Four of the five largest smart-contract platforms by total value locked have hard forks or consensus overhauls scheduled for the second half of 2026. Ethereum's Glamsterdam targets a 200-million gas-limit floor — roughly triple current capacity — with mainnet activation expected between September...
"That, to me, is this exciting step in the evolution of the protocol." — Anatoly Yakovenko, Co-Founder, Solana Labs, at Consensus Miami 2026
Four of the five largest smart-contract platforms by total value locked have hard forks or consensus overhauls scheduled for the second half of 2026. Ethereum's Glamsterdam targets a 200-million gas-limit floor — roughly triple current capacity — with mainnet activation expected between September and December. Solana's Alpenglow replaces its original TowerBFT consensus with the Votor/Rotor architecture, cutting finality from 12.8 seconds to approximately 150 milliseconds and freeing 75% of block space. Base's Beryl hard fork, activated June 25, introduced the B20 native token standard and shortened single-proof withdrawal finality from seven days to five. Avalanche continues building on its December 2024 Etna hard fork, which slashed the cost of launching a sovereign L1 by more than 99%.
Bitcoin is the notable outlier. No soft fork has reached the deployment stage since Taproot in November 2021. The most contentious proposal, BIP-110, faces an August 2026 deadline with miner signaling at 0.00% of hashrate as of mid-July.
The common thread across the four upgrading networks: the 2026 cycle focuses less on adding features and more on stability, parallel execution, institutional-grade infrastructure, and measurable reductions in operating costs.
Ethereum's Glamsterdam hard fork entered its final devnet phase in June 2026, locking the EIP bundle that core developers intend to carry through public testnets to mainnet. Holly Atkinson, chief product and technology officer at 1inch, told Cointelegraph that many in the ecosystem view Glamsterdam as Ethereum's most significant upgrade since The Merge in September 2022.
Two proposals anchor the upgrade:
EIP-7732 — Enshrined Proposer-Builder Separation (ePBS). This pulls block-building duties into the consensus layer, formally separating the validator that proposes a block from the builder that constructs the execution payload. MEV-Boost becomes optional rather than required once the protocol natively settles builder payments. The change makes the block-production pipeline more transparent and harder to game.
EIP-7928 — Block-Level Access Lists (BALs). BALs enable validators to process unrelated transactions in parallel by declaring which state elements each transaction touches. According to The Defiant, this expands the data propagation window from approximately 2 seconds to roughly 9 seconds, unblocking Ethereum's ability to safely handle higher throughput and accommodate more data blobs for L2 networks.
Together, these two proposals clear the path for a 200-million gas-limit floor — roughly tripling current L1 capacity from the ~60 million range. Proponents say this unlocks up to 10,000 TPS-equivalent throughput under realistic workloads. A separate gas-repricing package (EIP-2780) is projected to reduce standard ETH transfer costs by up to 71%, with broader L1 fee reductions of approximately 78.6%.
The mainnet activation date has not been set. Developers must first validate the upgrade on Holesky and Hoodi testnets. Based on the two-to-four-month public-testnet seasoning that recent forks required, a window of September to December 2026 is the working base case.
Solana co-founder Anatoly Yakovenko confirmed at Consensus Miami 2026 in May that the Alpenglow consensus upgrade could ship "as early as next quarter," targeting Q3 2026. A community validator test cluster launched May 11, 2026, and as of mid-July has been running for two months with dozens of external, production-grade validators distributed globally, according to CoinDesk.
Alpenglow is a ground-up replacement of Solana's original consensus primitives:
Under optimal network conditions, the architecture targets finality as low as 100 milliseconds, approaching what Yakovenko described as the physical limits of information transfer — "near the speed of light around the globe."
Moving validator votes off-chain frees an estimated 75% of current block space, creating room for application-layer throughput without raising hardware requirements.
The next milestone is Agave v4.2, scheduled for release on August 17, 2026, introducing foundational features including eXpress Data Path (XDK) networking and BLS key support. If testnet milestones proceed on schedule, the mainnet migration — referred to internally as the "Alpenswitch" — could occur in late September or October 2026.
Coinbase's Base network activated the Beryl hard fork on June 25, 2026, at 18:00 UTC. Unlike the still-pending Ethereum and Solana upgrades, Beryl is already live and in production. It bundled three changes:
B20 Native Token Standard. B20 tokens are ERC-20 compatible but implemented as Rust precompiles running inside the node software rather than as EVM bytecode. The standard is designed for stablecoin, real-world asset (RWA), and long-tail token issuers and includes a compliance toolkit with transfer strategies, freeze/seizure capability, role-based access control, memos, and supply caps. Executing token logic as precompiled Rust rather than interpreted EVM bytecode reduces gas costs and increases execution speed.
Shortened Withdrawal Finality. The single-proof dispute-game finalization window was reduced from 7 days to 5 days. The dual-proof fast path (TEE + ZK), introduced in the earlier Azul upgrade, remains at 1 day.
Reth V2 Integration. The upgrade to Reth V2 reduces node disk usage by 50% and increases throughput by 33%, according to Base documentation.
Beryl positions Base as a compliance-ready L2 environment. The B20 token standard — with its built-in freeze, seizure, and access-control features — is designed to satisfy regulatory requirements that generic ERC-20 tokens cannot meet without external wrapper contracts.
