On 15 November 2026 at 09:46 UTC, depositing ETH into Arbitrum through its official bridge would have cost about four cents of Ethereum gas, by our own reading of the network. Getting the same ETH back out through that bridge takes about a week. And the most expensive mistake a newcomer can make costs no gas at all: clicking a search ad that opens a copy of a bridge website. This guide shows how to bridge to a layer 2 in the order that keeps the first two numbers predictable and the third risk out of reach.
It is a hands-on lesson. You will move a little ETH and a little USDC from Ethereum to Arbitrum One, Base or OP Mainnet: first on the Sepolia test network, where nothing has value, then on mainnet with a few dollars.
Two earlier lessons did the theory. Our guide to layer 2 networks explains what a rollup is, why the official bridge contract on Ethereum holds the money and why you need a little ETH on the layer 2 to pay its fees. The previous lesson of this path, modular blockchains and app-chains explained, showed why Ethereum now has dozens of these networks, each with its own bridge. Here we use them.
Three ways to bridge to a layer 2, and what each one asks of you
There are three ways to get funds onto a layer 2. None is always the best: the right one depends mostly on where your money sits today and how soon you may need it back.
| Route | Time in | Time out | Who you rely on | Cost (15 Nov 2026 readings) |
|---|---|---|---|---|
| Exchange withdrawal straight to the L2 | minutes, set by the exchange | deposit back to the exchange on the same L2 | the exchange | the exchange’s withdrawal fee |
| Canonical (official) bridge | about 1 to 30 minutes | about a week on Arbitrum, 7 days on OP Mainnet, 5 days on Base | the rollup itself | Ethereum gas: about $0.04 to $0.27 for an ETH deposit |
| Third-party or intent bridge | seconds to minutes | seconds to minutes | the bridge operator’s contracts and capital | a bridge fee plus gas |
Route 1: withdraw from an exchange straight to the layer 2
If your ETH or USDC sits on a centralised exchange, you may not need a bridge at all. Many large exchanges list Arbitrum, Base or OP Mainnet as withdrawal networks for ETH and USDC; the list differs from one exchange to another and changes over time, so check yours. Base’s own documentation says that from a Coinbase account you can withdraw directly to Base, with no bridge needed.
The screen is the usual withdrawal form with one extra field: the network. Your wallet address is the same 0x address on Ethereum, Arbitrum, Base and OP Mainnet, so the address field does not protect you from a wrong choice. The network field does. Pick the layer 2 you want, read the exchange’s minimum amount and fee, and send a small amount first.
The reverse trip needs more care. Before you send funds from a layer 2 to an exchange, open the exchange’s deposit page and confirm that it accepts that coin on that exact network. Sending to an exchange deposit address over a network the exchange does not list for that coin is one of the classic ways to lose access to funds.
Route 2: the rollup’s canonical bridge
The canonical bridge is a set of smart contracts on Ethereum written by the team behind the rollup. When you deposit, your ETH is locked in that contract and the same amount is credited to you on the layer 2. Its security is the rollup’s own security: you add no new party to trust. The price is time on the way out, because withdrawals wait for a challenge or proof window that lasts days.
One detail matters for safety. The contract is official, but the website in front of it is not always run by the rollup team. Arbitrum runs bridge.arbitrum.io. Base’s documentation states that the bridge that used to sit at bridge.base.org has been deprecated. For deposits from Ethereum it now lists independent front ends (Superbridge and Brid.gg), with a disclaimer that Coinbase takes no responsibility for how they operate: a listing, not a guarantee. For OP Mainnet, the Optimism documentation links to a Superchain bridges page on app.optimism.io. The rule that follows: open the bridge from the network’s own documentation, never from a search result.
Route 3: third-party and intent bridges
Fast bridges, Across and Relay among them, do not wait for the rollup. A relayer or solver pays you on the destination chain from its own funds, then gets repaid later from your deposit; Base’s documentation puts the delivery time at typically seconds to minutes. You pay a fee for that speed and you rely on the bridge’s code and operators instead of the rollup alone. Base’s documentation says so in plain terms: fast bridges do not inherit Base’s security.
How these designs work, and the incidents some of them had in 2026, are covered in our lesson on cross-chain bridges and intents, which also ends with a ten-point safety checklist. This lesson turns that checklist into concrete steps rather than repeating it. For the large bridge thefts of the past, Ronin, Wormhole and Nomad among them, see our history of DeFi hacks and exploits.
