Mapping the Strawmap: Ethereum’s Big Course Correction

Key Takeaways

• The Strawmap is the most significant strategic reframing of Ethereum's L1 since The Merge. It marks an explicit course correction away from the rollup-centric scaling approach that defined the 2022-2025 cycle and back toward base-layer scaling as the priority.

• Ethereum is still dominant on metrics that measure capital at rest (TVL, stablecoins, RWAs), but materially behind faster and cheaper chains on metrics that measure active usage (network fees, application fees, DEX volumes). The Strawmap is an attempt to close the latter gap without eroding the credible neutrality that anchors the former.

• The roadmap is organized around five "north stars" through the end of the decade: Fast L1 (sub-second finality); Gigagas L1 (10,000 TPS on the base layer); Teragas L2 (~10 million TPS across rollups); Post-Quantum L1; and Private L1. The Ethereum Foundation frames the Strawmap as "an accelerationist coordination tool," not a prediction, and its later years should be treated as directional rather than scheduled.

• Glamsterdam (H1 2026) is the first planned fork and clears the runway for L1 scaling through enshrined proposer-builder separation and block-level access lists, targeting a post-fork gas limit floor of 200 million versus the current 60 million. Hegotá (H2 2026, timing uncertain) is anchored by FOCIL for protocol-enforced censorship resistance, with native account abstraction, Verkle trees, and optional execution proofs among possible additional features.

• The longer-horizon zkEVM transition is the technical foundation for the gigagas L1 north star and is advancing rapidly. Real-time proving latency has already dropped 45x in two years. Post-quantum cryptography work is tightly coupled with the same zk research, meaning progress on one front advances the other, and positions Ethereum to potentially ship quantum resistance before Bitcoin.

• The Strawmap does not directly address ETH value accrual, and the gap between the technical roadmap and the token thesis is the central tension for investors. The Strawmap addresses the supply side of the ETH thesis across every lens. Whether demand materializes to fill the capacity it creates remains the open question. AI agent activity and tokenized assets are the most concrete near-term candidates.

• Execution risk is the binding constraint. Sequencing across the seven planned forks is deliberate and dependencies are real, meaning any delays will compound. The credibility of the post-2026 timeline rests heavily on Glamsterdam and Hegotá shipping on schedule.

Introduction

Ethereum entered 2026 in a challenging position. The 2022-2025 cycle was the first in which it faced genuine competition from other general-purpose smart contract blockchains that not only delivered faster speeds and cheaper costs but achieved genuine adoption. Ethereum’s failure to scale the base layer (L1) while prioritizing layer-2 networks (L2s) drove user activity, and in turn developer interest, elsewhere. While L2s received some of this interest, Solana (a non-EVM chain) was the primary beneficiary.

Despite these headwinds, Ethereum remains the default general-purpose smart contract chain for asset issuance and settlement, evidenced in continued dominance on metrics like total value locked (TVL), and stablecoin and real-world asset (RWA) issuance. But in terms of active usage, Ethereum has lost substantial ground to faster and cheaper alternatives measured in network fees, application fees, and decentralized exchange (DEX) volumes. The result, while native token ETH has maintained its status as the second largest cryptocurrency by market cap uninterrupted since 2018, it has seen its overall market cap share decline.

The Ethereum community is well aware of these trends. Over the past year, its leadership and core developers have responded with a marked increase in shipping cadence and a sweeping reassessment of the protocol's direction. The Strawmap, released in March 2026, is the culmination of these changes and the most significant evolution of Ethereum's L1 since The Merge, when it transitioned to proof of stake in 2022.

This report examines the Strawmap's technical vision, assesses the near-term forks that will test Ethereum's ability to deliver on its ambitions, and evaluates what the roadmap means for the protocol's competitive positioning in the coming decade.

Ethereum’s Asset Issuance Moat

Ethereum's TVL market share has held remarkably steady at roughly 55%-60% since mid-2022, even as ETH's price has underperformed BTC and other L1 native tokens.

TVL alone is a highly imperfect metric. On the surface it tells little, given its reliance on token price fluctuations, recursive leverage (the same ETH can appear multiple times across staking, lending, and restaking), and the lack of a standardized methodology across data providers. That said, the composition of Ethereum's TVL and the structural reasons its maintained dominance is more informative than the headline number.

Lending applications, for example, account for ~$23 billion of Ethereum TVL, approximately 51% of the total (down from ~$32 billion and nearly 60% earlier in 2026, a direct result of KelpDAO exploit). Lending, and DeFi more broadly, was one of the first crypto application verticals to find product-market-fit and is a major revenue driver across the crypto landscape. Liquidity begets liquidity, and Ethereum's dominance persists here because of the depth of its collateral markets, oracle infrastructure, and surviving multiple market crashes. This creates a trust premium that newer chains can't quickly replicate, particularly for the largest allocators whose risk tolerance is the lowest.

Even in the wake of the KelpDAO exploit, Ethereum's lending layer demonstrated resilience. The Aave lending market's TVL fell nearly $10 billion within a week, yet a coordinated "DeFi United" relief initiative assembled more than 14 ecosystem participants (including Mantle, Aave DAO, Lido, EtherFi, and many others) and raised over $160 million in ETH to cover the shortfall within days.

Stablecoin issuance tells a similar story. Ethereum still hosts approximately 50% of all stablecoin market cap, down from near-100% before 2019, but stable over the last two years. The primary share taker has been Tron, not Solana, reflecting Tron's dominance in emerging market payments and peer-to-peer transfers rather than DeFi competition. Solana and Base (an Ethereum L2) have recently emerged as visible but still small slivers.

Ethereum also hosts over 60% of all tokenized real-world assets. While Ethereum's share declined from 92% in April 2024 to 63% by April 2026, the absolute value of RWAs issued on Ethereum grew more than 12x over the same period. The share compression reflects a broader market expansion as competing chains grew 71x off an extremely small base. But issuers have not migrated away from Ethereum, which retains the institutional trust premium (tokenized treasuries, private credit, DeFi-integrated RWA) while other chains absorb volume at the margin. This is arguably some of the stickiest capital on Ethereum, because institutional RWA issuers choose a chain through months of legal review, custodian integration, and compliance sign-off.

