Ethereum and Solana are deploying their largest protocol upgrades simultaneously. Ethereum's Glamsterdam hard fork, targeting H1 2026, introduces enshrined proposer-builder separation (ePBS) via EIP-7732 and block-level access lists (BALs) via EIP-7928, aiming for 10,000 TPS and a 78.6% gas fee r...
"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 and Solana are deploying their largest protocol upgrades simultaneously. Ethereum's Glamsterdam hard fork, targeting H1 2026, introduces enshrined proposer-builder separation (ePBS) via EIP-7732 and block-level access lists (BALs) via EIP-7928, aiming for 10,000 TPS and a 78.6% gas fee reduction. Solana's Firedancer client, developed by Jump Crypto over three years, is live on mainnet with 20% of validator stake after 100 days in production, while the companion Alpenglow consensus upgrade targets a reduction in transaction finality from 12.8 seconds to 150 milliseconds.
The two chains are attacking different bottlenecks. Ethereum is restructuring its block production pipeline to reduce MEV extraction — estimated at over $3 billion annually — and enable parallel transaction execution. Solana is replacing its validator software entirely and overhauling its consensus mechanism to close the gap between theoretical throughput (1 million TPS in controlled tests) and real-world performance. Both upgrades carry execution risk: Ethereum's Glamsterdam testnet is still transitioning from Devnet-4 to Devnet-5, and Firedancer shares zero codebase with Solana's incumbent Agave client, making consensus divergence a non-trivial concern.
The economic stakes are substantial. Ethereum controls approximately 68% of DeFi TVL at $45 billion. Solana holds $6.7–9.2 billion but generates higher daily chain fees ($1.03 million vs. approximately $182,000 for Ethereum L2s) and has overtaken Ethereum in DEX volume, processing $117 billion. The upgrade race will determine whether Ethereum's L1 can recapture activity lost to L2s and competing chains, or whether Solana's hardware-optimized approach continues eroding Ethereum's application-layer dominance.
Ethereum's Glamsterdam fork — named by combining Gloas and Amsterdam — is the network's first upgrade under its new biannual hard fork cadence, following the successful delivery of Pectra and Fusaka in 2025. The Ethereum Foundation's DevOps team has tested three proposed EIPs on Devnet-4 and is transitioning to Devnet-5, with public testnets planned for spring 2026 and a June 2026 aspirational mainnet target.
The upgrade clusters around three priorities:
1. Enshrined Proposer-Builder Separation (EIP-7732)
Currently, 80–90% of Ethereum block production depends on off-chain builders operating through relay infrastructure. ePBS moves this coordination into the protocol itself, replacing trusted intermediaries with an in-protocol commit-reveal flow. The design introduces explicit deadlines, payload commitments, and fallback behavior for builder non-delivery. Researchers estimate ePBS could reduce MEV extraction by up to 70%. However, Buterin himself acknowledged that ePBS "ensures that block builder centralization does not creep into staking centralization, but it leaves the question: what do we do about block builder centralization?"
2. Block-Level Access Lists (EIP-7928)
BALs require block producers to declare which accounts and storage slots each transaction will access. This declaration enables Ethereum clients to preload data and execute independent transactions concurrently across multiple CPU cores. The feature is a prerequisite for parallel transaction execution — a capability Solana has had since launch through its Sealevel runtime.
3. Gas Repricing (EIP-7904)
The upgrade realigns gas costs with actual computational resources consumed on modern hardware. Many gas prices were set years ago and no longer reflect execution costs. The recalibration produces a 78.6% reduction for both simple ETH transfers and complex smart contract interactions. The gas limit itself increases from 60 million to 200 million per block, targeting an effective throughput of approximately 10,000 TPS, up from roughly 1,000 TPS on the base layer today.
A further 25+ non-headliner EIPs are under consideration. The Ethereum Foundation explicitly declined inclusion of FOCIL inclusion lists, reduced slot times, multidimensional gas metering, and post-quantum signature verification for this fork.
Solana's upgrade strategy differs fundamentally from Ethereum's. Rather than modifying protocol rules through a hard fork, Solana is replacing its validator software and consensus mechanism in parallel.
Firedancer: The Client Replacement
Firedancer, written in C by Jump Crypto's engineering team led by Chief Scientist Kevin Bowers, shares zero codebase with Solana's incumbent Agave client. Bowers stated: "We didn't just tweak code, we rebuilt the whole engine." The client demonstrated 1 million TPS on commodity hardware in controlled tests. After going live on mainnet, Firedancer has run for 100 days, produced over 50,000 blocks, and now runs on validators representing 20% of network stake.
The client diversity argument is central to Firedancer's value proposition. Solana previously depended on a single client implementation, making correlated bugs a systemic risk. Validators running Firedancer report no performance degradation compared to Agave, and the two clients maintain consensus without divergence. However, Jito-Solana — a modified Agave client optimized for MEV extraction — still holds 72% of network stake, presenting a concentration concern that Firedancer's adoption has only begun to address.
