Fusaka and PeerDAS

What shipped

The Fusaka upgrade is the most consequential change Ethereum shipped in 2025, and it went live on 3 December 2025 at 21:49 UTC. Ethereum names its upgrades and ships them in coordinated hard forks; Fusaka is the one that landed at the end of the year, and its aim was narrow and important: to make the network cheaper to run and cheaper to build on without weakening the decentralisation that gives it its value. The headline component is a technology called PeerDAS.

How PeerDAS works

PeerDAS, short for peer data availability sampling, changes how the network handles blob data, the temporary data that Layer 2 rollups post to Ethereum when they settle their transactions. Before the Fusaka upgrade, every node had to download and store all of that blob data in order to verify it. PeerDAS removes that requirement. Nodes instead sample small portions of the data and rely on the network as a whole to hold the rest, which lets the network verify far more data than any single machine keeps. The result is a theoretical increase of up to eight times the previous blob capacity.

Blob capacity is the resource that sets the price of rollup transactions. Rollups compete for room to post their data to Ethereum, and when that room is scarce the cost is passed on to users as higher fees. By raising the amount of data the network can carry without raising what any one node must store, the Fusaka upgrade loosens that constraint, which is the mechanism by which it makes Layer 2 activity cheaper.

The gas limit and its guardrail

Alongside the data change, the Fusaka upgrade raised the default block gas limit target toward 60 million, increasing the amount of computation each block can carry. Lifting that ceiling on its own would widen the door to denial-of-service attacks, in which a single very heavy transaction clogs the network. To close that gap, the upgrade shipped EIP-7825, a per-transaction gas cap of 16.78 million, deliberately at the same time as the ceiling rise. The pairing was deliberate design: the network gained headroom and a new guardrail in the same release, so the two changes could be reasoned about together rather than one arriving before the other was ready. The cap ensures no individual transaction can consume a disproportionate share of a block, so the extra capacity is available to many users rather than exploitable by one.

The two costs that decide it

The plain-English way to read the Fusaka upgrade is through two costs. The first is the cost of running a node, which determines how many independent people can afford to help verify and secure the network. The second is the cost for rollups to post their data, which determines how cheap Layer 2 transactions can be. Those two costs are the constraints that decide whether Ethereum can grow its capacity while staying decentralised. Push node requirements too high and only well-funded operators remain, which concentrates control. Leave data too expensive and the rollups that carry most user activity struggle to keep fees low. Fusaka pushed on both at once.

Why it is significant

Its significance is that it treated capacity and decentralisation as a single problem rather than trading one off against the other. Most scaling shortcuts buy throughput by asking nodes to do more, which quietly raises the bar for who can run one; Fusaka pushed capacity up while pushing the storage burden on individual nodes down. For anyone tracking how Ethereum intends to carry far more activity without becoming a network only large operators can run, it is the clearest recent statement of the approach. The dates and details of what follows will be argued over, but the principle Fusaka set, that more capacity must not cost decentralisation, is likely to frame every upgrade after it. Our Ethereum explainer sets these upgrades in the context of the wider roadmap.