Before you start: five checks that take ten minutes
Most bridging losses do not come from broken cryptography. They come from a wrong network, a wrong website or an empty gas tank. Five checks remove most of that risk before you sign anything.
- Your wallet shows the network you think it shows. In MetaMask and similar wallets the network selector sits at the top; after adding a network, compare its chain ID with the table below.
- You opened the bridge from the network’s official documentation or a bookmark you made yourself, not from an ad, a reply on social media or a direct message.
- You hold ETH for gas on the sending side, and you know how you will pay gas on the receiving side. A canonical ETH deposit lands as ETH on the layer 2, which solves the second point; a USDC-only deposit does not.
- You know the withdrawal time of the route before you deposit, so a week-long wait does not come as a surprise when you need the money.
- You start with a test amount: a few dollars on mainnet after a rehearsal on Sepolia.
Chain IDs and explorers of the main networks
A chain ID is the number that tells your wallet which network it is signing for. Fake networks can copy a name and a logo but a check of the number against the project’s own documentation exposes them. The figures below come from the public registry at chainid.network and match the network pages of the Arbitrum, Optimism, Base and ZKsync documentation, all read on 15 November 2026; each network also returned the same number when asked directly over its public RPC.
| Network | Chain ID | Block explorer |
|---|---|---|
| Ethereum Mainnet | 1 | etherscan.io |
| Arbitrum One | 42161 | arbiscan.io |
| OP Mainnet | 10 | optimistic.etherscan.io |
| Base | 8453 | basescan.org |
| ZKsync Era | 324 | explorer.zksync.io |
| Ethereum Sepolia (test) | 11155111 | sepolia.etherscan.io |
| Arbitrum Sepolia (test) | 421614 | sepolia.arbiscan.io |
| OP Sepolia (test) | 11155420 | sepolia-optimism.etherscan.io |
| Base Sepolia (test) | 84532 | sepolia.basescan.org |
A registry only helps if the network you compare is the real one. In September 2026 a fake “GIWA” bridge collected deposits for a chain whose mainnet did not exist yet; the cross-chain lesson tells that case in full. The lesson for you: the reference is always the project’s own documentation, not a list a stranger sends you.
Step 1: rehearse the whole trip on Sepolia
A testnet rehearsal costs nothing and shows you every screen before real money is involved. Sepolia is Ethereum’s main test network, and Arbitrum, OP Mainnet and Base each run a Sepolia version of their layer 2 with its own test bridge.
First you need test ETH. Faucets give it away in small doses, but several now ask for proof that you are a real user; one popular faucet wants your address to hold a small balance of mainnet ETH first, as our step-by-step DeFi starter explains. Base’s faucet page lists several faucets for Base Sepolia, including the Coinbase Developer Platform faucet (up to 0.1 test ETH per 24 hours, plus test USDC) and others from infrastructure providers, most limited to one request a day. Circle runs its own faucet for test USDC.
Then do the rehearsal in this order:
- Add Sepolia and the Sepolia version of your target layer 2 to your wallet, and check both chain IDs against the table above.
- For Arbitrum, open the Arbitrum bridge with your wallet connected to Ethereum Sepolia: Arbitrum’s quickstart notes that the test networks only show up in the bridge when the wallet is on Sepolia. For OP Sepolia or Base Sepolia, use the bridge front end that the network’s documentation points to.
- Deposit a small amount of Sepolia ETH, note the transaction hash, and time how long the credit takes on the layer 2.
- Make a test transaction on the layer 2, for example sending a fraction of that ETH to a second address you own, to see how cheap layer 2 gas is.
- Start a withdrawal back to Sepolia and write down the waiting time the app shows you.
A testnet will not show you real fees or the scam sites that copy mainnet bridges. It can, however, teach you the wait: the Optimism tutorial notes that withdrawal times on OP Sepolia currently match mainnet’s, because fault proofs are being tested there. Other test networks may differ, so plan around the mainnet figures later in this guide and the countdown each app shows.
Step 2: a small real deposit of ETH through the canonical bridge
On mainnet, take the amount you would not mind waiting a week to recover, say the equivalent of 20 dollars, and bridge ETH through the canonical bridge. The screens look much the same on every front end.
- Open the bridge from the network’s documentation and connect your wallet. Check the address bar once more.
- Choose Ethereum as the source and your layer 2 as the destination. The interface should name the destination network clearly, for example Arbitrum One rather than another Arbitrum chain.