Avalanche's most significant structural change, the Etna hard fork (part of the broader Avalanche9000 initiative), activated on December 16, 2024. Its centerpiece, ACP-77, restructured how sovereign chains relate to the Primary Network: subnets were renamed to Avalanche L1s, validators were decoupled from the Primary Network validation requirement, and the base cost of launching a new L1 dropped to 1.33 AVAX per validator per month — down from the previous 2,000 AVAX staking requirement.
The cost reduction exceeded 99% and has materially lowered the barrier for institutions and projects to deploy custom blockchains on the Avalanche ecosystem.
In H2 2026, Avalanche's roadmap is less about a single branded fork and more about a series of performance upgrades targeting the C-Chain. The objective: making Avalanche's C-Chain one of the fastest EVM environments available. The network's growth strategy is oriented toward institutional participants, tokenized asset issuers, and regulated financial products.
Unlike Ethereum and Solana, which are shipping discrete hard forks with defined EIP bundles, Avalanche is executing a continuous-improvement model — incremental upgrades that compound without requiring a single coordinated network halt.
Bitcoin has had no consensus-layer upgrade since the Taproot soft fork activated in November 2021. Proposals including OP_CAT (BIP 347), OP_VAULT, CheckTemplateVerify, and LNHANCE have not reached agreement on an activation path.
The most contentious proposal in 2026 is BIP-110, which would temporarily restrict non-financial data on the Bitcoin blockchain for one year. According to CoinDesk, the proposal faces an early August deadline at block 961,542 with miner signaling at 0.00% of total hashrate as of mid-July. Only approximately 5 EH/s out of a network total near 940 EH/s has signaled support at any point — roughly 0.5%.
BIP-110 uses a user-activated soft fork mechanism with a 55% miner-signaling threshold, well below the traditional 95%. Major figures including Michael Saylor and Adam Back have publicly opposed the proposal. The measure appears likely to create only a small minority chain rather than a network-wide change.
The contrast is stark: Ethereum, Solana, Base, and Avalanche are each shipping structural upgrades that alter execution environments, consensus mechanisms, or token standards. Bitcoin's governance model, by design, makes such changes extraordinarily difficult to coordinate.
| Network | Upgrade | Status | Target Date | Key Metric | |---------|---------|--------|-------------|------------| | Ethereum | Glamsterdam | Final devnet | Sep–Dec 2026 | Gas limit: 60M → 200M (~3.3x capacity) | | Solana | Alpenglow | Testnet (2 months) | Sep–Oct 2026 | Finality: 12.8s → 150ms (~85x faster) | | Base | Beryl | Live (Jun 25) | Completed | Withdrawals: 7 days → 5 days; disk -50% | | Avalanche | Post-Etna | Continuous | Ongoing | L1 launch cost: 2,000 AVAX → 1.33 AVAX/mo | | Bitcoin | None scheduled | Stalled | N/A | Last fork: Taproot (Nov 2021) |
Throughput is the new battleground. Ethereum's 200M gas limit and Solana's 75% block-space liberation both address the same constraint: applications need more room. The approaches differ — Ethereum parallelizes execution via BALs; Solana moves consensus voting off-chain — but the economic effect is similar: more capacity at lower marginal cost per transaction.
Compliance infrastructure is moving on-chain. Base's B20 token standard embeds freeze, seizure, and role-based access controls directly into the precompile layer. This is a departure from the permissionless token model and signals that regulated asset issuers are a primary target for L2 growth.
Speed-of-finality competition is intensifying. Solana's 150ms target, if achieved at mainnet, would set a new benchmark for L1 finality among major networks. For comparison, Ethereum's slot time remains at 12 seconds, and even with ePBS, block confirmation patterns are unlikely to change in Glamsterdam.
Bitcoin's governance model is functioning as designed — and paying the cost. The inability to pass BIP-110 with 0.00% miner signaling is not a failure; it is the system's conservative design working as intended. The trade-off is that Bitcoin cannot ship the kind of structural upgrades its competitors are deploying in 2026.
The "upgrade cycle" is diverging. Ethereum and Solana ship discrete hard forks. Avalanche uses continuous delivery. Base operates under Coinbase's coordinated governance. Bitcoin requires near-unanimous consensus. These different governance models produce materially different rates of protocol evolution.
H2 2026 marks the densest concentration of major protocol upgrades since Ethereum's Merge in September 2022. The four networks pursuing upgrades — Ethereum, Solana, Base, and Avalanche — represent more than $700 billion in combined market capitalization and the majority of DeFi total value locked.
The upgrades share a common theme: the infrastructure layer is maturing. Gas-limit expansions, sub-second finality, compliance-native token standards, and dramatically cheaper chain deployment all point toward the same market demand — institutional-grade infrastructure capable of supporting regulated financial products at scale.
The economic implications are significant. Tripling Ethereum's gas capacity dilutes per-transaction fee revenue unless transaction volume grows proportionally. Solana's off-chain voting eliminates what was effectively a tax on block space. Base's precompiled tokens reduce gas costs for issuers. Each upgrade redistributes value within its respective ecosystem.
Whether these upgrades deliver on their stated objectives will be measurable by Q1 2027, when mainnet data for Glamsterdam and Alpenglow will provide the first real-world performance benchmarks against their respective testnets.