Solana and Ethereum are shipping their largest consensus and execution upgrades simultaneously. Solana's Alpenglow, entering its mainnet activation window between August and October 2026, replaces the TowerBFT consensus mechanism with Votor/Rotor, targeting 100-150ms finality — down from 12.8 sec...
"So the Alpenglow release is basically due sometime this year, I think next quarter. That, to me, is this exciting step in the evolution of the protocol." — Anatoly Yakovenko, Co-Founder, Solana Labs (Consensus Miami 2026)
Solana and Ethereum are shipping their largest consensus and execution upgrades simultaneously. Solana's Alpenglow, entering its mainnet activation window between August and October 2026, replaces the TowerBFT consensus mechanism with Votor/Rotor, targeting 100-150ms finality — down from 12.8 seconds. Ethereum's Glamsterdam, with Sepolia testnet activation set for August 3 and mainnet deployment projected around September 16, introduces enshrined proposer-builder separation (ePBS) via EIP-7732 and block-level access lists (EIP-7928) to enable parallel execution and a path toward 200 million gas limits.
Both upgrades attack the same constraint — throughput — through fundamentally different architectures. Solana rebuilds its consensus layer from scratch to collapse finality to sub-second. Ethereum restructures its execution layer to decouple block building from block proposing and enable concurrent transaction validation. Neither upgrade is incremental; both represent multi-year engineering bets reaching production in the same quarter.
The economic implications are material. Solana's Agave 4.2 companion release, targeting the week of August 17, cuts on-chain rent by approximately 90% and increases maximum transaction size from 1,232 to 4,096 bytes. Ethereum's gas limit roadmap targets a 3.3x increase from 60 million to 200 million, aiming for approximately 10,000 TPS on the base layer. For protocols, validators, and capital allocators, the next 90 days will determine which architectural thesis — monolithic speed or modular parallelism — delivers measurable economic throughput gains first.
Alpenglow is the most significant consensus change in Solana's six-year history. Approved under governance proposal SIMD-0236 with 98.27% validator support, the upgrade replaces two core subsystems:
TowerBFT → Votor. Solana's existing consensus mechanism requires 32 rounds of validator voting to finalize a block, with each vote submitted as an on-chain transaction consuming block space. Votor collapses this to one or two rounds. When 80% or more of validator stake is active, finality is achieved in a single round — approximately 100ms. Votor moves the entire voting process off-chain, using direct messaging and BLS signature aggregation. The result: approximately 75% of current block space occupied by vote transactions becomes available for user activity.
Turbine → Rotor. The block propagation layer changes from a tree of relay nodes to a single relay layer with erasure coding and stake-adjusted bandwidth allocation. The design minimizes network latency during block distribution.
The upgrade went live on a community test cluster on May 11, 2026. Mainnet activation requires validators to register BLS keys and the Validator Admission Ticket (VAT) system to be active. Anza, the engineering firm maintaining Solana's primary client, has not committed to an exact date within the August-October window.
Current Solana metrics for context: the network processes approximately 1,899 TPS on average (non-vote TPS between 1,600 and 3,800), with daily non-vote transactions averaging 102.7 million as of June 2026. Block time sits at 404ms with time-to-finality at 12.8 seconds. Q1 2026 handled 10.1 billion transactions, the highest quarterly figure in Solana's history, generating $89.5 million in network fee revenue.
Alongside Alpenglow, Anza is shipping Agave v4.2 with mainnet feature activations beginning the week of August 17, 2026. Three changes define this release:
Rent reduction (~90%). On-chain storage costs drop by approximately 90%. This directly lowers the capital required to deploy programs and maintain accounts on Solana, reducing barriers for developers building complex on-chain applications.
Transaction size increase (1,232 → 4,096 bytes). A 3.3x expansion of the maximum transaction payload. The current 1,232-byte ceiling has constrained developers building multi-instruction transactions, forcing them to split operations across multiple calls. The new limit enables more complex atomic operations within a single transaction.
Slot time reduction (400ms → 200ms). Implemented under SIMD-0525, the network's base time unit — the window in which a block leader processes transactions — is halved. Combined with Alpenglow's consensus changes, this creates a compounding effect on throughput.
A fourth component, the eXpress Data Path (XDK), enables kernel-bypass networking for validators. According to reporting from ETHNews, the hardware requirements for running a competitive validator increase as a result — raising questions about validator centralization.
Ethereum's Glamsterdam upgrade, which entered its final devnet phase in June 2026, takes a structurally different approach. Rather than replacing consensus, it restructures execution.
EIP-7732: Enshrined Proposer-Builder Separation (ePBS). Currently, block builders construct blocks and proposers select them through an off-protocol relay system (MEV-Boost). EIP-7732 formalizes this separation directly in the protocol, eliminating reliance on centralized relays. The data propagation window extends from 2 seconds to approximately 9 seconds, giving builders more time to construct optimal blocks. This targets the builder centralization problem — where a small number of sophisticated builders capture disproportionate MEV.
EIP-7928: Block-Level Access Lists (BALs). Every block will include a map of every account and storage slot it touches, along with post-execution state values. Nodes receive this map upfront, enabling parallel disk reads and parallel transaction validation. This is Ethereum's mechanism for achieving concurrent execution without abandoning its sequential transaction model.