- Type the amount. Leave a margin of ETH on Ethereum for this transaction’s gas and for any later withdrawal claim.
- Read the review screen: amount, destination, estimated time and the Ethereum gas fee. If a fast or third-party route is offered by default, decide whether you want it; on aggregators the route can switch between bridges.
- Confirm in your wallet. A plain ETH deposit asks for one signature and no token approval.
- Wait. The Optimism documentation says deposits usually complete within 1 to 3 minutes; Arbitrum’s quickstart gives roughly 15 to 30 minutes depending on congestion. Base gives no figure, but its deposits do not wait for the withdrawal window.
When the credit arrives, your wallet balance on the layer 2 goes up by the amount you sent. The canonical contracts take no cut on an ETH deposit: what you pay is Ethereum gas. A third-party front end or a fast route can add its own fee, and the review screen is where it shows.
What a deposit cost on 15 November 2026
We asked an Ethereum node, at 09:46 UTC on 15 November 2026, how much gas each canonical bridge would use for a simple ETH deposit, and priced it at that minute’s gas price and the Coinbase spot price of ETH ($1,826.66). These are our readings at one moment, labelled as examples; your wallet shows the live figure before you sign.
| Reading, 15 Nov 2026 09:46 UTC | Value |
|---|---|
| Ethereum block | 26,419,597 |
| Ethereum gas price | 0.2198 gwei |
| Arbitrum One ETH deposit | 93,000 gas, about 0.0000204 ETH, about $0.04 |
| OP Mainnet ETH deposit | 662,368 gas, about 0.000146 ETH, about $0.27 |
| Base ETH deposit | 659,419 gas, about 0.000145 ETH, about $0.26 |
| Gas price on Arbitrum One | 0.020 gwei |
| Gas price on Base | 0.006 gwei |
| Gas price on OP Mainnet | 0.001 gwei |
The OP Mainnet and Base deposits use more Ethereum gas because their contract on Ethereum burns gas to pay in advance for the layer 2 part of the operation; the exact figure depends on the settings of the front end you use. Ethereum gas was unusually cheap that morning. On a busy day at 10 gwei, the same deposits would cost about $1.70 on Arbitrum and about $12 on OP Mainnet or Base at the same ETH price: our own arithmetic, not a forecast.

Once the funds are on the layer 2, gas is a different world. According to growthepie data for 10 November 2026, the median transaction fee was $0.0018 on Base against $0.030 on Ethereum.
Step 3: check the transfer on both block explorers
A bridge transfer leaves traces on two chains, so it has two explorers and often two transaction hashes. Reading them takes two minutes and is the only way to know where your funds really are. If explorers are new to you, start with our guide to reading a block explorer; here is the bridge-specific part.
- On Etherscan, open your deposit hash. The status should read success, the value should match your amount, and the “To” field should be a bridge contract of the network you chose, not an unknown address. The Arbitrum bridge sends ETH deposits to Arbitrum’s Inbox contract; OP Mainnet and Base front ends call the standard bridge or portal contract of that network. Etherscan labels the official ones by name, which is a quick check.
- On the layer 2 explorer (Arbiscan, Basescan or Optimistic Etherscan), search your own address. The deposit appears as an incoming credit, sometimes on a tab for transactions coming from Ethereum rather than the main list.
- In the wallet, switch to the layer 2 network. If the balance is right on the explorer but not in the wallet, the wallet is showing another network.
- Save both hashes somewhere you can find them. They are your proof if you need support, and your records for tax purposes.
Step 4: moving USDC, and why the contract address matters more than the name
Stablecoins add a trap that ETH does not have: several different tokens can carry the name USDC on the same layer 2. Circle issues native USDC directly on Arbitrum, Base and OP Mainnet. Before that, bridges locked USDC on Ethereum and minted their own copies on layer 2s, and some of those copies still circulate under names such as USDC.e. On Base, the older bridged token is called USDbC, and Base’s documentation describes it as a separate legacy token not issued by Circle. Its advice: check the contract address, not the ticker.
