Ethereum is executing its most significant strategic reversal since The Merge. After years of championing a "rollup-centric roadmap" that delegated scaling to Layer-2 networks, co-founder Vitalik Buterin declared in February 2026 that this vision "no longer makes sense." The Ethereum Foundation h...
"If you create a 10,000 TPS EVM where its connection to L1 is mediated by a multisig bridge, then you are not scaling Ethereum." — Vitalik Buterin, Ethereum Co-Founder
Ethereum is executing its most significant strategic reversal since The Merge. After years of championing a "rollup-centric roadmap" that delegated scaling to Layer-2 networks, co-founder Vitalik Buterin declared in February 2026 that this vision "no longer makes sense." The Ethereum Foundation has responded with an aggressive two-fork 2026 agenda — Glamsterdam in H1 and Hegotá in H2 — that will enshrine proposer-builder separation into the protocol, push gas limits toward 200 million, lay the groundwork for ZK-EVM validation, and add forced transaction inclusion to harden censorship resistance.
The timing is not coincidental. Ethereum's mainnet fee revenue has collapsed over 90% since the Dencun upgrade, with daily gas fees falling from peaks above $23 million to roughly $500,000. Layer-2 networks like Coinbase's Base earned $94 million in profit while returning just $4.9 million to the mainnet. ETH trades below $2,000 — down 36% year-to-date — its worst start to a year on record. With $3.8 billion in ETF outflows over five weeks and a Fear & Greed Index at 11, Ethereum's leadership has concluded that the network must scale itself or risk becoming a settlement layer that nobody pays to use.
This report analyzes the technical architecture of Ethereum's 2026 upgrade slate, the economic logic behind the L1 pivot, and what it means for the $30+ billion Layer-2 ecosystem that built its business model on Ethereum's previous scaling philosophy.
The numbers tell a stark story. Ethereum's mainnet fee revenue has experienced a structural collapse since the March 2024 Dencun upgrade introduced blob transactions. Where L2 networks once contributed $113 million annually to Ethereum's mainnet in 2024, that figure plummeted to approximately $10 million in 2025 — a decline exceeding 90%.
The mechanism is straightforward: blobs made L2 data posting radically cheap, which was the explicit design goal. But the economic consequence was that L2s captured nearly all the value they generated. Base, Coinbase's flagship L2, exemplifies the dynamic — $94 million in profit against $4.9 million in blob fees returned to the settlement layer. Arbitrum and Optimism followed similar patterns. Ethereum's average transaction fees dropped to approximately $0.14 as of late January 2026, the lowest since 2017.
Meanwhile, the macroeconomic environment has turned hostile. Trump's February 15% global tariff announcement triggered a 5% single-day crypto crash. Bitcoin's worst five-month losing streak since 2018 dragged the entire market, with $2.56–3.2 billion in liquidations and $3.8 billion in ETF outflows over five weeks. ETH, already weakened by the fee revenue collapse and ongoing inflation (0.8% post-Dencun), fell below $2,000 for the first time since 2023.
The Ethereum Foundation's February 18, 2026 "Protocol Priorities Update" should be read against this backdrop. It is not merely a technical document — it is a strategic response to an economic emergency.
Glamsterdam (a portmanteau of Gloas + Amsterdam) headlines two structural changes that address block production centralization and execution bottlenecks.
Enshrined Proposer-Builder Separation (ePBS). Today, Ethereum's block construction relies on an off-chain relay infrastructure — primarily Flashbots' MEV-Boost — where proposers outsource block building to specialized builders through trusted intermediaries. Over 90% of Ethereum blocks flow through this system. ePBS moves this market directly into the consensus layer, replacing the relay-dependent model with an in-protocol commit-reveal mechanism.
The implications are significant. Validators will be able to use nearly the full 12-second slot for block verification, compared to the milliseconds available under the current relay-dependent system. Builder commitments become enforceable at the protocol level, standardizing MEV handoff rules and reducing the trust assumptions that make the current system fragile. Censorship risk declines as block construction becomes more transparent and less dependent on a handful of relay operators.
Block-Level Access Lists (BALs). The second Glamsterdam headliner targets execution efficiency. BALs (EIP-7928) allow parts of a block to specify which state they will access, enabling parallel verification by nodes. Today, nodes must process transactions sequentially because they cannot predict state dependencies in advance. BALs preload the needed state for a block, eliminating repeated disk reads — a major execution bottleneck that currently limits safe gas limit increases.
BALs are further along in development than ePBS, with multiple client implementations already underway. Together, these two proposals form the scaffolding for everything Ethereum wants to do with gas limits and throughput in 2026 and beyond.
The second 2026 fork, Hegotá, will headline the most controversial proposal in Ethereum's recent history: Fork-Choice Enforced Inclusion Lists (FOCIL, EIP-7805).
FOCIL works by pseudorandomly selecting a committee of 16 validators per slot, each of which scans the public mempool and assembles a list of transactions (capped at 8 kilobytes) that should be included in the next block. Attesters then refuse to vote for any block that omits transactions from these lists without valid reason. If a builder censors a FOCIL-listed transaction that could have fit, the block fails to achieve attestation and cannot become canonical.
