Sending money cheaply over an Ethereum layer 2 seemed like a given. On October 4, that given shook, because the price of blob gas spiked by around 400% and forced rollups to pay far more than expected to publish data. The network did not go down and no one lost funds, but the episode left an uncomfortable lesson. Cheap data space has a ceiling, and demand already knows how to hit it.
Blobs are the format layer 2s have used to publish their transactions on Ethereum since the Dencun upgrade. They run on their own fee market, separate from regular gas, which only gets more expensive when demand exceeds the target set by the protocol. Pectra, the major subsequent upgrade, expanded that target to give rollups more room. Even with more space, the first serious wave of demand was enough to quintuple the price within days.
The question running through the sector as of October 15 is whether that was a passing scare or the first warning of a structural ceiling. Rollups live by publishing cheaply and charging a margin, so every blob spike bites into their business. Understanding what blobs are, why their price multiplies so quickly and what limits Pectra set makes it possible to gauge what comes next.
What are blobs and why do they have their own price?
A blob is a data package that travels attached to a block without going through the virtual machine. Rollups use it to record their transactions on layer 1 while paying far less than with regular data. The official documentation of Ethereum describes this model as its path to scaling.
No one sets the blob price by hand. Each block has a blob target, and if demand exceeds it, a base fee rises automatically until consumption cools. If demand eases, the fee drops to almost zero. It is a thermostat, not a counter, and it reacts with the coldness of a formula.
That coldness explains the violence of the move. When several rollups publish above the target at the same time, the fee does not rise a little. It multiplies block after block until someone gives way. A 400% rise sounds like a catastrophe, but in this market it is the thermostat doing its job.
Dencun activated blobs in March 2024, more than two years ago, and since then the price had slept near zero. That prolonged calm explains why the October spike caught so many by surprise.
How did the October 4 rise come about?
The rise was not caused by a single culprit with a name and surname. More publications than usual coincided, with several rollups settling batches at a high rate over the same weekend. Aggregate demand exceeded the protocol's target and the base fee chained automatic increases until it stabilized well above its previous level.
The calendar helps place the scare. Pectra had recently expanded the available capacity, which generated confidence that there would be plenty of room. That confidence encouraged more activity, and the activity ate up the new cushion within days. It is the old pattern of every capacity expansion in shared infrastructure.
The data released on October 4 put the jump at around 400% compared with previous levels. The figure should be read as an order of magnitude rather than an exact quote, because the blob price changes with every block. What matters is not the exact figure, but that the market proved it can come under strain despite the expansion.
Pectra's limits and where they pinch
Pectra raised the number of blobs per block and adjusted the cadence at which the protocol pursues its target. On paper, the network can move more layer 2 data per second than before, with spikes absorbed without spilling over into regular gas. In practice, the episode showed that a target does not mean a guarantee.
The technical nuance is that the target marks the point where the price neither rises nor falls. Sustained overshooting triggers the fee, and only excess supply brings it back down. Pectra moved that point higher, but did not eliminate the mechanism. As long as demand can concentrate during peak hours, there will be spikes.
For rollups, the limit that hurts is not physical but economic. Every batch published during an expensive hour cuts into their margin or forces them to raise fees for the end user. Those who publish on a rigid schedule suffer more than those who can wait for the night-time valley to settle.
A party whose bill keeps getting steeper
The payment chain is short and has no shock absorbers. The rollup pays the blob gas in ether, passes it on in the fees it charges its users or absorbs it from its treasury. When the price multiplies fivefold overnight, one of the three legs cracks.
Large rollups have the treasury to withstand an expensive weekend without touching their fees. Small ones, by contrast, operate with thin cushions and publish on rigid schedules. A prolonged spike puts them before an ugly dilemma. Either they raise prices and lose users, or they hold prices and burn cash.
End users felt the episode unevenly. Those operating during off-peak hours barely saw changes, while those who needed immediate confirmation paid the premium. It is the same hourly lottery as regular gas, transferred to the layer 2 data market.
What the spike teaches about Ethereum's design
The episode validates a design decision that was much debated at the time. Separating the blob market from regular gas prevented the fever from spreading to the whole network. While blobs were getting more expensive, ordinary layer 1 transactions continued without disruption. The firewall worked as intended.
It also teaches that automatic fee markets punish synchrony. If all rollups publish at the same time, they compete with each other and the thermostat penalizes them all. Spreading publications out over time makes the whole cheaper, but it requires coordination that no one imposes today. Each rollup optimizes for itself and the aggregate pays the bill.
The third lesson looks to the future of scaling. As long as layer 2 usage grows faster than blob capacity, spikes will recur. The roadmap contemplates further expansions, with Glamsterdam already testing much larger blocks on Sepolia. The race between demand and capacity has only just begun.
| Phase |
Blob capacity |
Effect on price |
What it showed |
| Before Dencun |
No blobs, expensive data |
High fees on L2 |
The problem that had to be solved |
| Dencun |
Blobs with initial target |
Price near zero for months |
The calm before demand |
| Pectra |
Expanded target |
Bigger cushion, but finite |
The expansion fills up quickly |
| October 4 spike |
Demand above target |
Rise of around 400% |
The thermostat works and it bites |
Rollups await the next spike
Teams are already adjusting their publishing strategies. Delaying non-urgent batches to cheap hours, compressing data better and sharing publishing windows are the immediate levers. None requires changing the protocol, only operating with more cost discipline.
In the medium term, service-level commitments appear. Some rollups are studying maximum-fee guarantees for their frequent users, taking on the spike risk themselves in exchange for loyalty. Others prefer to pass the price on as is and compete on transparency. Both stances are legitimate and the market will judge them.
Coordination among rollups would be the elegant solution and the hardest one. A shared publishing calendar would flatten demand without expanding anything, but it would ask competitors to cooperate on their main cost. Until that happens, every spike will find rollups publishing at the same time and overpaying.
Meanwhile, attentive users already check Ethereum's scaling roadmap and public blob price dashboards before operating. Checking the fee before accepting a commission avoids surprises and rewards patience with visible savings.
A useful scare for an Ethereum that scales
Nothing broke on October 4 and that is the good news. The network absorbed the demand, the price did its job and funds were never at risk. Fee markets exist precisely for this, to ration a scarce resource without administrators in moments of strain.
The bad news is that the scare has a repeat date. Every cycle of intense layer 2 activity will push the blob price against its ceiling again, and Pectra only bought time. The next expansion will arrive with Glamsterdam and its giant test blocks, but demand has already shown it knows how to fill every new gap.
Anyone who uses layer 2s daily can draw a practical conclusion. Off-peak hours exist in the world of blobs too, and operating with patience is cheaper than demanding immediacy. October 4 reminded us that on Ethereum, as on the highway on a Friday afternoon, the price is set by other people's haste.