Gas limit path: 60M → 200M. The combination of ePBS and BALs creates the technical foundation for raising Ethereum's gas limit from approximately 60 million to 200 million per block. At 200 million gas, Ethereum targets approximately 10,000 TPS on the base layer — compared to the current effective rate of roughly 25.78 TPS (Q1 2026 data). This increase remains subject to testing, governance, and security evaluation.
Timeline: Sepolia testnet activation was tentatively set for August 3, 2026, with mainnet deployment projected around September 16.
Ethereum's current L1 gas price sits at approximately 0.139 gwei as of August 7, 2026. A simple ETH transfer costs $0.10-$0.20. Layer 2 networks handle more than 90% of Ethereum's total transaction throughput.
| Metric | Solana (Pre-Alpenglow) | Solana (Post-Alpenglow Target) | Ethereum (Current) | Ethereum (Post-Glamsterdam Target) | |---|---|---|---|---| | Finality | 12.8 seconds | 100-150ms | ~13 minutes (2 epochs) | ~13 minutes (unchanged) | | Block time | 400ms | 200ms (Agave 4.2) | 12 seconds | 12 seconds (unchanged) | | L1 TPS (actual) | 1,600-3,800 | TBD | ~25.78 | Target: 10,000 | | Avg. tx cost | ~$0.00025 | Lower (90% rent cut) | $0.10-$0.20 | Lower (higher gas limit) | | Max tx size | 1,232 bytes | 4,096 bytes | ~128KB (calldata) | ~128KB (unchanged) | | Consensus votes | On-chain (~75% of block space) | Off-chain (BLS aggregation) | Off-chain (attestations) | Off-chain (ePBS formalized) |
The comparison reveals different optimization targets. Solana is collapsing finality by 85x (12.8s → 150ms). Ethereum is expanding throughput by potentially 388x (25.78 → 10,000 TPS target), though this is a long-term roadmap target dependent on gas limit governance decisions.
Both networks face client diversity dynamics that affect upgrade risk.
Solana: Agave-derived clients (including Jito variants) account for approximately 86% of staked SOL. Firedancer, Jump Crypto's clean-room C/C++ implementation, runs approximately 2% of stake on its full client, with Frankendancer (the hybrid) at approximately 12%. Combined Jump-derived clients represent roughly 14-40% of stake depending on measurement methodology. The 33% stake-weight threshold — where client diversity becomes operationally meaningful for consensus safety — has not been definitively crossed by non-Agave clients. Firedancer published v1.0.0 to testnet on June 12, 2026.
Solana operates with 906 active validators and a Nakamoto coefficient of 19.
Ethereum: Maintains broader client diversity across execution layer (Geth, Nethermind, Besu, Erigon, Reth) and consensus layer (Prysm, Lighthouse, Teku, Nimbus, Lodestar) clients. No single client holds a supermajority of validator stake. This diversity model, while slower to coordinate upgrades, reduces the risk of a single-client bug halting the network.
The Alpenglow upgrade introduces additional coordination complexity: validators must register new BLS keys specifically for the Votor consensus mechanism. This creates a migration gate that could slow adoption.
Fee revenue redistribution. Solana's elimination of on-chain vote transactions frees approximately 75% of current block space. If demand fills this capacity, fee revenue from user transactions could increase substantially — though lower per-transaction costs from rent reduction and competition could offset total revenue growth. Q1 2026 fee revenue was $89.5 million; the post-upgrade trajectory depends on demand elasticity.
Validator economics. Solana's Agave 4.2 increases hardware requirements through XDK kernel-bypass networking. Validators must run more capable machines, raising the capital barrier to entry. Ethereum's Glamsterdam does not materially change validator hardware requirements in the near term, though a 200M gas limit would eventually increase state growth and storage demands.
Developer costs. Solana's 90% rent reduction and 3.3x transaction size increase directly reduce deployment costs. Ethereum's fee reduction depends on gas limit governance — the technical capability exists, but validators must vote to increase limits.
Stablecoin and settlement implications. Both upgrades target institutional settlement use cases. Solana's sub-second finality makes it competitive with traditional payment rails for point-of-sale and remittance flows. Ethereum's higher throughput could absorb more tokenized asset settlement volume on L1 without forcing activity to L2s.
Solana risks:
Ethereum risks:
Shared risk: Both upgrades are shipping within weeks of each other in Q3 2026. If either encounters a critical bug post-activation, the other network may absorb displaced activity — creating an asymmetric competitive dynamic.
The Q3 2026 upgrade cycle is the most consequential simultaneous infrastructure change in L1 history. Solana is rebuilding consensus to achieve sub-second settlement. Ethereum is restructuring execution to scale throughput by orders of magnitude. Neither network is making incremental improvements; both are shipping foundational architectural changes.
The comparative result will not be determined by technical specifications alone. Finality speed matters only if applications and capital flows demand it. Throughput capacity matters only if transaction volume fills it. The next 90 days will produce the first real-world performance data for both upgrades — data that will determine whether the theoretical gains survive contact with mainnet economics.