Activity Migrates Elsewhere

The picture reverses when you look at activity-based metrics. Network-level fees, the metric most directly tied to ETH's value accrual through EIP-1559 burns, have collapsed. Ethereum's overall share of total crypto network fees has fallen from over 90% in early 2020 to a range of 10%-20% today.

Solana and Tron now individually generate comparable or greater network fees than the Ethereum L1. The network’s decline reflects two overlapping forces. First, activity migrated as Ethereum L1 transactions remained expensive and slow relative to alternatives at precisely the moment when the cycle's dominant activity (memecoin and perpetuals trading) demanded cheap and fast execution. Users went where the UX matched the use case. Second, Ethereum deliberately reduced its own fees. EIP-4844 cut L2 data posting costs by roughly 99%, and successive gas limit increases expanded L1 capacity, reducing per-transaction costs even for activity that stayed on the base layer. Lower fees per transaction mean less ETH burned per transaction, even at constant volume. The less ETH burned, the more supply remains on the market, which, all else equal, weighs on the price.

Application-layer fees, fees generated by applications on a blockchain rather than the protocol itself, show a similarly steep decline. Ethereum held near-total dominance through early 2021 but has since fallen below 33% with fee generation fragmenting rapidly across Solana, Binance Smart Chain, Base, and Hyperliquid. Application fees reflect genuine user demand. They fill a chain's blockspace and drive base-layer fee revenue (and depending on the urgency of the transaction, additional priority fees that can juice staking yields for validators). Critically, they attract the developer talent that sustains an ecosystem. Teams build where users pay.

Application fees reflect genuine user demand and attract the dev talent that sustains an ecosystem. Teams build where users pay.

Again the same story is evident in DEX volume. Ethereum's share has fallen from 60%-70% to approximately 20%, with the most dramatic shift occurring since late 2024. Solana has surged from negligible DEX volume to roughly 25%-30% market share in under 18 months, briefly exceeding Ethereum. Base and Arbitrum (Ethereum L2s) have also captured meaningful share. This complicates the competitive narrative. The Ethereum ecosystem isn't shrinking as fast as the L1 charts suggest, but less value is flowing back to ETH.

The Ethereum ecosystem isn't shrinking as fast as the L1 charts suggest, but less value is flowing back to ETH.

It is worth being clear about what the activity migration does and does not imply. Some L2s have found genuine product-market fit, particularly with institutions that want the security and liquidity of Ethereum while retaining the ability to customize execution environments and meet regulatory requirements. Base, Arbitrum, and a growing set of permissioned or app-specific L2s are real businesses with real users, and they remain the most viable path for many of the use cases that institutions want to deploy. The critique that follows is not of L2s themselves but of a roadmap that prioritized L2 scaling at the expense of L1 throughput and, in doing so, broke ETH's value accrual mechanics without delivering the unified user experience the rollup-centric vision promised. The Strawmap reframes L1 scaling as complementary to the L2 ecosystem rather than competitive with it. A larger, faster L1 restores fee revenue and burn while continuing to serve as the settlement and data availability layer that L2s depend on.

The Contradiction

These charts present a contradiction at the center of Ethereum's current position. ETH dominates metrics that measure capital at rest but is losing on metrics that measure active transactional usage. ETH the asset has suffered because of the latter.

The divergence makes sense once you understand what drives ETH's token economics. EIP-1559, implemented in Ethereum's August 2021 London hard fork, burns base fees from every L1 transaction. As a result, more L1 usage means more ETH burned, creating deflationary pressure. ETH is currently inflating at ~0.8% annualized, though cumulative supply growth since the Merge (the 2022 switch from proof-of-work to proof-of-stake consensus) is just 0.27%. Supply has expanded by roughly 1.1 million ETH since the Merge, erasing the deflationary narrative that had been a core part of the old “ETH as sound money” thesis.

Ethereum is, in short, a network that retains the trust of capital but has lost the attention of users. The Strawmap is an explicit attempt to win that attention back by scaling the ecosystem to the point where the cost and speed disadvantages that drove users to competing chains no longer apply.

The Ethereum network retains the trust of capital but has lost the attention of users. The Strawmap aims to win it back.

Setting the Stage: What 2025 Delivered

Ethereum's protocol development entered a new phase in 2025 with the delivery of two major hard forks in a single calendar year, the first time that had happened since 2021. Pectra, activated in May, delivered improvements to wallet functionality, validator economics, and data “blob (binary large object) capacity. Fusaka followed in December with a focus on scaling through data availability sampling for blob verification, raising the gas limit, and establishing mechanisms to increase blob throughput and stabilize blob fee revenue.

Pectra improved the user and validator experience. Fusaka attempted to address the value accrual crisis that had plagued Ethereum since the Dencun upgrade in March 2024, when the introduction of cheap blob space for L2 data posting caused L1 fee revenue to collapse, breaking ETH's deflationary narrative even as ecosystem usage grew. Since 2024, Ethereum chain fees have increased quarter-over-quarter only twice.

In retrospect the 2025 updates did little to move the needle on Ethereum activity and value accrual. They prioritized feature releases and L2 scalability over L1 scalability. However, they did demonstrate that Ethereum core developers were capable, and increasingly focusing on, shipping at faster cadence to ensure Ethereum’s continued relevance and technical primary versus emerging competitors.

Ethereum’s Strawmap Is Unveiled

In February 2026, the Ethereum Foundation published its L1 Strawmap, a unified visual roadmap for Ethereum's Layer-1 protocol development through the end of the decade. The document originated from an internal EF workshop in January 2026.

Ethereum Strawmap

The Strawmap marks the first time the Ethereum Foundation has placed its multi-year technical vision on a single page. Previously, Ethereum's roadmap existed as a collection of Ethereum creator Vitalik Buterin’s blog posts, researcher tweets, and ethereum.org diagrams that covered individual development tracks. The EF is explicit that the Strawmap is "not a prediction" but rather "an accelerationist coordination tool, sketching one reasonably coherent path among millions of possible outcomes." It assumes human-first development, with a caveat that AI-accelerated R&D could significantly compress timelines.

These qualifications are important and should be taken seriously. Ethereum's roadmap has changed frequently and will almost certainly change again in response to shifting dynamics. For this reason, in addition to the Strawmap’s structure, below we only cover the nearest-term upcoming hard forks planned for 2026, Glamsterdam and Hegotá, as well as the ultimate objectives Ethereum core developers are building toward.

