If you’ve used Ethereum for more than five minutes, you’ve felt it: that moment where you’re ready to hit “Confirm” and the wallet hits you back with a fee that makes you rethink your life choices.
So when people say “Ethereum fees are basically free now,” it can sound like hype. Ethereum is still a global settlement network with scarce blockspace. Scarcity usually means fees.
And yet, in 2025, something real shifted.
Not because Ethereum magically stopped charging for computation, but because the part of Ethereum you feel as a user changed. Between two major network upgrades in 2025, Ethereum got materially better at the two things that dominate everyday cost and UX:
-
Making rollups (Layer 2s) dramatically cheaper to run
-
Making wallets capable of “gasless” experiences by default
That combination is why a growing chunk of Ethereum activity now feels like “tap button, pay pennies (or nothing), move on.”
Let’s break down what actually changed, and why it matters.
The key idea: Ethereum is no longer “the place you do every transaction.”
Ethereum’s scaling plan has been telling the same story for a while: don’t cram everything into Layer 1 (mainnet). Instead:
-
Use Layer 2 rollups for most user transactions (payments, swaps, NFTs, app actions)
-
Use Ethereum Layer 1 as the settlement and security layer that rollups anchor to
This matters for fees because rollups bundle tons of user transactions, then post compressed proofs and data back to Ethereum.
So your “gas fee” for a rollup is mostly the cost of publishing data to Ethereum, divided among many users.
That’s why the biggest lever for cheaper fees is not “make Ethereum blocks infinite.” It’s “make rollup data cheaper and more scalable.”
That’s exactly what Ethereum doubled down on in 2025.
Upgrade #1 (May 2025): Pectra made wallets smart enough to hide gas
Ethereum’s Pectra upgrade activated on the mainnet in May 2025.
One of the most user-facing changes was a big step toward mainstream “smart wallet” behavior without forcing everyone to migrate to a new wallet contract.
The headline UX change: your normal account can temporarily behave like a smart account
Pectra included EIP-7702, which allows a regular externally owned account (EOA) to delegate execution logic, enabling smart-account-like features.
In plain English: the account you already use can now do things that used to require a full smart contract wallet, like:
-
Sponsored (gasless) transactions: an app or wallet provider pays the gas for you
-
Batching: approve + swap in one action, or multiple app actions in one confirmation
-
Session keys: “stay logged in” experiences that don’t require constant signing in
-
Better recovery and security patterns (when wallets implement them)
None of this removes gas at the protocol level. Ethereum still charges for execution. What it changes is who pays, and how often you have to think about it.
Why does this make transactions feel free?
Most users don’t care whether fees are eliminated or abstracted. They care if:
-
The transaction goes through
-
They aren’t surprised by costs
-
They don’t have to keep ETH around just to pay fees
With EIP-7702-style flows, an app can say: “You’re paying in USDC, we’ll cover the gas.” Or: “This is an onboarding step, we’ll sponsor it.”
That is a psychological and practical breakthrough. It removes the biggest “first-time user” friction: “Wait, I need ETH to use my USDC?”
Upgrade #2 (December 2025): Fusaka made rollup data cheaper and easier to scale
If Pectra was the “wallet UX” upgrade, Fusaka was the “make rollups cheaper at scale” upgrade.
Fusaka activated on the Ethereum mainnet on December 3, 2025, and its centerpiece is PeerDAS (EIP-7594).
What PeerDAS does (without the PhD version)
Rollups post data to Ethereum using “blobs,” a data format introduced earlier to reduce the cost of rollup publishing. The limit on cheaper rollup transactions is often: how much blob data Ethereum can safely handle without forcing every node to download and store everything.
PeerDAS changes the data availability model so the network can scale blob throughput without pushing hardware requirements through the roof.
More blob capacity and safer scaling translate into:
-
Lower rollup operating costs
-
More predictable rollup fees under load
-
Room for rollups to handle big usage spikes without fees exploding
Fusaka also introduced a concept called Blob Parameter Only (BPO) forks, which are smaller, config-style upgrades designed to safely increase blob throughput after PeerDAS.
