In the second half of 2026, protocol upgrades are emerging as a key variable for the blockchain market, rather than prices.

Cointelegraph reported on July 2 that Ethereum, Solana and Avalanche are preparing major upgrades, and that Coinbase's Base recently implemented the Beryl hard fork.

This year's upgrades are focusing less on adding features and more on stability, predictable governance and institutional-grade infrastructure. Haesky Group senior researcher Tim Sun stressed that the foundation to handle large financial demand has become important.

On Ethereum, Glamsterdam is in focus. The upgrade is being tested on developer networks and is expected to be applied to the mainnet in the second half of 2026.

The goals are improved scalability, stronger Layer 1 and better usability. It aims to increase processing speed and data throughput while reducing database bloat. ePBS focuses on easing control over transaction ordering concentrated among a small number of builders and relays.

Solana will change its consensus structure with Alpenglow. The upgrade was approved through governance procedures in September 2025 and is expected to be deployed in late 2026 with the Agave 4.1 validator client. It will introduce a new voting component, Votor, focusing on cutting finality time from about 12.8 seconds to 100 to 150 milliseconds under optimal conditions. It is also expected to reduce network load by eliminating on-chain voting transactions.

Base activated the Beryl hard fork last week. Just before that, there was an incident in which block production stopped for about 2 hours due to incorrect blocks. Jesse Pollak said user funds were not affected. Beryl includes the B20 native token standard, a reduction in the withdrawal finalisation period to 5 days from 7 days, and Reth V2 integration.

Avalanche is placing more weight on restructuring aimed at institutional demand and tokenised-asset issuers than on a single hard fork. A recent Etna hard fork replaced the existing subnet model with a sovereign Avalanche Layer 1 and lowered the cost of launching a dedicated blockchain by more than 99 percent. Progmat moved more than $2 billion in tokenised assets to a dedicated Avalanche L1.

Bitcoin is cited as an exception. It has had no major soft fork since Taproot in 2020, and proposals such as OP_CAT, CheckTemplateVerify and LNHANCE have not reached agreement on an activation path.

Discussions are also continuing on BIP-360 for a transition to quantum resistance, but Cointelegraph reported that the likelihood of actual implementation within 2026 is low.