Architecture of the Roadmap

The Strawmap organizes development along three layers, each of which handles a distinct function.

The Consensus Layer governs how the network agrees on the canonical chain, including block proposals, attestations, and finality. The Data Layer ensures that the data required to verify the chain is available and accessible, including blob data that Layer-2 rollups post to Ethereum. And the Execution Layer is where transactions are processed, including the Ethereum Virtual Machine (EVM), smart contract execution, state transitions, and gas metering.

At the highest level, Ethereum developers aim to scale (increase throughput, reduce costs), improve UX (for developers and users), and harden the L1 (enhance security, censorship resistance, and decentralization) through development of these layers.

Five North Stars

The Strawmap articulates five "north stars," end-state goals that define what Ethereum should look like by the end of the decade. They provide a directional framework for achieving scale, improving UX, and hardening the L1.

Fast L1 – Finality in Seconds: Ethereum produces blocks every 12 seconds with finality taking approximately 13 minutes. The target is block times as low as two seconds (although exact targets remain unclear) and finality in 6-16 seconds. The motivation here is to make the L1 viable for high-demand applications and institutional settlement where finality matters most.

Gigagas L1 – 10,000 TPS on the Base Layer: Today Ethereum processes roughly 15-30 transactions per second (TPS) on the L1 and has a 60 million gas limit for 12-second blocks. Attaining 1 gigagas per second (1 billion gas units) and 10k TPS represents a monumental increase in L1 capacity, equating to a 200x increase in the gas limit. Scaling to these levels is dependent on breakthroughs in zero-knowledge (zk) computation that would enable the use of zk for proving (often referred to as zkEVM and discussed in further detail below).

This is arguably the north star with the most direct implications for ETH as an asset because it makes the Ethereum L1 a viable platform for onchain activity at scale (a far cry from the days when an NFT mint led transactions to cost hundreds of dollars). More L1 transactions means more gas burned through EIP-1559 and the potential for ETH to again become disinflationary. Success, however, depends heavily not just on technical execution, but also on Ethereum as a platform generating the demand necessary to meet the TPS required.

Teragas L2 – 10 Million TPS Across Rollups: While gigagas addresses L1 computation, teragas addresses data availability for L2s to increase throughput. It targets 1 gigabyte/second of blob data, enough for ~10m TPS across the rollup ecosystem. The role of L2s in the Ethereum ecosystem is in the process of being redefined. The EF's Platform team published a reframing in March 2026 acknowledging that L2s previously existed primarily to scale Ethereum, but now exist to offer differentiated capabilities (custom features, compliance environments, and specialized execution that the L1 cannot provide). Vitalik has been blunter, recently saying the original rollup-centric vision where L2s were "branded shards" of Ethereum "no longer makes sense," and L2s that exist only to be "Ethereum but cheaper" need to find a different value proposition as the L1 scales. At the end of March, during the EthCC conference, a new initiative, the Ethereum Economic Zone (EEZ), was announced which aims to enable synchronous composability between mainnet and connected rollups without bridges. Again, beyond technical challenges, the primary question here remains if there will be demand commensurate with the massive increase in supply.

Post-Quantum L1 – Centuries of Cryptographic Security: Ethereum's cryptography relies on mathematical problems quantum computers could theoretically solve. This north star envisions hash-based alternatives secure against quantum attacks. The risk creates urgency for high-value assets on Ethereum and for institutions evaluating the chain as long-term infrastructure. Notably, the EF's current zk research is tightly coupled with post-quantum work, meaning progress on the gigagas path simultaneously advances quantum resistance (discussed in further detail below).

Private L1 – Privacy as Default Infrastructure: Today most Ethereum transaction are public. This north star envisions protocol-level shielded transfers where sender, receiver, and amount are cryptographically hidden while remaining verifiable. The EF's position is that privacy should be the default, not an opt-in feature. This represents both a philosophical commitment to self-sovereignty and a practical recognition that privacy is a prerequisite for real-world adoption, particularly among institutions loath to tip their hands to competitors. The EF's recently established DeFi team has explicitly positioned privacy as "base infrastructure: first for payments of all tokens, then for more complex use cases like trading and lending."

The Ethereum Foundation's position is that privacy should be the default, not an opt-in feature.

Glamsterdam: Clearing the Runway (H1 2026)

Glamsterdam, targeted for the first half of 2026, is the next fork on the roadmap. Its two headliners are enshrined proposer-builder separation (Consensus Layer) and block-level access lists (Execution Layer).

Enshrined Proposer-Builder Separation (EIP-7732)

Every 12 seconds, a randomly selected Ethereum validator is chosen to propose the next block. In practice, nearly all validators outsource the actual construction of blocks to specialized firms called builders, who compete to harvest maximal extractable value (MEV) through sophisticated transaction ordering. This outsourcing is commonly referred to as proposer-builder separation (PBS) and implemented through MEV-Boost, off-chain software built by Flashbots (for a full overview of the PBS architecture, refer to Galaxy Research’s prior work here).

While PBS has been an effective method for increasing validator profitability and block-building efficiency, it imposes constraints on Ethereum scaling and introduces centralization concerns. Today, over 90% of blocks are primarily routed through three relays, creating a centralization dependency with no protocol-level fallback.

The bigger issue is a scaling bottleneck in the protocol itself. An Ethereum block bundles consensus data and transaction data together as a single object. Validators must download, execute, and vote on the block all within roughly the first four seconds of a slot so that their attestation can be aggregated and propagated during the remaining eight seconds. That caps how many transactions can fit in a block. If you increase the block size (gas limit), validators can't finish processing in time, so the system breaks.

EIP-7732 aims to fix this by splitting the block into two pieces that are transmitted and processed at different times. In the first phase of the slot, the proposer publishes a lightweight consensus block (which is only 5%-10% of total block size). This is a commitment to a specific builder's bid. Validators can attest to this header quickly without seeing any transaction data.

