | 0.117 dollars over 28 sampled transfers, 8 of which paid zero through staked or deployer Energy | C-15 | |
| Solana, USDC, conditions fee | 0.00051 dollars in both sessions, the fixed 5,000-lamport base fee from the docs | C-19 | |
| Solana, USDC, observed median | 0.00053 dollars over 19 transfers, priority median of 50 lamports | C-23 | |
| BSC, USDT, conditions fee | 0.0025 then 0.0013 dollars from the gasPrice suggestion of 0.1 to 0.05 gwei times 34,515 gas, a named stand-in because feeHistory is unpopulated on public nodes | C-26 | |
| BSC, USDT, observed median | 0.0062 dollars over 24 sampled transfers, batch-inclusive | C-26 | |
| Base, USDC, conditions fee | 0.0006 dollars in both sessions, base fee pinned at the documented 5 M wei minimum | C-32 | |
| Base, USDC, observed median | 0.0046 dollars over 30 transfers; the simple subset read 0.0007 dollars over 6; the L1 data fee is not exposed in receipts | C-33 | |
| Arbitrum, USDC, conditions fee | 0.0020 dollars in both sessions, 0.020 gwei times 40,451 gas plus 180 median L1 gas, the L1 component about 0.1 percent | C-35 | |
| Arbitrum, USDC, observed median | 0.0068 dollars over 19 transfers, L1 component quantified | C-36 | |
| Polygon, USDC, conditions fee | 0.00135 then 0.0013 dollars, 247.2 to 239.3 gwei times 56,231 gas, with the issuer Gas Station V2 reading 246.47 gwei in three-way agreement | C-39 | |
| Polygon, USDC, observed median | 0.0080 dollars over 50 transfers; a 90-transfer window read 0.0043 dollars, batch-share variance labeled | C-39 | |
| Hyperliquid, USDC, conditions fee | 0.00056 then 0.0005 dollars, the 0.10 gwei floor times 67,757 plateau gas, a documented deviation above the under-60k filter | C-43 | |
| Hyperliquid, USDC, observed median | 0.0058 dollars over 50 sampled transfers | C-44 | |
| Token prices used for USD conversion | ETH 2,515.31 dollars, TRX 0.338547, SOL 102.40, BNB 726.62, POL 0.097291, HYPE 79.41, CoinGecko same-session pull | C-11 |
What this table measures
One metric, stated precisely. The median network cost of moving a native major stablecoin at median network conditions on the pull date. USDC is the reference token everywhere except two chains. On Tron the reference is USDT TRC-20. On BSC it is the chain-dominant USDT, under the issuer-doc caveat the BSC page records.
Every figure in the table carries its method, its providers, and its date. The pulls ran 11 to 12 September 2026, two sessions about two hours apart, from public RPC endpoints with second-provider checks throughout [2] [8] [12]. Dollar conversion used a same-session price pull [3].
Two numbers per chain
The conditions fee is the reproducible figure. It multiplies the feeHistory median base fee by the simple-transfer gas amount, the median gas of sampled transfers under 60,000 gas [2]. A reader with an RPC endpoint can re-derive it. That is why it is the comparison row.
The observed median is what sampled transfers actually paid. It includes batched transactions, exchange pricers, and priority behavior. It varies up to about 2 times across sampling windows [15]. It is labeled and dated, and it is never the comparison standard, because comparing tip-inclusive medians across chains compares mixed things.
The named method notes
Four chains need notes, and the notes are part of the data.
Tron prices transfers in Energy, not gas. The documented full-burn rate would cost 2.18 dollars [4]. That is a ceiling. The observed median was 0.117 dollars, and 8 of 28 sampled transfers paid zero through staked or deployer Energy [5]. A second provider read the identical rate [6]. Whether a given transfer pays little or nothing depends on Energy holdings the table cannot see.
BSC public nodes no longer populate feeHistory after the Fermi changes. The conditions row uses the gasPrice suggestion as a named stand-in, stated here rather than silently dropped [9]. The chain’s page carries the same note.
Base receipts do not expose the L1 data fee, so the observed median understates what a full transfer costs [11]. Arbitrum receipts do expose it, and it ran about 0.1 percent of the total [12]. The two rollups are not symmetric in what their receipts show.
Hyperliquid’s simple transfers sit at a 67,757 gas plateau, above the under-60k filter used elsewhere. The plateau repeated across both providers, and the deviation is documented rather than hidden [16] [17].
Same-day variance, stated honestly
The conditions metric moved within 7.3 percent across the two same-day sessions. Ethereum moved 7 percent down. Base, Arbitrum, and Hyperliquid sat at protocol floors. Polygon moved 3 percent, and its issuer-run Gas Station V2 oracle agreed with both node pulls on the base fee [14]. Solana’s base fee is protocol-fixed. BSC’s move reflects node-suggestion variance, not a chain change. Token prices moved under 1 percent [3].
Both sessions were the same day. Day-over-day drift is not bounded by this pull, and the page does not pretend otherwise. The freshness rule: fees are re-pulled at build, never stale-quoted. The gas definition explains the base fee machinery behind every row [1].
What this register refuses
No fee figure appears without its source and date. Entrant chains whose fee fields are not yet documented get no row here, never a guess. And the register states costs. It does not say which chain anyone should use. That decision belongs to the reader, and the chain pages carry each network’s full facts.
How much does it cost to send USDC?
At the 11 to 12 September 2026 pulls, the conditions fee ranged from 0.00051 dollars on Solana to 0.0056 dollars on Ethereum, with each chain’s row in the table above. Observed medians ran higher on every chain, from 0.00053 dollars to 0.21 dollars. These are dated observations of network cost, not standing prices. They move.
What is the cheapest way to send USDC?
This site does not hand out a cheapest. The table shows each chain’s dated conditions fee and observed median side by side, with the caveats stated. Solana’s row reads lowest at this pull. It also carries a fixed base fee and priority mechanics that differ from the EVM chains. Lowest at one pull on one metric is not a verdict.
Which network is best for sending USDC?
No verdict. Best depends on what the sender holds, where the recipient is, what finality they need, and what the day’s fees are. The chains hub and the per-chain pages carry those facts, each dated. This register supplies one input, the cost, on stated conditions.
If I send a large USDC amount on Solana, how much SOL do I need?
Transfer fees on Solana are paid in SOL. At the pull-date levels the docs fixed a 5,000-lamport base fee and the sampled transfers paid a median 0.00053 dollars total [7] [8]. The fee is charged per transaction, not per dollar moved. What an account also needs for rent and account state is outside this register, which does not state it.
Where this page sits
The chains hub frames the networks. The Ethereum and Solana pages carry those chains’ full facts. The gas term defines the unit behind every row. The layer 2 term explains why two chains in this table inherit their security from Ethereum. Every figure above carries its source and date.
Last verified