Ethereum’s Dencun upgrade activated on 13 March 2024. The part readers will feel is not a slogan about “scalability”. It is blobs — extra, cheaper data space aimed at layer-2 networks — and the sudden drop in what those networks charge to move a small amount.
Low fees are pleasant. They are not a personality trait of the chain, and they are not a forecast. Gas is still an auction. Dencun changed how much space rollups can buy. It did not abolish demand, sequencers, bridges, or the habit of declaring a new era every time the meter slows down.
What happened
Dencun (Cancun on the execution layer, Deneb on consensus) included EIP-4844, proto-danksharding. Rollups can post data in blobs that are cheaper than stuffing everything into ordinary calldata. Within hours, several layer-2 fee dashboards showed transfers that had cost dollars falling toward cents. That is a real product change. It is also easy to misread as “Ethereum is free now”.
The base layer is still a scarce block. If you are settling on Ethereum itself, you are still bidding against everyone else who wants inclusion. If you are using a rollup, you are using that rollup’s operator, its bridge, and its habit of explaining networks in a wallet dropdown.
Why it matters
Fee spikes are how newcomers decide the technology is a joke. Fee droughts are how insiders decide the technology has won. Both reactions skip the mechanism. This week is a good time to practise a test transaction on the network you actually intend to use, inspect a block explorer, and see a confirmation without the panic of a $40 “oops”.
If you are writing a business plan that assumes fees stay here, you are writing fiction. Protocol changes can alter the supply of space. Airdrops and mints can alter demand. Neither is a straight line.
Layer 2 does not make the base layer disappear
When activity moves to a rollup, users still eventually depend on Ethereum for settlement and for the security story on the brochure. They also depend on the rollup’s sequencer, bridges, and support culture. Cheap fees on a layer 2 with a confusing withdraw path can cost more than a dear fee you understood.
Collecting five new networks because they are cheap is how people misplace assets they later describe as “stolen by the blockchain”. Pick one, send a tiny test, wait for the full withdraw path to make sense, then stop.
What to watch
Watch whether activity moved to layer 2 rather than vanished. Watch whether a new token launch bids the base layer back up. Watch, if you care about users, whether wallets label networks clearly enough that cheap fees do not become expensive mistakes.
Not advice. Not a prediction. A reminder that the meter still exists even when blobs have made it spin more slowly for rollups.