In the second phase, the builder reveals the full transaction payload. A separate validator committee (the payload timeliness committee) confirms the execution data arrived on time, without validating its contents. Validators then have a much longer window, roughly 9 to 12 seconds instead of 4, to download, verify, and propagate the payload. The trust model shifts from "verify then attest" to "attest then verify.” Builders are economically committed to providing valid payloads through collateralized bids, so the system no longer depends on validators finishing execution before casting their votes. The same expansion applies to blob data distribution for L2s. The expected result is a 4-6x increase in the time available for processing transactions, enabling higher gas limits and creating the proving window needed for the eventual zkEVM transition.

ePBS also removes relays from the critical path while making the builder market protocol-governed and transparent. However, ePBS does not solve builder concentration itself. The same sophisticated builders that dominate today will likely continue to dominate, because their competitive advantage comes from offchain capabilities (MEV algorithms, order flow relationships, latency infrastructure) that the protocol cannot equalize. The FOCIL proposal (discussed in the Hegotá section) attempts to further constrain builder power, though with important caveats about its scope.

Whether ePBS eliminates relay dependency in practice remains an open question. Ethereum researchers, including several ePBS authors, have acknowledged that the economic incentives to use relays may persist, if builders prefer private channels and proposers favor relays' optimization services.

Nevertheless, at its core ePBS scales Ethereum slot pipelining. It will be enforced at the protocol level regardless of whether external relays continue to operate alongside it. It is also worth noting that ePBS originated as a decentralization improvement focused on block-building market structure and has since evolved into primarily an L1 scaling mechanism, a clear marker of how Ethereum's development priorities have shifted.

Block Access Lists (EIP-7928)

Ethereum processes transactions one at a time. When a validator receives a block, it executes each transaction sequentially because it doesn't know in advance which accounts or data each transaction will touch. Two transactions might modify the same account balance. Running them simultaneously could produce the wrong result. So, the safe default has always been serial execution.

This creates two bottlenecks. On the compute side, no matter how high the gas limit is raised, execution speed is limited by how fast a single thread can process transactions in order. On the data retrieval side, each transaction may require the validator to read data from disk (such as account balances), and it can't start those reads until the previous transaction finishes because it doesn't yet know what the next transaction will need.

EIP-7928 requires each block to include a complete rundown of every account and storage slot that each transaction in the block will touch, along with the resulting state changes. With this information available before execution begins, validators can do several things simultaneously that were previously sequential. They can prefetch all needed data from disk in parallel, rather than waiting for each transaction to trigger its own reads. They can identify which transactions are independent (historical analysis suggests 60%-80% of transactions in a typical block have no state overlap) and execute those in parallel. And they can compute the resulting state root in parallel rather than sequentially. The execution pipeline transforms from "sequential reads plus sequential compute" to "parallel reads plus parallel compute."

Parallel execution has become a de facto standard for high-performance blockchains. Ethereum has been the outlier by lacking it.

Parallel execution has become a de facto standard for high-performance blockchains. Solana pioneered it from inception. Newer chains including Monad and MegaETH have also made it a central design feature. Ethereum has been the outlier. The key architectural difference from chains like Solana is that block access lists (BALs) on Ethereum will operate at the block level rather than the transaction level. The builder assembles the access list for the entire block. Individual users and developers won't need to change how they construct transactions. This should make adoption less disruptive so that no application layer changes required but places more responsibility on the builder.

BALs should make gas limit increases more practical. Bigger blocks mean more computation per slot. Without parallel execution, processing them depends on single-thread performance, which pushes hardware requirements up. BALs will let commodity multi-core machines handle the additional load. BALs also have a less obvious but equally important connection to the zkEVM path. Today, a ZK prover must execute every transaction sequentially to generate the witness (execution trace) needed for proof generation. BALs provide a deterministic dependency graph before execution begins, which will allow provers to parallelize witness generation and shard proof work across independent transaction clusters.

While the implementation of BALs is not the single gating dependency for gigagas (that distinction belongs more directly to zkEVM integration and real-time proving), they materially accelerate the proving pipeline that gigagas ultimately depends on.

Other Notable Changes

Fast Confirmation Rule. Originating from EF research in 2024 and now being implemented across consensus clients, this mechanism provides a strong guarantee against transaction reversion after a single slot (~13 seconds), thereby reducing bridge and exchange deposit times by 80%-98%. It does not require a hard fork and can roll out independently of Glamsterdam.

Gas Repricing. A suite of EIPs aim to realign gas costs, which haven't been updated in years, with actual computational costs on modern hardware. The net effect is an estimated 70%-80% reduction in gas fees for common transaction types ranging from simple transfers to DeFi interactions, while increasing costs for currently underpriced operations that could become attack vectors at higher gas limits. The repricing also serves a less obvious purpose of aligning gas costs with proving costs for the eventual zkEVM transition. The central piece is EIP-8037, which raises the cost of writing new state so that a higher gas limit doesn't translate into unbounded state growth.

Ethereum After Glamsterdam

Together, ePBS and BALs promise to restructure the block lifecycle. ePBS restructures how blocks are built and proposed, expanding the time available for processing from ~2 seconds to 9-12 seconds. BALs restructure how blocks are executed and verified, making that expanded time dramatically more productive through parallelization.

Glamsterdam targets a post-fork gas limit floor of 200 million, more than 3x the current 60 million ceiling. That number was firmed up at the Soldøgn interop in Svalbard the week ending May 2, 2026, where roughly 100 core contributors aligned on 200M as the credible target. Final approval still runs through AllCoreDevs calls, but the working number is no longer aspirational. While a far cry from the throughput and costs of the leading high-performance chains, Glamsterdam represents a meaningful step forward and a clear signal of core developers' renewed focus on scaling the L1.

Hegotá (H2 2026, Timing Uncertain)

Hegotá's only confirmed headliner at the time of writing is FOCIL (EIP-7805). The rest of the fork's scope remains under discussion. Discussion opened last month for non-headliner proposals.

Timing is also uncertain. As of the EF's April 2026 checkpoint, Glamsterdam itself is proving "trickier and more slow-going than anticipated," and Hegotá's timeline depends on Glamsterdam shipping first. Whether both forks land in 2026 is an open question.

Fork-choice Enforced Inclusion Lists (EIP-7805)

Ethereum's block production is highly concentrated. Over 90% of blocks are built by a small number of specialized builders through the MEV-Boost relay system. This creates a risk of censorship because builders can selectively exclude transactions from blocks. While individual participants may be legally compelled by governments to exclude certain transactions, widespread censorship undermines Ethereum’s value proposition of a credibly neutral, open financial system.