| Network | Native USDC contract (Circle) |
|---|---|
| Ethereum | 0xA0b86991c6218b36c1d19D4a2e9Eb0cE3606eB48 |
| Arbitrum One | 0xaf88d065e77c8cC2239327C5EDb3A432268e5831 |
| Base | 0x833589fCD6eDb6E08f4c7C32D4f71b54bdA02913 |
| OP Mainnet | 0x0b2C639c533813f4Aa9D7837CAf62653d097Ff85 |
| Sepolia (test) | 0x1c7D4B196Cb0C7B01d743Fbc6116a902379C7238 |
| Arbitrum Sepolia (test) | 0x75faf114eafb1BDbe2F0316DF893fd58CE46AA4d |
| Base Sepolia (test) | 0x036CbD53842c5426634e7929541eC2318f3dCF7e |
| OP Sepolia (test) | 0x5fd84259d66Cd46123540766Be93DFE6D43130D7 |
These addresses come from Circle’s USDC contract list, read on 15 November 2026. Compare at least the first and last six characters with what your wallet or the bridge shows, and import the token in your wallet by its address if it does not appear.
Native USDC moves between chains through Circle’s own transfer protocol, CCTP, which burns the tokens on one chain and mints them on the other, so you receive native USDC rather than a copy. Many bridge interfaces use it behind the scenes. Its mechanism and timings are in the cross-chain lesson; how USDC differs from its main rival is in our USDT vs USDC comparison.
USDC is a token, so a deposit through a token bridge usually takes two transactions: an approval that lets the bridge contract move your USDC, then the deposit itself. Each costs gas. Approve the amount you are bridging rather than an unlimited amount when the interface lets you choose. And keep a little ETH on the destination layer 2, because USDC cannot pay for gas there.
Coming back: withdrawals through the canonical bridge take days
Deposits are fast because Ethereum already knows about them. Withdrawals are slow because Ethereum has to be convinced that the layer 2 really says you own the funds. Optimistic rollups give anyone a window to challenge a wrong state; rollups that also use proofs can shorten it when the proofs check out. The user-facing numbers, as of 15 November 2026:
| Network | Withdrawal through the canonical bridge | What you have to do | Source |
|---|---|---|---|
| Arbitrum One | about a week (challenge period 6 days 8 hours; the app tells users to count 7 to 8 days) | start on Arbitrum, then press Claim on Ethereum | Arbitrum quickstart, ETHBEAT |
| OP Mainnet | 7 days | start on OP Mainnet, prove on Ethereum (after up to an hour), finalize on Ethereum after the window | Optimism docs |
| Base | 5 days; 1 day when both its TEE and ZK proofs back the state | through a third-party front end on Base’s official contracts | Base docs |
| Fast or intent bridges | seconds to minutes | one transaction and a fee | Base docs |
On Arbitrum, the protocol’s challenge period is set to 6 days and 8 hours, according to ETHBEAT (formerly L2BEAT), but the bridge app adds the time for the withdrawal to be posted and tells users to count 7 to 8 days. When the countdown ends, nothing happens on its own: you return to the bridge and press Claim, which is an Ethereum transaction you pay for.
On OP Mainnet, the Optimism bridging tutorial lays out three moves: start the withdrawal on OP Mainnet, prove it on Ethereum once the network has posted its state, then finalize it on Ethereum after the 7-day window. That is two Ethereum transactions to pay, so keep some ETH on Ethereum for them. Base, which left the OP Stack in February 2026 and now runs its own proof system, gives 5 days in its documentation, or 1 day when both of its proof types back the result.
ZK rollups such as Starknet, ZKsync Era, Linea and Scroll follow their own rules. Their exit delays are set by each project and have changed more than once, so read the figure in the official bridge app on the day you start, not in an old article.
If you cannot wait, a fast bridge brings funds back to Ethereum in minutes for a fee. The trade is the one described in route 3: you swap the rollup’s guarantee for the bridge operator’s. On a small amount the fee and the extra trust may be worth a week saved. On a large one, the week is what the stronger guarantee costs.
Funds that “don’t arrive”: six causes and what to do
A transfer that seems lost is usually just in a place you are not looking. Go through these causes in order before you contact anyone.
- The wallet shows the wrong network. Switch to the destination layer 2 and look again; a balance on Base does not appear while the wallet is set to Ethereum.
- The token is not displayed. Wallets do not always list a token automatically. Import it by the contract address from the issuer’s page, as in the USDC table above, never from an address someone sends you.
- The deposit is still pending. Compare the time elapsed with the network’s deposit time. On Etherscan, a pending or failed deposit tells you where the delay sits.
- On Arbitrum, a token deposit that carries extra instructions travels as a “retryable ticket”. If it fails to execute automatically, Arbitrum’s documentation says the ticket stays available for up to one week for manual redemption, after which it expires and the escrowed ETH goes to the refund address. Plain ETH deposits rarely hit this, but token transfers through some interfaces can; the bridge app shows a retry button when it happens.
