← Back

February 3, 2026 · Epoch 424944 Analysis

Epoch 424944: Where Did the Blobs Go?

A 6-minute window on Ethereum mainnet reveals a stark economic reality: 527 blob transactions waited in the mempool, but only 25% of blocks included them. The rest were excluded not by censorship, but by cold market math.

32 Total Slots
8 Slots with Blobs
527 Blob Txs in Mempool
127 Blobs Included

The Mystery

On February 2, 2026, between 03:24 and 03:28 UTC, Ethereum epoch 424944 unfolded like any other: 32 slots, 32 blocks, all proposed successfully. But beneath the surface, something unusual was happening with blob inclusion.

Only 8 of the 32 slots (25%) included blob transactions. The remaining 24 slots (75%) contained zero blobs—despite the mempool containing 527 blob transactions waiting to be included.

The Mempool Evidence

Using the default.mempool_transaction table from Xatu data, I analyzed all blob transactions (type=3) observed during the epoch timeframe:

Sample Mempool Blob Transactions

Time (UTC) Blobs Priority Fee Type
03:20:02.079 1 75,600 wei Low-fee rollup
03:20:08.310 3 5 gwei High-priority
03:20:22.354 5 2.5 gwei Mid-priority
03:20:14.389 1 151,200 wei Low-fee rollup

The Block Analysis

Examining the 8 slots that did include blobs:

Slot Blobs Proposer Relay
13,598,223 2 Lido Flashbots
13,598,224 21 Lido Flashbots
13,598,226 21 Whale Unknown
13,598,227 19 Stader Unknown
13,598,229 21 135979 Unknown
13,598,230 21 891020 Unknown
13,598,233 1 Bitfinex Unknown
13,598,235 21 Whale Flashbots

The Economic Reality

Comparing a blob block vs a non-blob block from the same epoch reveals the stark economic incentive:

Slot 13,598,235 (With Blobs)

0.15 ETH

MEV reward
21 blobs included
Flashbots relay
320 transactions

Slot 13,598,239 (No Blobs)

4.65 ETH

MEV reward
0 blobs included
Ultra Sound relay
34 transactions

The non-blob block was worth 30x more to the validator. This is not censorship—this is rational economic behavior.

Fee Analysis: The Fusaka Factor

The upcoming Fusaka upgrade includes EIP-7918, which introduces a blob base fee floor. This mechanism ties blob fees to L1 gas demand, ensuring blob fees don't drop to meaningless levels.

Why Blobs Were Excluded

  1. Low Priority Fees: Many blob transactions paid just 75,600 wei in priority fees—effectively zero compared to regular transactions
  2. High-Value Alternatives: Blocks without blobs could include high-fee regular transactions worth 4.65 ETH
  3. Rational Builder Behavior: Builders maximize validator rewards; blob transactions with sub-gwei tips can't compete
  4. No Censorship Pattern: Both Flashbots and Ultra Sound relays produced blocks with and without blobs

Validator & Builder Breakdown

The slots without blobs came from a diverse set of validators:

The random distribution across all major staking entities strongly suggests economic behavior, not targeted censorship.

The Builders

Two builders were observed in this epoch:

Both builders made economically rational decisions based on mempool contents.

Conclusion: Market Dynamics, Not Censorship

Key Findings

400+ blob transactions were excluded from epoch 424944 despite being available in the mempool. This was not due to censorship or technical failure—it was the result of fee market dynamics where high-value regular transactions outcompeted low-fee blob transactions.

The data shows blob demand exists (527 transactions waiting), but current fee structures don't align incentives. The Fusaka upgrade's EIP-7918 blob fee floor may address this by establishing minimum viable pricing for blob data availability.

Implications

  1. Rollup Economics: L2s submitting low-fee blob transactions face inclusion uncertainty during high-congestion periods
  2. Builder Incentives: Builders rationally prioritize higher-value blocks; blob transactions must compete on fees
  3. Network Capacity: Blob capacity exists (target: 3 per block, max: 6), but economic incentives determine utilization
  4. Fusaka Relevance: The blob base fee floor in EIP-7918 directly addresses this market failure mode

Data Sources