While governments may compel individual participants to exclude certain transactions, widespread censorship undermines Ethereum’s value proposition of a credibly neutral, open financial system.

ePBS addresses part of this problem by bringing block building onchain and removing relays from the critical path. But ePBS alone doesn't prevent a builder from censoring transactions. It just makes the market more transparent, and as discussed above is primarily a scaling mechanism. Fork-choice enforced inclusion lists (FOCIL) would add the enforcement layer.

The mechanism works across two slots. During slot N-1, a randomly selected Inclusion List (IL) committee observes the public mempool and each member independently publishes a signed list of valid pending transactions they believe should be included. These lists are broadcast over the P2P network. Once submitted, a builder aggregates the union of all flagged transactions and must include them in their proposed block for the given slot N. When the block is released, the full attester set (not just the IL committee) checks whether the payload satisfies the inclusion list conditions before voting. If the builder excluded flagged transactions without valid justification (the transaction became invalid; the block is full), attesters withhold their votes, and the block is not included in the chain.

The design is committee-based and conditional. Multiple randomly selected validators contribute overlapping lists rather than a single proposer dictating inclusion. This makes it far harder for any one actor to manipulate what transactions get included. FOCIL doesn't guarantee that any specific transaction lands in any specific block. Transactions that become invalid between the IL freeze and block production are legitimately excludable. What it does is make sustained, systematic censorship economically irrational. If at least one committee member is honest and connected to the public mempool, censored transactions should reach the chain within one to two slots.

FOCIL reflects a broader convergence across blockchains toward protocol-enforced censorship resistance. Solana is pursuing the same goal through the R&D lab Anza’s recently released Constellation plan. Constellation introduces a multiple concurrent proposers (MCP) protocol that would allow roughly 16 randomly selected proposers to simultaneously submit transactions every 50 milliseconds, with attesters enforcing inclusion through quorum requirements.

The approaches differ in architecture. FOCIL is a lighter-weight addition to Ethereum's ePBS structure, while Constellation is a fundamental redesign of Solana's block production. Both signal that the industry has converged on the same conclusion. Transaction inclusion guarantees need to be enforced at the protocol level, not left to the goodwill of block producers.

Other Candidates

Frame Transactions (EIP-8141). Native account abstraction is widely considered one of the most consequential pending upgrades for Ethereum UX. Today, interacting with Ethereum requires holding ETH for gas and managing a private-key-based externally owned account (EOA), a friction point that has limited mainstream onboarding. Native account abstraction would enable custom verification logic, wallet recovery without seed phrases, gas payments in stablecoins (removing the ETH-or-nothing onramp), and post-quantum signature migration on a per-wallet basis. Vitalik championed this plan as a Hegotá headliner but Ethereum developers have downgraded it to "considered for inclusion" after client teams argued the complexity risked delaying the fork. This is the latest in a pattern of account abstraction deferrals dating to 2016.

Quick Slots (EIP-8198). This proposal makes slot duration a runtime configuration and uses that infrastructure to reduce slot time from 12 second to eight. The 8-second target is aspirational. The EIP explicitly phases the work into building infrastructure, characterizing bottlenecks, and iteratively reducing. Gas limits scale proportionally, so this is a latency improvement, not a capacity upgrade. Whether it lands in Hegotá or waits a later upgrade is unclear.

Verkle Trees. Replaces Ethereum's Merkle Patricia Trie with a structure that reduces proof sizes by ~90%, enabling stateless verification. Verkle trees are a precondition for zkEVM integration. They were deferred from multiple prior forks due to migration complexity.

Optional Execution Proofs (EIP-8025). Validators would opt into proof-generating or proof-verifying modes, a testing ground for the zkEVM transition (discussed in further detail below).

Beyond 2026: The Gigagas Path and Prover Architecture

The Strawmap extends through forks I*, J*, and beyond to the end of the decade, but specifics become increasingly speculative past Hegotá. Per above, even the majority of the specific Hegotá implementations remain unclear. One of the most consequential long-term developments is the introduction of the zkEVM, the technical foundation of the gigagas L1 north star, and the proving infrastructure on which Ethereum's post-quantum migration will be built.

A New Architecture

Today, every Ethereum validator independently re-executes every transaction in every block to verify correctness. This redundant computation is a scalability constraint, whereby increasing validator computational capacity results in more expensive validator infrastructure that reduces the number of individuals who can run a validator.

The zkEVM aims to strike a balance through a new architecture with four distinct roles:

• Builders (an existing role) construct blocks by ordering transactions. They also assemble a witness representing the slice of state data each transaction will read, bundled with cryptographic proofs tying that data to the parent block's state root. Under ePBS, the builder market moves onchain.

• Proposer (existing) select a winning builder's bid and publish a lightweight consensus commitment to that block. The builder then reveals the full payload to the network. Proposers no longer need to execute blocks to propose them. They attest to the commitment, and execution is handled downstream.

• Provers (a new role) execute the block against the witness and generate a zero-knowledge proof that the execution was done correctly and produces the claimed new state root. This is computationally intensive, requiring specialized GPU-heavy hardware. The EF targets a maximum capex of $100,000 and power draw of 10 kW for on-premises proving. Only one prover needs to be live for the network to function, and the role is expected to be served by a competitive market where provers are compensated through transaction fees, MEV sharing, and incentives.

• Validators (existing) verify the proof and attest to the block. Proof verification takes milliseconds on consumer hardware regardless of block complexity. Validators no longer need to run a full execution layer client or store the chain's 100GB+ state on disk. In effect, validators become "zkAttesters," verifying proofs rather than re-executing transactions. This lowers the barrier to entry for validators, protecting the diversity of the validator set even as the gas limit scales dramatically.

The EF's goal is full, uncompromising EVM-equivalence, called a "Type 1" zkEVM. This means existing applications, developers, and tooling require zero modifications. Some competing approaches sacrifice compatibility for proving efficiency, but Ethereum is explicitly choosing the harder path to avoid fragmenting its application ecosystem.

The Proving Challenge

The EF's zkEVM team is focused on making zero-knowledge virtual machine technology performant enough for mainnet use. The target is "real-time proving," generating proofs within the slot time so that the proving step doesn't delay finality. The EF's working definition sets the current bar at under 10 seconds latency for 99% of blocks, at least 128-bit cryptographic security, proof sizes under 300 KB, and no trusted setups. Proving latency has dropped from 16 minutes to 16 seconds over the past two years, a 45x improvement. In November 2025, the SP1 Hypercube (a zero-knowledge virtual machine developed by Succinct Labs) achieved real-time proving of 99.7% of Ethereum blocks in under 12 seconds on a cluster of just 16 NVIDIA RTX 5090 GPUs (roughly $50k-60k in hardware). This is a 12.5x reduction in GPU count from the May 2025 benchmark, well under EF's $100k capex ceiling. SP1 Hypercube went live on mainnet in February 2026.

