The Fusaka Upgrade: How Ethereum Tripled Blob Capacity and Slashed L2 Fees by 60%
Ethereum just completed the most aggressive data throughput expansion in its history — and most users have no idea it happened.
Between December 2025 and January 2026, three coordinated hard forks quietly tripled Ethereum's blob capacity while slashing Layer-2 transaction fees by up to 60%. The upgrade, codenamed Fusaka (a portmanteau of "Fulu" and "Osaka"), represents a fundamental shift in how Ethereum handles data availability — and it's only the beginning.
From Bottleneck to Breakthrough: The Blob Revolution
Before Fusaka, every Ethereum validator had to download and store 100% of blob data to verify its availability. This created an obvious scalability ceiling: more data meant more bandwidth requirements for every node, threatening the network's decentralization.
Fusaka's headline feature, PeerDAS (Peer Data Availability Sampling), fundamentally restructures this requirement. Instead of downloading complete blobs, validators now sample just 8 of 128 columns — roughly 6.25% of the total data — using cryptographic techniques to verify the rest is available.
The technical magic happens through Reed-Solomon erasure coding: each blob is mathematically extended and split into 128 columns distributed across specialized subnets. As long as 50% of columns remain accessible, the entire original blob can be reconstructed. This seemingly simple optimization unlocks an 8x theoretical increase in blob throughput without forcing nodes to scale their hardware.
The BPO Fork Sequence: A Masterclass in Careful Scaling
Rather than shipping everything at once, Ethereum's core developers executed a precise three-part rollout:
| Fork | Date | Target Blobs | Max Blobs |
|---|---|---|---|
| Fusaka | December 3, 2025 | 6 | 9 |
| BPO-1 | December 17, 2025 | 10 | 15 |
| BPO-2 | January 7, 2026 | 14 | 21 |
This Blob-Parameter-Only (BPO) approach allowed developers to collect real-world data between each increment, ensuring network stability before pushing further. The result? Blob capacity has already more than tripled from pre-Fusaka levels, with core developers now planning BPO-3 and BPO-4 to reach 128 blobs per block by mid-2026.
Layer-2 Economics: The Numbers That Matter
The impact on L2 users is immediate and measurable. Before Fusaka, average L2 transaction costs ranged from $0.50 to $3.00. Post-upgrade:
- Arbitrum and Optimism: Users report transaction costs of $0.005 to $0.02
- Average Ethereum gas fees: Dropped to approximately $0.01 per transaction — down from $5+ during peak 2024 periods
- L1 batch submission costs: Reduced by 40% for L2 sequencers
The ecosystem-wide statistics tell a compelling story:
- L2 networks now process approximately 2 million daily transactions — double Ethereum mainnet volume
- Combined L2 throughput has exceeded 5,600 TPS for the first time
- The L2 ecosystem handles over 58.5% of all Ethereum transactions
- Total Value Secured across L2s has reached approximately $39.89 billion
The EOF Saga: Pragmatism Over Perfection
One notable absence from Fusaka tells its own story. The EVM Object Format (EOF), a sweeping 12-EIP overhaul of smart contract bytecode structure, was removed from the upgrade after months of heated debate.
EOF would have restructured how smart contracts separate code, data, and metadata — promising better security validation and lower deployment costs. Supporters argued it represented the future of EVM development. Critics called it over-engineered complexity.
In the end, pragmatism won. As core developer Marius van der Wijden noted: "We don't agree, and we're not coming to an agreement about EOF anymore, and so it has to go out."
By stripping EOF and focusing exclusively on PeerDAS, Ethereum shipped something that worked rather than something that might have been better but remained contentious. The lesson: sometimes the fastest path to progress is accepting that not everyone will agree.
Network Activity Responds
The market has noticed. On January 16, 2026, Ethereum L2 networks recorded 2.88 million daily transactions — a new peak driven by gas fee efficiency. The Arbitrum network, specifically, has seen its sequencer throughput reach 8,000 TPS under stress tests following its "Dia" upgrade optimized for Fusaka compatibility.
Base has emerged as the clear winner in the post-Fusaka landscape, capturing the majority of new liquidity while many competing L2s have seen their TVLs stagnate. The combination of Coinbase's distribution advantage and sub-penny transaction costs has created a virtuous cycle that other rollups struggle to match.
The Road to 10,000 TPS
Fusaka is explicitly positioned as a stepping stone, not a destination. The current roadmap includes:
June 2026: Blob count expansion to 48 through continued BPO forks
Late 2026 (Glamsterdam): The next major named upgrade, targeting:
- Gas limit increases to 200 million
- "Perfect parallel processing" for transaction execution
- Further PeerDAS optimizations
Beyond: The "Hegota" fork slot, expected to push scaling even further
With these improvements, L2s like Base project they can reach 10,000-20,000 TPS, with the entire combined L2 ecosystem scaling from current levels to over 24,000 TPS.
What This Means for Builders
For developers and infrastructure providers, the implications are substantial:
Application Layer: Sub-penny transaction costs finally make microtransactions viable. Gaming, social applications, and IoT use cases that were economically impossible at $1+ per transaction now have breathing room.
Infrastructure: The reduced bandwidth requirements for node operators should help maintain decentralization as throughput scales. Running a validator no longer requires enterprise-grade connectivity.
Business Models: DeFi protocols can experiment with higher-frequency trading strategies. NFT marketplaces can batch operations without prohibitive gas costs. Subscription models and per-use pricing become economically feasible on-chain.
The Competitive Landscape Shifts
With L2 fees now competitive with Solana (often cited at $0.00025 per transaction), the narrative that "Ethereum is too expensive" requires updating. The more relevant questions become:
- Can Ethereum's fragmented L2 ecosystem match Solana's unified UX?
- Will bridges and interoperability improve fast enough to prevent liquidity balkanization?
- Does the L2 abstraction layer add complexity that drives users elsewhere?
These are UX and adoption questions, not technical limitations. Fusaka has demonstrated that Ethereum can scale — the remaining challenges are about how that capacity translates to user experience.
Conclusion: The Quiet Revolution
Fusaka didn't make headlines the way The Merge did. There were no dramatic countdowns or environmental impact debates. Instead, three coordinated hard forks over six weeks quietly transformed Ethereum's economics.
For users, the difference is tangible: transactions that cost dollars now cost pennies. For developers, the playground has expanded dramatically. For the broader industry, the question of whether Ethereum can scale has been answered — at least for the current generation of demand.
The next test comes later in 2026, when Glamsterdam attempts to push these numbers even higher. But for now, Fusaka represents exactly what successful blockchain upgrades should look like: incremental, data-driven, and focused on real-world impact rather than theoretical perfection.
BlockEden.xyz provides enterprise-grade RPC nodes and indexing infrastructure for Ethereum and all major L2 networks. As the ecosystem scales, we scale with it. Explore our API marketplace to build on infrastructure designed for the multi-rollup future.