The Empire Strikes Back: How Ethereum's 'Fusaka' Upgrade Redefines the Data Availability War

The Core Narrative

Fusaka — a combination of Fulu (Consensus Layer) and Osaka (Execution Layer) — is Ethereum’s most aggressive attempt to reclaim dominance in the Data Availability (DA) market.

Mainnet Target Date: December 3, 2025.

Over the past two years, Alt-DAs such as Celestia, EigenDA, and Avail gained momentum as low-cost DA solutions for rollups. Fusaka changes the equation by fundamentally increasing Ethereum’s own DA throughput and lowering its native DA cost.


PeerDAS (EIP-7594): The Game Changer

Before Fusaka, Ethereum nodes had to download entire blobs to verify data availability — limiting scalable data throughput to ~3–6 blobs per block.

Sampling Instead of Downloading

PeerDAS (Peer Data Availability Sampling) introduces:

  • Random sampling instead of full blob downloads
  • Reed-Solomon erasure coding for recoverability
  • Lower bandwidth requirement per node
  • Massive DA scaling, targeting 64+ blobs per block in future updates

Because nodes only download tiny random pieces, the network can safely increase total blob capacity without overwhelming home stakers.


BPO Forks (EIP-7892): Ethereum’s New “Volume Knob”

BPO = Blob-Parameter-Only fork

Before Fusaka, increasing blob capacity required a full hard fork.

Now Ethereum can adjust blob limits without a major network upgrade.

Scheduled Capacity Increases

  • Dec 3, 2025 — Fusaka mainnet launch
  • Dec 9, 2025 — BPO1 (+67% blobs)
  • Jan 7, 2026 — BPO2 (+133% total increase)

This introduces a flexible system that lets Ethereum match DA capacity to real-time demand.

Data Throughput Roadmap (Blobs per Block)

1
2
3
4
5
6
7
8
9
10
11
Current (Dencun)    : ▇▇▇ (3 Target / 6 Max)
└─ ~0.75 MB max per block

Fusaka Launch (Dec) : ▇▇▇▇▇▇▇▇ (9 Target / 18 Max)
└─ ~2.2 MB max per block (3x Increase)

Fusaka BPO2 (Jan) : ▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ (15 Target / 30 Max)
└─ ~3.7 MB max per block (5x Increase)

Future Goal (PeerDAS): ▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇... (64+ Target)
└─ ~8.0 MB+ max per block (21x+ Increase)

This roadmap illustrates Ethereum’s aggressive scaling strategy, moving from the current 3-6 blob capacity to a future target of 64+ blobs per block through PeerDAS implementation.


Impact on Alt-DAs (Celestia, EigenDA, Avail)

Fusaka doesn’t destroy Alt-DAs — it compresses their margins.

Side-by-Side Comparison

Feature Ethereum (Fusaka) Alt-DAs (Celestia, EigenDA, Avail)
Throughput High (PeerDAS scaling) Very High (DA-first chains)
Cost Low Lowest
Security Gold Standard (ETH L1) Smaller validator sets / restaking
Use Case General rollups + L2s Ultra-high throughput apps

The Squeeze

General-purpose rollups like Arbitrum, Optimism, Base, and zkSync will see fees fall significantly.
Their incentive to migrate to Alt-DAs decreases.

The Pivot for Alt-DAs

  • Celestia will target bandwidth-heavy use cases (gaming, social, on-chain compute).
  • EigenDA benefits indirectly because Fusaka increases the L1 gas limit (EIP-7935), lowering proof costs.

Other Goodies in Fusaka

These upgrades ship alongside the DA improvements:

1. EIP-7935 — Gas Limit Increase

Block gas limit doubles from 30M → ~60M, increasing L1 computation capacity.

2. EIP-7951 — Passkey Support

Adds secp256r1 cryptography, enabling FaceID/TouchID wallet authentication.

3. EIP-7934 — Max Block Size Cap

Caps blocks at 10 MB to prevent data-flood attacks.

4. EIP-7825 — Per-Transaction Gas Cap

Prevents one transaction from consuming an entire block.


Summary

Fusaka isn’t just an upgrade — it’s a defensive maneuver.

By introducing PeerDAS, BPO forks, and gas limit expansions, Ethereum positions itself as:

  • Cheaper
  • Higher throughput
  • More flexible
  • More competitive against modular DA chains

For 90% of rollups, Ethereum’s DA becomes “good enough,” reducing the economic pressure to migrate to external DA layers.

External DA still matters — but only for applications that require extreme bandwidth (e.g., gaming, AI, fully on-chain compute).

Ethereum has effectively said:

“We can be modular too — and we’re not giving up the DA market.”