These benchmarks should be read carefully. They reflect results on target hardware under controlled conditions. Independent teams attempting to replicate real-time proving on available cloud infrastructure have reported gaps. Framework documentation remains sparse, critical proof aggregation steps require GPU memory (27GB+) that exceeds standard cloud offerings, and a meaningful fraction of mainnet blocks fail to prove due to size and transaction-type incompatibilities. GPU supply itself is also constrained by AI-driven demand, and ASIC alternatives that could bypass the bottleneck are still in production.

Security is another challenge. The EF's December 2025 security foundations update revealed that some zkEVM implementations had overestimated their cryptographic security levels. In other words, the mathematical assumptions underpinning certain proof systems turned out to be weaker than believed. The updated roadmap requires all participating zkEVM teams to meet standardized security thresholds by end of 2026. On the integration side, zkVM Standards v0 (February 2026) standardizes the instruction set so execution clients write integration code once for multiple prover backends. Two execution clients (Reth and Ethrex) are fully spec-compliant, with the teams behind the Geth and Nethermind clients in active development to meet the standards. Real-time progress can be tracked here.

Centralization Implications

The prover role also introduces a new centralization question. If generating proofs requires specialized GPU hardware costing $30k-100k, won't proving concentrate among a few well-resourced operators? This mirrors the builder centralization dynamic in which economies of scale favor larger players.

The system is being designed so a concentrated prover set does not compromise security the way a concentrated builder set can. Provers cannot manipulate blocks because they operate on the trustless witness input described previously. The witness contains all the state data needed to execute the block, but it comes bundled with cryptographic proofs (Merkle proofs today, Verkle proofs in the future) that tie every piece of data back to the parent block's state root (a hash the entire network has already agreed on through consensus). If a prover receives a tampered witness, the proofs won't match the state root and the prover rejects it. And if a prover attempts to generate a fraudulent proof of incorrect execution, the mathematics of zk-proofs prevent it. A valid proof can only be produced from correct computation. The result is that provers can neither manipulate their inputs nor fake their outputs.

Beyond this fundamental security property, the mitigation strategy has several additional layers. A multi-client approach means no single prover stack dominates. The cost bar is expected to come down as hardware improves and software optimizes. And the system aims to operate on a 1-of-N liveness model. Only one honest prover needs to be live for the network to function. The EF has emphasized that proving should remain "viable outside of data center infrastructure” meaning home-based or small-operator proving is an explicit design goal.

The Ethereum Foundation says proving should remain "viable outside of data center infrastructure,” meaning home-based or small-operator proving is an explicit design goal.

Post-Quantum L1: Ethereum Setting the Standard

Ethereum and Bitcoin rely on elliptic curve cryptography (ECDSA) to secure wallets and authorize transactions. A sufficiently powerful quantum computer could reverse this cryptography using Shor's algorithm by deriving a wallet's private key from its public key and stealing funds. Current estimates for "Q-Day" center around 2032, though timelines are highly uncertain.

Ethereum's exposure spans all three of its protocol layers. Each relies on different cryptographic schemes built on math that quantum computers could theoretically break. Organizations including the National Institute of Standards and Technology (NIST) have already finalized standards for post-quantum cryptography. However, protecting the network is not as simple as updating Ethereum's cryptography to incorporate these signatures, primarily due to their size.

At the consensus layer, Ethereum validators sign attestations using BLS signatures (named after the cryptographers Boneh, Lynn, and Shacham), which are compact (96 bytes). They have a useful mathematical property in that thousands can be natively combined into one small signature. The quantum-resistant hash-based replacement under development (leanSig) runs approximately 3,000 bytes per signature and lacks this property. With a million validators producing 32,000 signatures per slot, raw post-quantum signatures would balloon to roughly 30 megabytes per slot, about 30x the current footprint. That would choke the network and raise bandwidth requirements to the point where running a home validator becomes impractical. The EF's answer is leanVM, a minimal zero-knowledge virtual machine that adds an aggregation step. Before a block is published, an aggregator compresses thousands of individual post-quantum signatures into a single compact proof, achieving roughly 250x compression without raising hardware requirements for validators.

The execution and data layers take different approaches. Users will migrate to quantum-safe wallet authentication gradually through account abstraction, which lets individual wallets upgrade their signature logic without a protocol-wide change. At the data layer, replacing the quantum-vulnerable KZG (Kate, Zaverucha, and Goldberg) commitment scheme is still in active research. Real-time updates can be tracked here.

The EF's zk research is tightly coupled with post-quantum work. The hash-based cryptographic primitives being developed for quantum resistance are the same class of primitives that power the STARK proving systems central to the zkEVM path. Progress on one front is expected to simultaneously advance the other, making the post-quantum and gigagas north stars mutually reinforcing rather than competing for engineering bandwidth.

The ETH/BTC Angle: Quantum as a Relative Advantage

More than a defensive upgrade, post-quantum security is a potential differentiator for Ethereum. The EF is actively collaborating with Blockstream researchers Mikhail Kudinov and Jonas Nick (the same people optimizing hash-based signatures for Bitcoin) to ensure the aggregation technology works across ecosystems. If both chains adopt the same standard, Ethereum's work would effectively set the default for the industry, much as Satoshi's choice of the secp256k1 elliptic curve did a generation earlier.

While Bitcoin's primary challenge is governance (see Galaxy Research prior report on Bitcoin’s Quantum progress), Ethereum’s is engineering. Ethereum must upgrade three cryptographic layers versus Bitcoin's one. But the EF has the coordination infrastructure, funding, and track record of shipping complex protocol changes that give it the credibility that it can achieve its objective.

Bitcoin also faces the politically radioactive question of what to do with the estimated 6–7 million BTC whose public keys are exposed onchain. That exposed supply includes roughly 1 million BTC attributed to Satoshi. The community will eventually face a forced choice between freezing or burning those coins (breaking the property rights narrative that underpins Bitcoin's identity), implementing rate-limited recovery mechanisms like Hourglass (adding significant complexity), or leaving them as a bounty for whoever builds a quantum computer first (risking sell pressure). None of these options are appealing. Ethereum faces a much smaller version of the same dilemma.

