Ethereum and Solana are simultaneously executing the largest protocol overhauls in their respective histories. Ethereum's Glamsterdam hard fork entered its final devnet phase on June 16, 2026, locking in ten EIPs headlined by Enshrined Proposer-Builder Separation (EIP-7732) and targeting a 200 mi...
"Probably the largest fork we've had since the Merge. It will change a lot of assumptions about Ethereum and set us up for much more scaling in the future." — Parithosh Jayanthi, Core Developer, Ethereum Foundation
Ethereum and Solana are simultaneously executing the largest protocol overhauls in their respective histories. Ethereum's Glamsterdam hard fork entered its final devnet phase on June 16, 2026, locking in ten EIPs headlined by Enshrined Proposer-Builder Separation (EIP-7732) and targeting a 200 million gas-limit floor — more than triple current levels. Solana's Alpenglow consensus rewrite, live on a community test cluster since May 11, replaces both Proof of History and TowerBFT with two new components (Votor and Rotor), targeting transaction finality of 100–150 milliseconds versus the current 12.8 seconds.
The two upgrades represent fundamentally different scaling philosophies converging on similar goals: higher throughput, lower latency, and reduced value extraction by intermediaries. Glamsterdam prioritizes MEV reform and parallel execution within Ethereum's existing validator architecture. Alpenglow prioritizes raw consensus speed, compressing finality by roughly 100x. Combined, the two chains secure approximately $75 billion in DeFi TVL and process the majority of on-chain economic activity. The outcome of these concurrent upgrades will define the competitive positioning of the two dominant smart-contract platforms through 2027.
Ethereum's Glamsterdam hard fork consolidates ten Ethereum Improvement Proposals into a single activation event. Core developers confirmed on June 16 that all planned EIPs are running on devnets, with public testnet deployment and mainnet activation expected in H2 2026 — current best estimate is late August, according to CoinDesk reporting.
The two headline proposals:
EIP-7732 — Enshrined Proposer-Builder Separation (ePBS). This moves the separation between block builders and block proposers from off-chain relay infrastructure (currently dominated by a handful of entities) into the core protocol. The stated objective is to reduce centralization risk in block construction and narrow the attack surface for maximal extractable value (MEV). According to developer estimates cited by The Defiant, ePBS could reduce MEV extraction by up to 70%, resulting in fewer sandwich attacks and front-running on DeFi protocols.
EIP-7928 — Block-Level Access Lists (BALs). Blocks declare in advance which accounts and smart-contract storage slots they will access. This allows Ethereum clients to execute non-overlapping transactions in parallel — a structural shift from the network's current sequential execution model. The practical effect: L1 throughput increases without raising the gas limit, by using existing capacity more efficiently.
Gas Limit Expansion. EIP-8037, finalized in May 2026, sets a fixed cost per state byte and establishes a separate gas reservoir for state growth. This was the final prerequisite for client teams to model sustainability at a 200 million gas-limit floor — up from the current approximately 60 million. At the target gas limit, Ethereum L1 throughput could reach an estimated 10,000 transactions per second, according to developer roadmap documents.
Gas Repricing. Jayanthi stated that Glamsterdam "will majorly change the cost of actions on Ethereum. High-level compute gets cheaper and state gets more expensive." The repricing alters the economic calculus for every smart contract on the network — compute-heavy operations (mathematical calculations, signature verification) become cheaper, while operations that write to or expand on-chain state become more costly. This is a deliberate architectural choice: incentivize computation, discourage state bloat.
Current Ethereum L1 gas prices sit at approximately 0.196 gwei as of June 22, with average transaction fees at $0.18, according to Etherscan data. Glamsterdam's repricing could further compress fees on compute-intensive transactions while increasing costs for state-expanding operations.
Alpenglow replaces Solana's two foundational consensus mechanisms — Proof of History (PoH) and Tower BFT — with a clean-sheet design comprising two components:
Votor collapses the current 32-round confirmation process into one or two rounds. When 80% or more of validator stake is online (which Solana typically achieves), finality is reached in a single round — approximately 100 milliseconds. This compares to the current 12.8-second finality under TowerBFT.
Rotor replaces Turbine for block propagation. Turbine uses a multi-layer tree structure with a fanout of 200 nodes. Rotor uses a single-hop relay model where shreds (erasure-coded data packets) propagate in one step rather than cascading through tree layers. Rotor is natively compatible with multicast infrastructure such as DoubleZero.
Anza, the Solana Labs spinout leading Alpenglow development, confirmed on May 11 that the upgrade is live on a community test cluster. Solana co-founder Anatoly Yakovenko stated at Consensus Miami 2026 on May 7 that "the Alpenglow release is basically due sometime this year, I think next quarter." The current roadmap targets Agave client version 4.1 for Q3 2026, followed by security audits in Q4, with mainnet activation expected before year-end.
The finality improvement — from 12.8 seconds to 100–150 milliseconds — represents a roughly 100x compression. At that speed, Solana finality would approach the physical latency constraints of global internet routing, a point Yakovenko has referenced publicly.
Solana's upgrade cycle includes a concurrent infrastructure change: Jump Crypto's Firedancer validator client. Written entirely in C as a clean-room implementation (not derived from the existing Agave/Solana Labs codebase), Firedancer began producing mainnet blocks in May 2026 after over 100 days of continuous testnet operation and 50,000+ validated blocks.