The proposal was originally considered for Glamsterdam but was pushed back to reduce fork scope. Its inclusion in Hegotá carries strong support from core developers and from Buterin himself, who described his broader vision as building a "cypherpunk principled, non-ugly Ethereum" — a layer that enshrines censorship resistance, ZK compatibility, and efficient consensus directly into the protocol.
Not everyone agrees. Privacy Pools founder Ameen Soleimani has argued that FOCIL creates legal risks for U.S.-based validators who would be protocol-mandated to include transactions they might otherwise be compelled to exclude under sanctions law. The tension between censorship resistance and regulatory compliance remains unresolved.
The Ethereum Foundation's 2026 priorities explicitly target gas limits "toward and beyond 100 million," up from the current 60 million (itself raised from 45 million in November 2025 after validator signaling).
Analysis from Lookonchain and other researchers suggests the gas limit could reach 200 million following ePBS and BALs deployment, effectively tripling L1 throughput without changing slot times. The roadmap envisions a gradual, monitored increase — with defined observation periods between increments — to ensure network stability.
The longer-term vision is more radical. In 2026, ZK-EVM clients are expected to become viable for a portion of the validator set, allowing attestation via zero-knowledge proofs rather than full re-execution. This sets the stage for Buterin's stated target of approximately 10,000 transactions per second on L1 — a figure that, if achieved, would fundamentally alter the economics of when and why applications should deploy on L2s rather than the base layer.
The scaling track is led by Ansgar Dietrichs, Marius van der Wijden, and Raúl Kripalani, who merged the previously separate "Scale L1" and "Scale Blobs" efforts into a single unified program — an acknowledgment that execution capacity and data availability throughput are deeply intertwined and must advance together.
Ethereum's third development track — "Harden the L1" — is anchored by the Trillion Dollar Security (1TS) initiative, a cross-ecosystem effort to upgrade security to the standard required for a network that aspires to secure trillions in value.
The most forward-looking component is post-quantum readiness. Ethereum's core cryptography relies on elliptic curve signatures (secp256k1) that could theoretically be broken by sufficiently powerful quantum computers. While no such machine exists today, the 1TS initiative is designing migration paths to quantum-resistant signature schemes that can be embedded in the protocol years before they become necessary.
In February 2026, the Ethereum Foundation held a dedicated "Trillion Dollar Security Day" at Devconnect, bringing together practitioners from across the stack to assess current security posture and identify near-term priorities. Additional focus areas include post-execution transaction checks, trust-minimized RPC patterns, and hardened client diversity requirements.
Buterin's February declaration that the rollup-centric roadmap "no longer makes sense" sent shockwaves through the L2 ecosystem. His reasoning rested on two pillars: progress toward L2 decentralization has been "slower and more difficult than expected," and Ethereum's own L1 is now cheap enough to scale directly.
The implication is not that L2s become irrelevant — but that their value proposition must evolve. Networks connected to Ethereum through multisig-controlled bridges, Buterin argued, are "not scaling Ethereum" regardless of their throughput. True Ethereum scaling, in his framework, means block space "backed by the full faith and credit of Ethereum" — guaranteed valid, uncensored, and irreversible as long as the base layer functions.
This reframing pressures L2s to accelerate their decentralization timelines or risk being classified as separate networks that merely use Ethereum for settlement. It also opens the door for "native rollups" — execution environments that inherit Ethereum's security properties directly rather than through bridges and fraud proofs managed by centralized operators.
For the L2 industry that now exceeds $30 billion in combined market capitalization, this is an existential redefinition. Networks like Arbitrum, Optimism, and Base must demonstrate genuine alignment with Ethereum's security model — or acknowledge that they are building independent chains with an Ethereum data availability subscription.
Ethereum's 2026 agenda represents the network's most consequential year since The Merge. The simultaneous pursuit of L1 scaling, censorship hardening, and security upgrading — all while ETH trades at multi-year lows — is either a masterclass in building through a downturn or an overambitious engineering sprint that risks execution failure across multiple fronts.
The economic logic is sound. A settlement layer that captures less than 5% of the value generated by its own ecosystem is not sustainable. The Ethereum Foundation's three-track approach — Scale, Improve UX, Harden the L1 — directly addresses this by making L1 competitive on throughput while enshrining the security and censorship resistance properties that theoretically justify L1 fees.
But ambition is not delivery. Glamsterdam's ePBS requires cross-client coordination, extensive testnet validation, and consensus-layer changes that touch the most sensitive part of Ethereum's infrastructure. FOCIL's legal implications for U.S. validators remain unresolved. And the gas limit escalation assumes that BALs and parallel verification will work as theorized under real-world load.
What is clear is that Ethereum has stopped waiting for Layer-2s to solve its scaling problem. Whether the network can solve it on its own — and recapture the economic value that has drained to L2s, competitors, and apathy — will define whether ETH deserves to trade at $2,000 or at multiples of it. The 2026 forks will provide the answer.