If you care about fees, that last part is a big deal. It means Ethereum can keep tuning the “cheap data lane” for rollups more quickly and safely than by doing full, feature-packed hard forks every time.
What this means for real users
When rollups can publish more data per unit cost, fees on L2s can drop into ranges like:
-
fractions of a cent for simple transfers
-
a few cents for more complex actions
-
and in many “sponsored” flows, effectively $0 to the user
Again: Ethereum mainnet is not suddenly free. But for the majority of consumer-style activity, it’s increasingly normal for the fee to be either tiny or invisible.
So are Ethereum transactions actually free now?
Not literally.
Here’s the honest breakdown:
Layer 1 still has fees
If you send ETH on mainnet, swap on mainnet, or interact with a mainnet contract directly, you’re still bidding for limited blockspace. Fees can still spike.
Layer 2 is where “nearly free” becomes true
On modern rollups, especially for simple actions, costs can be extremely low because:
-
transactions are batched
-
Data is compressed
-
And Ethereum has been making rollup data cheaper and more scalable
“Gasless” is real, but it’s a product choice
When you see a transaction that costs you nothing, someone is still paying for it. Usually:
-
the app (customer acquisition cost)
-
the wallet provider
-
a relayer/paymaster setup
-
or the protocol treasury
Pectra makes it easier to build these experiences and to offer them more naturally.
What you’ll notice in practice (even if you never read an EIP)
These upgrades show up as everyday improvements:
1) You stop “keeping ETH around just for gas.”
Apps can sponsor transactions or let you pay fees in the token you’re using, which removes a classic onboarding tax.
2) You click less, sign less, approve less
Batching and smarter account flows reduce the “death by 6 popups” feeling.
3) Fees don’t jump as violently during busy moments
More scalable rollup data capacity helps smooth out the worst spikes, especially on L2s.
4) On-chain UX starts to resemble normal apps
When transactions are cheap and abstractable, teams can design flows that feel like Web2 without breaking decentralization.
Why 2025 was the tipping point
Ethereum has been “the rollup chain” in theory for a while, but 2025 is when two major pieces clicked together:
-
Fusaka (PeerDAS) strengthens the cheap data pipeline, enabling rollups to scale fees without sacrificing decentralization.
-
Pectra (EIP-7702) makes it dramatically easier for wallets and apps to offer gasless, low-friction experiences for regular accounts.
You need both.
Cheap execution without good UX still feels clunky. Great UX without cheap execution still feels expensive.
Together, they make Ethereum feel different.
What builders should do to take advantage of this
If you’re building apps, wallets, or onboarding flows, 2025’s upgrades are basically an invitation to redesign the “first 5 minutes” experience.
A few practical directions teams are taking:
-
Sponsor the first transaction (account creation, first swap, first mint)
-
Batch common flows (approve + action)
-
Use session-based permissions for repeat interactions
-
Move default user activity to L2s and use L1 only when needed
-
Treat gas as an implementation detail instead of a user-facing feature
The competitive bar is rising fast. When a user can do the same action for $0.01 elsewhere, your app charging $8 on mainnet will feel broken, even if it’s “technically normal.”
The fine print: what to stay aware of
Even in this cheaper world, a few realities remain:
-
Mainnet will always be premium: it’s the settlement layer.
-
Not all L2s are equal: decentralization, uptime, and security models vary.
-
Sponsored gas introduces new trust and abuse surfaces: apps need rate limits, fraud controls, and good account security patterns.
-
Bridging and withdrawals can still cost more depending on how you move assets across layers.
“Nearly free” is a user experience outcome, not a law of physics.
Where this goes next
If 2025 was about “cheap L2 transactions + gasless UX becomes normal,” the natural next chapter is:
-
even more scalable data availability
-
continued improvements to wallet safety and recovery
-
smoother cross-L2 movement so users stop caring which rollup they’re on
The direction is pretty clear: Ethereum is pushing hard toward a world where the average user doesn’t think about gas fees at all, the same way most people don’t think about TCP/IP packets.
And after 2025, that world feels a lot closer than it did just a couple of years ago.