As of May 2026, approximately 26% of Solana validators were running either Firedancer or its hybrid predecessor Frankendancer. No double-sign incidents or chain halts were reported during the transition.
Firedancer's significance is architectural redundancy: Solana becomes the first major L1 to run two independent validator codebases in production. If one client has a bug, the other continues producing blocks. Firedancer also pushes real-world throughput to 5,500+ TPS and aims for a theoretical ceiling of 1 million TPS at full rollout.
| Metric | Ethereum (Current) | Ethereum (Post-Glamsterdam) | Solana (Current) | Solana (Post-Alpenglow) | |---|---|---|---|---| | L1 TPS | 15–30 | Up to ~10,000 (est.) | 3,000–5,000 | 5,500+ (Firedancer) | | Finality | ~12 minutes | ~12 minutes* | 12.8 seconds | 100–150 ms | | Gas Limit / Block Capacity | ~60M gas | 200M gas (target) | N/A (continuous) | N/A (continuous) | | Avg. Transaction Fee | $0.18 | TBD (repricing) | <$0.01 | <$0.01 | | DeFi TVL | ~$70B | — | ~$4.9B | — | | Validator Clients | 5+ (Geth, Nethermind, etc.) | 5+ | 2 (Agave, Firedancer) | 2 | | MEV Mitigation | Off-chain (MEV-Boost) | On-chain (ePBS) | Scheduler-based | Scheduler-based |
Note: Glamsterdam does not change Ethereum's slot/epoch finality timing directly. The 12-minute finality refers to economic finality under Casper FFG. Execution throughput increases, but consensus finality remains on the existing schedule.
Fee Revenue Redistribution. Ethereum's gas repricing alters value flows across the ecosystem. Protocols with heavy compute requirements (AMMs, options pricing, oracle aggregation) may see reduced execution costs, potentially shifting fee revenue toward state-heavy applications (NFT minting, new contract deployments). The net effect on total Ethereum fee revenue — currently depressed at sub-$0.20 average transactions — is unclear. Higher throughput could increase total transaction volume, offsetting lower per-transaction costs.
MEV Market Restructuring. Ethereum's ePBS moves a multi-billion-dollar MEV market from off-chain relay infrastructure into the protocol layer. The current MEV-Boost relay system, dominated by a small number of operators, has been a persistent centralization concern. Enshrining proposer-builder separation reduces the trust assumptions required for fair block construction, but also eliminates a revenue stream for existing relay operators.
Solana's Speed Premium. Sub-200ms finality, if achieved on mainnet, has direct implications for institutional adoption. High-frequency trading firms, payment processors, and cross-chain bridge operators have historically cited Solana's finality time as a limiting factor versus traditional financial infrastructure. Compressing finality to near-internet-latency speeds could expand Solana's addressable market in time-sensitive applications, though the network's $4.9 billion TVL and depressed SOL price (trading around $72-81 in June 2026, down 50%+ from 2025 peaks) suggest the market has not yet priced in these upgrades.
Infrastructure Cost Basis. Both upgrades carry meaningful infrastructure costs. Ethereum's 200M gas limit increases storage and bandwidth requirements for node operators. Solana's Alpenglow and Firedancer require validators to run and maintain more complex client software. The economic burden falls on the validator set — a cost dynamic consistent with the broader trend of rising infrastructure overhead in proof-of-stake networks.
Glamsterdam Risks. The 200 million gas limit has not been tested under adversarial mainnet conditions. Client teams model state growth at 120 GiB per year under the new ceiling, but real-world usage patterns may deviate. Gas repricing could create unexpected economic distortions — protocols optimized for current gas schedules may require contract migrations. The upgrade's scope ("the largest fork since the Merge," per Jayanthi) inherently increases the surface area for implementation bugs.
Alpenglow Risks. Replacing both PoH and TowerBFT simultaneously removes two battle-tested mechanisms. Votor's single-round finality depends on 80%+ stake being online — an assumption Solana generally meets, but the network has experienced multiple outages in prior years. Rotor's single-hop propagation model has not been tested under the kind of adversarial network conditions (DDoS, geographic partition) that Turbine's tree structure was designed to handle. The transition from testnet to mainnet under real economic stakes is the primary risk vector.
Concurrent Upgrade Risk. Both networks are deploying transformational changes within the same calendar quarter. Any major incident on either chain during the upgrade window could create cross-chain contagion effects through shared DeFi infrastructure (bridges, oracles, multi-chain protocols).
The simultaneous deployment of Glamsterdam and Alpenglow represents the most consequential concurrent infrastructure upgrade cycle in smart-contract platform history. Ethereum is restructuring how blocks are built and priced. Solana is replacing how blocks reach consensus. Neither upgrade addresses the other chain's primary weakness — Ethereum remains slower at finality; Solana remains smaller in economic security and TVL. The competitive dynamic is not winner-take-all but rather a continuing divergence in architectural philosophy: Ethereum optimizes for economic security and composability at the base layer while offloading speed to L2s; Solana optimizes for raw L1 performance. Both chains are betting that their respective approaches will capture the next wave of on-chain economic activity. The data from H2 2026 mainnet performance will provide the first empirical test of these assumptions.