Most crypto investors still obsess over price charts. But in 2026, a growing share of attention is shifting back to improving the fundamentals of the protocols. Ethereum, Solana, and Avalanche are preparing some of their largest protocol upgrades in years, while Coinbase's Base network rolled out its Beryl hard fork in a bid to streamline the network with a native token standard and shorter withdrawal windows. Bitcoin development, however, remains frozen, with developers still arguing over controversial covenant proposals and post-quantum computing upgrades.
Protocol upgrades have historically focused on adding features, speed, and throughput. However, in 2026, the emphasis is shifting toward reliability, predictable governance, and institutional-grade infrastructure that can support large-scale financial use cases. Here are the top five major blockchain upgrades to watch in the second half of 2026.
Ethereum: Glamsterdam
Glamsterdam is arguably the most consequential upgrade this year, already being tested on devnets. According to Ethereum's public roadmap, Glamsterdam is designed to improve scalability, harden the layer-1, and make the network easier to use, with a mainnet launch expected sometime in the second half of 2026. The upgrade should improve processing speeds by allowing more transactions to be processed simultaneously, expand capacity so Ethereum can handle more data at higher throughput, and reduce database bloat. Those changes should make the chain better suited for stablecoin settlement and real-world asset use cases.
One of the key changes is enshrined proposer-builder separation (ePBS). Most validators still depend on a small set of specialized builders and relays, which concentrates control over transaction ordering. That setup amplifies maximal extractable value (MEV), censorship, and centralization risks. ePBS is designed to pull block building and proposing back into the protocol and make the process more transparent and accountable. However, some analysts note that ePBS does not eliminate MEV or fully solve builder centralization; practices like sandwich attacks may therefore migrate rather than disappear.
Solana: Alpenglow
Solana's biggest change this year is Alpenglow, a consensus upgrade that reworks the network's core protocol. Billed as the chain's most significant consensus upgrade yet, Alpenglow was overwhelmingly approved through a governance process in September 2025 and remains under development, expected to ship alongside the Agave 4.1 validator client release later in 2026. This upgrade will be a major tailwind that reinforces the 'internet capital markets' thesis even more strongly.
At its core, Alpenglow is designed to dramatically speed up how quickly the network reaches finality. Instead of relying on Solana's existing TowerBFT-based consensus mechanism, it introduces a redesigned system built around a new voting component called Votor. The practical impact is a major reduction in confirmation times, with finality targeted at roughly 100-150 milliseconds in optimal conditions, compared to around 12.8 seconds today. Beyond speed, the upgrade also removes onchain vote transactions, which currently account for a significant portion of network activity. By streamlining how validators communicate and agree on the state of the chain, Alpenglow is intended to make Solana both lighter and more efficient under load. Removing onchain vote transactions is a key improvement for institutional allocators because it cleans up validator economics and provides honest telemetry, which matters when underwriting SOL as a treasury asset.
Base: Beryl
Base's Beryl hard fork went live recently, following a short sequencer-related outage when block production stalled for around two hours due to an invalid block that triggered a temporary consensus failure. User funds were unaffected during the incident, but lessons learned from the episode will be used to further strengthen Base as a platform for global, 24/7 finance.
According to Base's documentation, Beryl introduces a set of changes aimed at tightening the network's performance and reducing friction at the edges. These include the B20 native token standard, a shortening of withdrawal finality from seven days to five, and integration with Reth V2, which is expected to reduce node storage requirements while improving execution efficiency. Base has been moving toward a more unified stack approach, giving it greater control over how the network is built and upgraded, and allowing changes to ship more quickly than under the earlier Optimism Superchain model. The trade-off is that liquidity, which once moved more freely across the broader Superchain ecosystem, may become more fragmented, even as Base deepens its integration with Coinbase's wider user base.
Avalanche: Octane
Avalanche's next chapter is less about a single branded hard fork than a broader push to improve performance while courting institutions and tokenized asset issuers. The recent Etna hard fork replaced the old subnet model with sovereign Avalanche L1s, cutting the cost of launching a dedicated blockchain by more than 99% and making the network more attractive to institutional players. This has already seen success, with Progmat—which accounts for roughly 63% of Japan's national security token market—migrating more than $2 billion in tokenized assets to a dedicated Avalanche L1. Additionally, the Avalanche Payments Collective, backed by firms including major asset managers, has been formed.
Avalanche is also pushing two upgrades aimed at making its C-Chain one of the fastest Ethereum Virtual Machine (EVM) environments. Streaming Asynchronous Execution is a way to separate transaction execution from consensus so the chain can run more continuously and size capacity closer to normal demand. For users, the practical effect should be higher throughput and lower, steadier fees during periods of heavy activity.
Bitcoin: OP_CAT and Quantum Debates
Bitcoin is the outlier here because its biggest developments in 2026 are not scheduled upgrades but a continuation of passionate debates over whether the protocol should become more programmable and how urgently it should be hardened against quantum threats. Bitcoin has not activated a major soft fork since Taproot in 2020, which upgraded Bitcoin's scripting to make transactions more flexible and improve privacy. Since then, discussion around covenant-related proposals such as OP_CAT, CheckTemplateVerify (CTV), and Lightning-focused ideas like LNHANCE has intensified. None of these changes has an agreed path to activation.
Researchers have also been debating BIP-360 and related proposals as ways to make it easier to migrate coins into quantum-resistant spending paths, if and when the quantum computing threat becomes real. Covenant proposals could unlock safer storage and richer scripting, but the subject remains divisive and subject to much debate. Those proposals could improve self-custody security, fee management, and protocols such as Lightning and Ark, while giving institutions more programmable custody logic directly on the L1. Bitcoin development is infamously slow, and any change to the protocol is pored over from every angle. There is general agreement that no covenant opcode is on track for activation this year, and reaching consensus on proposals like OP_CAT or CTV is still some distance away. On the post-quantum side, a full migration to quantum-resistant addresses and signatures would take years even under optimistic assumptions, making it unlikely that a quantum-resistance upgrade will be implemented before the end of 2026.
Source: Cointelegraph News