- A withdrawal is waiting for you. On Arbitrum and OP Mainnet, funds stay in the bridge until you claim or finalize them on Ethereum. Reconnect the same wallet to the same bridge and look at its transaction history.
- You withdrew from an exchange on a network the receiving side does not support, or deposited to an exchange on the wrong network. Only the exchange can sometimes recover such funds. Contact it through the support link inside its own app or website.
Never answer people who offer help in replies, private messages or comment threads after you post about a missing transfer. Scammers watch for those posts. No real support agent ever needs your seed phrase, and “recovery services” that promise to get funds back are a scam category of their own.
The scams that target people who bridge
Bridging makes you a target for one reason: you are about to sign a transaction that moves money to a contract, and you have probably searched for “bridge” a minute earlier. The common traps fall into four families.
- Fake bridge sites bought as search ads. They copy the look of the real front end, sometimes the domain with one letter changed, and ask you to connect and sign. Bookmark the official address once, from the network’s documentation, and use only the bookmark.
- Fake support. Messages on X, Discord or Telegram that offer to “sync” your wallet or fix a stuck transfer, often after you complained in public. They end with a request for your seed phrase or a signature on a malicious site.
- Fake networks and fake tokens. A network added from a random site can point your wallet to a chain that is not what its name says; a token with the right ticker can be worthless. Chain ID and contract address checks catch both.
- Approvals to unknown contracts. An approval lets a contract spend a token from your wallet, often without limit and without an end date. A drainer site only needs one such signature.
If you have approved a contract you do not trust, you can withdraw the approval. Tools such as Revoke.cash or the token approval checker on Etherscan list your approvals and let you cancel them, at the cost of a little gas per cancellation. The full routine, with simulation and a separate wallet for experiments, is in our DeFi security routine. A later lesson covers token approvals and revoking them in more depth.
One rule covers everything above: never type your seed phrase into a website, a form or a chat, whatever reason it gives. A bridge never needs it, and neither does any support desk.

Taxes and records when you bridge
No tax authority in the United States or the United Kingdom has written a rule specifically about bridging from Ethereum to a layer 2. What exists are general rules on moving assets between wallets you own, plus the question of the fee.
For US taxpayers, the IRS digital assets page says that transferring digital assets between wallets or accounts you own or control does not by itself require a “Yes” answer to the digital asset question on Form 1040, unless you paid a transaction fee with digital assets. Paying a transfer fee in crypto is itself listed among the transactions that do. The same page lists “paying a transfer fee with digital assets” as a disposal, so the gas you pay in ETH for a deposit or a claim is a small disposal of ETH to record. Whether a particular bridge counts as moving the same asset or as exchanging ETH for a wrapped token is not settled by a specific rule; our guide to US crypto taxes gives the general framework.
For UK taxpayers, HMRC’s cryptoassets manual says at CRYPTO22100 that there is no disposal when you keep beneficial ownership throughout, for example when you move tokens between public addresses you control. For tokens moved from one distributed ledger to another, the manual’s Corporation Tax section for businesses, CRYPTO41260, gives HMRC’s view that whether this is a disposal “will depend on the facts”. It is written for companies and is not a rule on layer 2 bridges. That is the position to state; our UK crypto tax guide covers how disposals are computed.
Whatever the jurisdiction, keep records: both transaction hashes, the date, the amount, the gas paid and its value in dollars or pounds at the time. A spreadsheet line per transfer is enough.
A one-page checklist for your next bridge transfer
- Decide the route: exchange withdrawal, canonical bridge or fast bridge, and know its time out before you go in.
- Open the bridge from the network’s documentation or your own bookmark.
- Check the network and chain ID in the wallet.
- Keep ETH for gas on Ethereum, and get some ETH onto the layer 2 if you are moving only USDC.
- Send a test amount, then the rest.
- Read the review screen, including the route and the gas fee.
- Check both explorers and save both hashes.
- For USDC, compare the contract address with Circle’s list.
- Approve only what you bridge; review old approvals from time to time.
- Never share a seed phrase, and ignore unsolicited “support”.
Every step above rests on a layer 2 doing what it promises. The next lesson of this path looks at how far you can trust that: the risks specific to layer 2 networks, such as upgrade keys, centralised sequencers and missing proofs, and how to read their ratings on ETHBEAT.