In the current state of affairs, it is likely that Ethereum will develop, test, and implement quantum resistance before Bitcoin, leveraging the same upgrade infrastructure that delivered the Merge, Pectra, Fusaka, and subsequent forks. In a world where quantum concerns grow louder by the month, the mere existence of a credible, funded, and actively executing roadmap is itself a narrative advantage.

Private L1: From Opt-In to Default

Ethereum transactions, unless modified, are fully public. As Vitalik wrote in April 2025, the cryptocurrency community "undervalued privacy for what is ultimately a bad reason: before ZK-SNARKs, we had no way to offer privacy in a decentralized way, and so we downplayed it." The Tornado Cash sanctions in August 2022 were the catalyst for a reassessment. Over 70% of Ethereum blocks began filtering sanctioned transactions, demonstrating that censorship and surveillance on Ethereum were operational realities.

The response has been a full-stack push across three layers. At the application layer, Railgun, a zk-SNARK-based protocol for private DeFi, hit $4.5 billion in cumulative volume by early 2026, with Vitalik using it and the EF staking 50,000 of the system’s RAIL tokens. Privacy Pools, co-authored by Vitalik, adds selective disclosure so users can prove funds aren't illicit while maintaining privacy. At the wallet layer, Kohaku, an open-source software development kit (SDK) and reference wallet unveiled at Devcon in November 2025, lets any wallet integrate shielded balances, per-dApp account isolation, and local light clients that eliminate surveillance by remote procedure call (RPC) providers. At the institutional layer, the EF rebranded its privacy team as the "Privacy Stewards of Ethereum" and launched a 47-member Privacy Cluster. EF researchers say they expect private transfers to be effectively solved at the application layer by late 2026.

All of this is happening at the application and wallet layers, deliberately avoiding consensus changes. Vitalik's privacy roadmap is explicitly "very light on Ethereum consensus changes." Unlike the gigagas or fast L1 north stars, Private L1 has no confirmed protocol-level implementation on the strawmap. Whether and how to embed privacy into L1 itself remains an open design question with no fork assignment. The spectrum ranges from Privacy Pools-style selective disclosure to unconditional privacy in the cypherpunk tradition. The EF has signaled a preference for the middle ground, privacy by default with optional transparency, but whether regulators accept that framework is unresolved.

Artificial Intelligence: Demand Driver and Risk Factor

In September 2025, the Ethereum Foundation launched a new internal team called dAI, led by Davide Crapis, with an explicit mandate to position Ethereum as "the preferred settlement and coordination layer for AIs and the machine economy." The strategy sits parallel to, rather than inside, the Strawmap. AI is not a north star on the technical roadmap, but the EF sees it as one of the most consequential forces shaping what Ethereum's infrastructure will be used for over the coming decade.

The dAI team's thesis is that Ethereum's role in an AI-mediated world is not computational competition with OpenAI or Google, but coordination. As Crapis framed it at NEARCON 2026, "Ethereum functions as a public, governance-less verification layer for AI." What Ethereum provides is identity, reputation, payment rails, and cryptographic verification for the autonomous agents that will increasingly transact on behalf of humans and businesses. The near-term deliverables are ERC-8004, a standard for AI agent identity and trust verification, and ERC-8183, co-developed with Virtuals Protocol, which defines commerce primitives for agent-to-agent transactions.

AI is not just a strategic initiative for the EF. It may also be one of the most concrete sources of future demand. Autonomous AI agents are already transacting onchain. The x402 protocol, governed by a foundation including Coinbase, Cloudflare, and Stripe, has processed over 140 million agent-to-agent transactions on the Base L2. Most of this activity is payments, but agents that evolve into autonomous economic actors will need financial services that traditional finance cannot provide. An AI agent cannot open a bank account or sign a brokerage agreement, but it can access Ethereum's full DeFi stack without anyone’s permission or a human identity. Ethereum is hardly alone in pursuing this opportunity. Nearly every major blockchain ecosystem is investing significant resources in becoming the preferred chain for AI agents, as are offchain companies with much better distribution.

AI also presents a direct challenge to the low-risk DeFi thesis this demand story depends on. Vitalik's argument that Ethereum's flagship use case should be simple, durable financial products requires DeFi protocols to be secure enough to trust with real capital. AI-assisted vulnerability discovery lowers the cost of finding smart contract bugs, and attackers have access to the same tools as defenders. Vitalik has repeatedly called bugs in code "Ethereum's biggest technical risk." And the risk is not theoretical. On April 18, 2026, an attacker exploited KelpDAO's cross-chain bridge to drain 116,500 rsETH (roughly $293 million), the largest DeFi exploit of 2026. The attacker deposited the stolen tokens as collateral on Aave V3, borrowed clean wrapped ETH, and left the protocol with an estimated $196 million in bad debt. Aave's TVL plunged from $26.4 billion to roughly $18 billion in 48 hours and at least nine protocols froze markets. The exploit did not compromise Aave's smart contracts, but it demonstrated how a single failure in cross-chain infrastructure can cascade through the interconnected lending stack.

The same properties that make Ethereum's DeFi stack the most credible financial system for autonomous agents also make it a high-value target. Whether the demand story outpaces the security risk is one of the most consequential open questions for Ethereum.

Assessing the Strawmap

The preceding sections walked through the Strawmap's technical implementations in detail, from near-term forks like Glamsterdam and Hegotá through longer-horizon bets on zero-knowledge proving, post-quantum cryptography, and protocol-level privacy. The scope is substantial. Seven forks through the end of the decade targeting an exponential increase in L1 throughput, sub-second finality, and cryptographic hardening across all three protocol layers.

Taken together, the Strawmap is Ethereum's attempt to scale the L1 to the point where it can truly serve as a credibly neutral global settlement layer, while simultaneously hardening the network's security and privacy foundations. Not all of these goals face the same odds of success, and not all carry the same weight for Ethereum's competitive positioning. The diagram below summarizes the primary tailwinds supporting the roadmap's execution and impact alongside the headwinds that complicate it.

