2026 Public Chain Upgrades: Analysis of Technical Roadmaps and Performance Landscape of Solana Alpenglow and Ethereum Glamsterdam

The crypto industry in 2026 is standing on a subtle, underlying watershed. After the narrative-driven cycle of the last round, the market’s evaluation criteria for public chains are shifting—from “whose concept is better” to “whose engineering capability is stronger.” The core variable driving this shift is a major round of upgrades pushed forward by two leading public chains at nearly the same time.

In May 2026, Solana underwent its most comprehensive consensus mechanism reform since mainnet launch—code-named “Alpenglow.” The Alpenglow upgrade officially entered the community validation node testing stage. In almost the same time window, Ethereum is also pushing its flagship upgrade for 2026—“Glamsterdam,” moving its parallel execution framework from the roadmap into the development network validation phase.

As of May 19, 2026, Gate data shows Solana is priced at $84.98, down approximately 48.95% over the past year; Ethereum is priced at $2,130.24, down about 15.58% over the past year. Both are in low-price ranges, and the rollout of large-scale technical upgrades is reshaping the market’s fundamental assessment framework for these two chains.



## In Spring 2026, the two chains press the upgrade button at the same time

Solana’s Alpenglow upgrade was activated on May 11, 2026 into the community testnet, marking the official entry of this, Solana’s largest-scale consensus reform in history, into practical validation node operations. The core development team, Anza, announced that validators can now perform “Alpenswitch” operations in the testnet, migrating from the current consensus architecture based on PoH and TowerBFT to a new consensus framework.

On Ethereum’s side, Glamsterdam is currently in the multi-client development network phase. ePBS (proposer-builder separation, embedded) has completed end-to-end testing, covering almost all client implementations. The Ethereum Foundation confirmed in May 2026 that the core goals of Glamsterdam have essentially been achieved, with a gas limit set at 200 million. Mainnet activation is expected in 2026 Q3; earlier plans had targeted June.

Although the two chains pressed the upgrade button around the same period, their technical route choices are sharply different: Solana chooses to “cut into” the underlying consensus mechanism, while Ethereum chooses structural optimization in the execution layer and the block-building layer.

## Engineering evolution along two different paths

### Solana Alpenglow timeline

The technical concept of Alpenglow traces back to research conducted at the Distributed Systems Lab of ETH Zurich by Professor Wattenhofer. It was later executed as an engineering implementation by the Anza team, with collaboration from Jump Crypto’s Firedancer team to advance multi-client compatibility.

In May 2025, the Alpenglow proposal was first publicly unveiled at the Solana Accelerate conference. In the same year, in September, it passed with overwhelming support of 98.27% in validator governance votes, with 1.05% voting against and 0.36% abstaining; the amount of staked participation in the vote accounted for 52% of total staking. In early 2026, Alpenglow entered the main branch of the Agave client to start private cluster testing. On May 11, it officially launched the community testnet. On May 15, the testnet expanded from 49 validators to 86 validators. Solana co-founder Anatoly Yakovenko said at the Consensus Miami conference that Alpenglow could be deployed to mainnet as early as 2026 Q3.

In addition, the Firedancer independent validator client developed by Jump Crypto officially went live on the Solana mainnet on May 16, 2026, beginning block production. It has already processed tens of millions of transactions and controls approximately 7% of the network’s staking weight. This milestone means Solana no longer relies solely on a single validator client, Agave. Increasing client diversity has important structural significance for improving network resilience.

### Ethereum Glamsterdam timeline

Glamsterdam is another key upgrade for Ethereum after it successfully delivered two hard forks—Pectra and Fusaka—in 2025. In February 2026, the Ethereum Foundation released the “Protocol Priorities Update for 2026,” listing Glamsterdam and Hegotá as the two major core upgrades of the year. It planned protocol evolution along three main lines: scaling, improvements to user experience, and L1 consolidation.

Glamsterdam’s originally scheduled mainnet time was June 2026, but it was pushed back to the third quarter. The upgrade includes two already confirmed Ethereum Improvement Proposals (EIPs): EIP-7732 (ePBS) and EIP-7928 (block-level access lists).

## Two reform paths for consensus architecture

Below, we break down the structural differences between the two upgrades across four dimensions: technical architecture, performance metrics, validator mechanisms, and MEV governance.