Solana's active validator count has declined to 789, down from 2,500 in 2023. The consolidation reflects rising operational costs (annual voting costs of 394 SOL per validator) and a market that increasingly favors fewer, better-capitalized operators.
Alpenglow: The Consensus Overhaul
Alpenglow replaces Solana's existing consensus with a protocol targeting 150-millisecond finality, down from 12.8 seconds — a 98.3% reduction. Testnet deployment is planned for Q1 2026, with mainnet activation targeted for Q2 2026. At 150ms finality, Solana would match or exceed the settlement speed of traditional payment networks.
| Metric | Ethereum (Post-Glamsterdam) | Solana (Post-Firedancer + Alpenglow) | |---|---|---| | Target TPS | ~10,000 | 10,000+ real-world; 1M theoretical | | Finality | 12–15 min (with L1 inclusion) | ~150 ms | | Gas Fee Change | 78.6% reduction | Already sub-cent | | Active Validators | ~1.06 million | 789 | | Total Staked Value | 35.86M ETH (~$68.8B at $1,920) | ~388M SOL (~$55.7B at $143.80) | | Staking Rate | 28.9% of supply | ~65% of supply | | Client Diversity | 5+ execution clients | 2 (Agave, Firedancer) | | Block Production | Proposer-builder separated (ePBS) | Validator-produced, Jito-optimized | | MEV Handling | Protocol-enshrined auction (ePBS) | Jito bundles (72% stake) | | Upgrade Status | Devnet-4 → Devnet-5 | Firedancer live (20% stake); Alpenglow testnet |
Ethereum's approach prioritizes protocol-level guarantees: ePBS enshrines MEV handling rules, and 1.06 million validators provide redundancy. Solana's approach prioritizes raw performance through hardware optimization and consensus redesign, accepting a smaller validator set.
The upgrade race is ultimately a contest over economic capture. Both networks are attempting to attract the transaction volume that generates fee revenue and justifies token valuations.
Fee Revenue Divergence
Solana currently generates $1.03 million in daily chain fees. Ethereum L2s collectively generate approximately $182,000 per day, while Ethereum L1 fees have collapsed as activity migrated to rollups. Ethereum's revenue problem — documented extensively in prior analysis of L1-to-L2 value extraction — is partly structural. Glamsterdam's gas repricing and throughput increase aim to make L1 execution competitive again, but the 78.6% fee reduction mechanically lowers per-transaction revenue even as it increases transaction count.
DeFi TVL and DEX Volume
Ethereum's $45 billion DeFi TVL dwarfs Solana's $6.7–9.2 billion. But TVL measures idle capital; DEX volume measures economic velocity. Solana processed $117 billion in DEX volume, overtaking Ethereum. If Glamsterdam's throughput gains enable Ethereum L1 to recapture DEX activity from L2s and competing chains, the TVL advantage becomes economically active rather than merely custodial.
Stablecoin Settlement
Solana holds $14 billion in stablecoin issuance, increasingly positioning itself as a retail settlement layer. With Alpenglow's 150ms finality, Solana could offer stablecoin settlement faster than Visa's authorization network. Ethereum's stablecoin dominance remains larger in absolute terms, but the Tron network — not Ethereum — handles the highest stablecoin transfer volume at $86 billion.
Institutional Positioning
Solana's U.S. spot ETF products reached $1 billion in assets by early January 2026. Ethereum ETFs drew $2.5 billion in net inflows (though price has not responded proportionally, as analyzed in prior reporting). Institutional capital allocation between the two networks increasingly depends on which upgrade cycle delivers measurable throughput gains first.
Ethereum Glamsterdam Risks
The June target remains aspirational. Devnet-5 testing has not begun. The transition from 60 million to 200 million gas limit per block increases state growth, potentially raising hardware requirements for node operators. The 25+ additional EIPs under consideration could delay the fork if scope creep occurs. ETH has declined 45% from its October 2025 high of $3,600 to $1,920, and 37 million ETH is staked with 3 million more in the validator queue — a further price decline could trigger validator exits.
Solana Firedancer + Alpenglow Risks
Firedancer's zero-shared-codebase design is both its strength and its risk. A consensus divergence between Firedancer and Agave would halt the network. Solana's validator count decline to 789 raises centralization concerns. Alpenglow's 150ms finality target has not been validated on mainnet. Jito-Solana's 72% stake share means MEV extraction is concentrated in a single client variant, and Firedancer does not natively integrate Jito's bundle processing.
The simultaneous deployment of Glamsterdam and Firedancer/Alpenglow represents the most significant parallel upgrade cycle in smart contract platform history. Ethereum is retrofitting parallel execution and MEV reform onto a protocol designed for sequential processing. Solana is replacing its infrastructure from the ground up to close the gap between theoretical and practical performance.
Neither approach is without trade-offs. Ethereum's 1.06 million validators provide unmatched redundancy but slow consensus changes. Solana's 789-validator set enables rapid iteration but compresses the margin for error. The data shows both networks are making credible engineering bets. The market will price the outcomes based on measurable throughput, fee economics, and application-layer activity — not upgrade announcements. Delivery timelines in Q2–Q3 2026 will determine which chain's economic thesis holds.