The sequencing is deliberate and the dependencies are real. This means any delays will compound. A failure to deliver on Glamsterdam or Hegotá would cascade, shifting entire timelines. There will be intense focus on the ability of core developers to deliver on the upcoming two forks as a measure of the timelines’ future accuracy. Moreover, uncertainty remains over what will and will not be included in future upgrades, with the community still divided on sequencing priorities and factions competing to have their upgrades included.

The Strawmap's own authors are now signaling that its later years should be read more loosely. At the Soldøgn interop, ACDE co-leads ran a session on the Ethereum core development process that explicitly debated the Strawmap's structure. Developers flagged the per-fork year assignments past 2026 as overcanonicalized and likely to be softened in future revisions. The "headliner" construct that organizes each fork around a flagship feature will be retained but loosened to accept "theme plus candidate EIP" as a viable pattern, rather than committing to specific EIPs years in advance. The implication for investors is that the post-Hegotá portions of the roadmap should be treated as directional rather than scheduled. The destinations (gigagas, post-quantum security, private L1) are unlikely to change, but the sequencing and timing almost certainly will.

Credible Neutrality and the EF Mandate

Ethereum's premium is credible neutrality, and the cost of that premium is speed. The Strawmap's technical difficulty exists because Ethereum is attempting to achieve performance parity with faster competitors without compromising the decentralization guarantees that constitute its core value proposition.

The EF Mandate, published in early 2026, makes this constraint explicit. The document commits the Foundation to a narrowly scoped role as "technological protocol stewards," explicitly stating that the EF is "NOT a Casino" and will not optimize for token price, spend on policy, or engage in activities outside direct protocol development. The Foundation is deliberately reducing its surface area, spinning out functions like institutional outreach to Etherealize, while narrowing its own mandate to research, coordination, and ecosystem support.

This posture is unusual in crypto. Most competing chains centralize governance around corporate foundations or venture-backed development teams that can move quickly but whose priorities may shift with funding cycles, leadership changes, or regulatory pressure. Ethereum's path dependency, from the 2014 whitepaper through the Merge to the current Strawmap, has produced something competitors cannot easily replicate: a credibly neutral settlement layer governed by a foundation that has committed in writing to not capturing the protocol for its own benefit, and to eventually dissolving.

Ethereum's premium comes from people believing it won't be captured. The Strawmap is the technical plan for closing the speed gap. The EF Mandate is the governance commitment to preserving the neutrality premium while doing so. Whether performance parity and credible decentralization can be achieved simultaneously is the central question execution of the Strawmap will answer.

ETH the Asset

The Strawmap does not directly address ETH value accrual. There is no box on the map labeled "fix the fee revenue problem." But investors need an answer to the question of how protocol improvements translate into token value, and the Strawmap leaves that inference to the reader.

The strongest version of the bull case is being articulated not by the EF but by institutional-facing entities like Etherealize, which frames ETH as the reserve asset of a tokenized economy. The argument is that if the world's assets move onchain and most tokenized instruments anchor to offchain issuers (Circle, Tether, BlackRock, Apollo), ETH serves as a credibly neutral bearer asset that can transact among all of them without introducing a new counterparty. It is a primary collateral, uncensorable, and native to the settlement layer where the assets are issued. Under this framing, ETH is less a tech stock and more analogous to a reserve currency for on-chain finance, valued on the scale of the economy it denominates rather than on the cash flows it generates.

The bear case is straightforward. None of this is working as advertised today. ETH has been net inflationary since April 2024. The staking yield (~3%) underperforms U.S. Treasuries on a risk-adjusted basis. Fee revenue has collapsed. Ecosystem growth has not mechanically translated into token value accrual, as the past two years have demonstrated. The Strawmap's implicit answer, that gigagas L1 and teragas L2 throughput combined with EIP-1559 mechanics will create a radically different fee revenue picture, is years away and dependent on demand that has not materialized.

The honest framing is that the Strawmap addresses the supply side of the ETH investment thesis across every lens. What remains unanswered is whether the demand will materialize to fill that capacity.

Conclusion

The EF Mandate opens with a line that could have been written in 2014, "Ethereum started as a question: what if digital life could be shared, yet still belong to its users?" As EF president Aya Miyaguchi noted upon the document’s release, "The principles we listed on the Mandate are not new. However, there were times we were too implicit about them."

That acknowledgment captures something important about where Ethereum stands. The vision has not materially changed since Buterin's 2013 whitepaper described a "next-generation smart contract and decentralized application platform." The technical approach has evolved dramatically, from sharding to rollups to the current zk pivot, but the destination has remained constant: A credibly neutral, permissionless, censorship-resistant world computer. What the Mandate and the Strawmap together represent is not a new vision but a new level of clarity about the old one, the principles made explicit, the technical path made concrete, and the timeline made specific for the first time.

Bitcoin proved that there is global demand for a credibly neutral digital money, an asset and network that no single entity controls. That proof of concept was catalyzed by the 2008 financial crisis, which provided an unmistakable demonstration of why such a thing needed to exist. Ethereum aims to fill an analogous role one layer up as the credibly neutral smart contract platform and the settlement layer for programmable finance.

There is a fundamental difference in what Bitcoin and Ethereum each demand from their communities. Bitcoin requires holders. Ethereum requires builders. Its value depends on continuous research, active protocol development, and coordinated upgrades that carry real consequences if executed poorly. A global catalyst on par with the 2008 financial crisis has not yet materialized for Ethereum. DeFi, NFTs, and DAOs have demonstrated the concept in miniature, but the convergence of AI, quantum computing, and geopolitical fragmentation is making the case for neutral infrastructure harder to dismiss.

The coming years represent what may be Ethereum's most consequential window. The L2 scaling detour, whatever its merits, created a value accrual problem that the market has punished severely. The Strawmap is an explicit course correction, a return to L1 scaling as the priority, with a technical roadmap that, if executed, would address the performance gap that drove users to competing chains in the first place. There are indicators that suggest the prerequisites for trust are already in place. The TVL, the stablecoins, the institutional familiarity, the developer ecosystem. What remains is to rebuild the infrastructure for usage.

Ethereum has done this before. It migrated consensus mechanisms while live, a feat equivalent to replacing an airplane’s engine mid-flight. It shipped two hard forks in a single year. It built an L2 ecosystem from scratch. Whether it can execute the Strawmap on the timeline and at the ambition level the plan describes will be a key factor in shaping whether the current period is remembered as a temporary trough or the beginning of a longer decline.