### Consensus reform: Reconstruction vs. optimization

Solana’s Alpenglow is a complete replacement of the consensus layer. It removes the two core components that have been running since Solana’s genesis—PoH and TowerBFT—and replaces them with two entirely new protocols: Votor, which is responsible for consensus voting and finality, and Rotor, which is responsible for block propagation.

In Solana’s original design, PoH plays the role of a “cryptographic clock,” providing timestamps for transactions through continuous hashing, enabling validators to reach agreement on event order without real-time communication. TowerBFT then uses timing proofs generated by PoH, locking in validator stances through 32 rounds of incrementally increasing votes. This design performed well during Solana’s early operation, but structural flaws have gradually become apparent: validator voting is handled as on-chain transactions, consuming roughly 75% of block space.

Alpenglow fundamentally reconstructs this mechanism. Votor compresses the finality process from 32 voting rounds into 1–2 rounds, achieving this through dual parallel paths. The fast path, when more than 80% of staked weight passes, can complete finality in about 100 milliseconds; the slow path, when 60%–80% passes, enters the second round and completes in about 150 milliseconds. Whichever path finishes first determines the network’s final result.

More critically, Votor moves the entire voting process off-chain. Validators communicate directly through a BLS signature aggregation mechanism, recording only an aggregated certificate of about 1,000 bytes on-chain—replacing the previous vote data of about 500 KB per slot. This means the roughly 75% of block space originally occupied by validator voting will be freed up for user transactions.

Ethereum’s Glamsterdam does not touch the consensus layer; instead, it focuses on structural parallelization changes in the execution layer. Its core mechanism is the block-level access list (EIP-7928). This mechanism allows nodes to pre-read read/write dependency relationships of transactions in blocks, assigning non-conflicting transactions to different CPU cores for parallel execution. At the same time, the gas limit is planned to increase from about 60 million to 200 million, and theoretical TPS could rise from roughly 1,000 toward ten-thousand levels.

These two routes reflect different engineering philosophies. Solana chooses to perform major surgery on the underlying consensus mechanism, pursuing an extreme performance breakthrough for every millisecond. Ethereum, while preserving existing consensus security, gradually “widens the lanes” at the execution layer, trading engineering robustness for predictable performance improvements.

### Generational differences in finality time

| Comparison dimension | Solana current architecture | Solana after Alpenglow | Ethereum current architecture | Ethereum after Glamsterdam |
| --- | --- | --- | --- | --- |
| Core consensus mechanism | PoH + TowerBFT | Votor + Rotor | PoS + Gasper | Maintains PoS + Gasper |
| Finality time | ~12.8 seconds | ~100–150 milliseconds | 12–15 seconds | 12–15 seconds |
| Voting mechanism | 32-round on-chain incremental voting | 1–2 rounds off-chain BLS aggregation | Casper FFG voting | Maintains Casper FFG |
| Block time | 400 milliseconds | Fixed at 400 milliseconds | ~12 seconds | ~12 seconds |
| Core bottleneck | Validator votes take up 75% of block space | Largely freed up | Serial transaction execution | Parallel execution gradually introduced |

Alpenglow compresses finality time from about 12.8 seconds to 100–150 milliseconds, achieving an improvement of roughly 80 to 100 times in latency. This metric pushes Solana’s transaction confirmation speed beyond Visa authorization-level performance. From an engineering perspective, 100-millisecond-class finality is a qualitative threshold—it not only means “faster,” but also means Solana in the latency dimension begins to have the capability to directly compete with traditional centralized financial infrastructure.

Ethereum does not pursue finality-time improvements of the same magnitude in Glamsterdam. Its strategy is to: reduce interaction costs at the user end through Layer-2 scaling, strengthen fairness in block building through ePBS, and lay the foundation for subsequent parallel execution via block-level access lists and a gas limit increase. This is an upgrade logic that prioritizes systemic structural adjustments rather than a single-dimension performance sprint.

### ePBS mechanism: Ethereum’s structural reform of block governance

Ethereum’s current “default mode” for block building is dominated by the MEV-Boost ecosystem—more than 80% to 90% of blocks are built and selected through a small number of relays. This forms a de facto concentration of power and amplifies the risks of transaction censorship.

Glamsterdam’s ePBS is designed to reshape block-building rules. Through EIP-7732, it directly embeds proposer-builder separation logic into Ethereum’s core protocol. Builders can participate in a permissionless manner; they only need to satisfy ETH staking requirements and commit to publishing blocks on time. Validators do not need to rely on external relays to select the best blocks. Multiple studies estimate that PBS embedded at the protocol level could reduce MEV extraction by about 70%.

Solana’s Alpenglow also addresses MEV, but along a different path. Under the current architecture, slot leaders can delay block production within a time window to sell more favorable transaction ordering to searchers. Alpenglow changes this game structure by introducing a timeout penalty mechanism: once a leader exceeds the timeout threshold, they not only lose the block-production rewards for that round, but also are deprived of leadership rights for subsequent slots. As Yakovenko described, the cost of delaying early slots is highest, while the cost of delaying the last slot is lowest, forming an asymmetric penalty structure.

## On-chain landscape: the economic gravity of the two public chains seen in data

Beyond the upgraded technical metrics, the two public chains’ on-chain economic activity shows distinct structural characteristics.

Ethereum maintains a clear advantage in total value locked (TVL). Its DeFi protocols have a TVL of about $45.4 billion, representing large amounts of long-term capital embedded in the ecosystem and compounding growth. Solana, in terms of transaction volume, demonstrates extremely high turnover speed. In the first quarter of 2026, its network processed about 25.3 billion transactions, compared with roughly 200 million on Ethereum’s mainnet in the same period.

The divide between these two economic models is becoming increasingly clear. Ethereum is a “capital-intensive” public chain—the moat comes from massive TVL settlement and institutional-grade liquidity infrastructure. Solana is a “speed-intensive” public chain—its competitiveness comes from extremely high transaction throughput and low-latency user experience. These two modes are not a zero-sum game, but in the same market environment they attract different types of capital and user groups.

## Breakdown of public opinion: what the market is arguing about

Around the two upgrades, the market discourse clearly shows several opposing viewpoints.

### Consensus disagreement one: Is the “big surgery” risk on Solana controllable?

Alpenglow achieved 98.27% validator support, one of the strongest proposals in Solana’s governance history. Behind this rare consensus, there is strong structural motivation: removing on-chain voting directly improves validator operating costs and increases the feasibility of participation for independent small nodes.

However, the opposing view also has logical grounds. Since Solana went live, it has experienced network outages multiple times. And the Firedancer client only entered mainnet production in mid-May 2026; it currently controls only about 7% of staking weight. Such a drastic consensus-layer replacement under real mainnet load may surface unforeseen boundary issues. Even though the testnet has expanded to 86 validators, the scale and performance under real economic incentives and adversarial conditions of thousands of validators on mainnet still differ in nature.

The core issue is that PoH provides not only timing information, but also the underlying agreement among validators on transaction ordering. The performance of Votor and Rotor in test environments still needs validation under real mainnet stress conditions.

### Consensus disagreement two: Is Glamsterdam “toothpaste squeezing” or “systemic engineering”?

Criticism from the Ethereum community mainly focuses on pacing. After the delivery of Pectra and Fusaka in succession, some of Glamsterdam’s core functions—such as ePBS and parallel execution—are still in a “gradual rollout” stage. Even if the gas limit is increased to 200 million, Layer-1 TPS improvements still rely on continued stacking of subsequent upgrades.

Supporters argue that this is precisely where Ethereum’s “engineeringized upgrade” route shines. With a twice-a-year hard-fork cadence, Ethereum breaks major changes that might otherwise cause network splits into modules that are verifiable and rollback-able, allowing ecosystem participants to form stable technical expectations. The successful delivery of the dual upgrades in 2025 has already validated the feasibility of this methodology at the engineering level.

### Consensus disagreement three: Is the SOL/ETH competition landscape undergoing a fundamental qualitative shift?

One view in the market is that if Alpenglow is successfully deployed on mainnet, Solana’s performance advantage will form a differentiated competitive barrier—especially in high-frequency trading and payments scenarios. Another view emphasizes that Ethereum’s TVL of about $45.4 billion provides structural demand support; this depth cannot be replaced by activity metrics alone.

## Industry impact analysis: how upgrades change the competitive landscape among public chains

### Impact on DeFi and transaction infrastructure

Solana’s 100–150 millisecond-class finality means that for the first time, a blockchain network has entered the same order of magnitude as centralized exchange order-matching systems in the latency dimension. This enables Solana’s native on-chain central limit order book to potentially compete with CEXs in latency, liquidity, and transaction fairness. The technical barriers for high-frequency trading strategies to migrate from CEXs to on-chain are gradually decreasing.

On Ethereum, implementing ePBS and increasing the gas limit will directly affect the quality of transaction execution in DeFi protocols. The expected improvement from reduced MEV extraction means that transaction losses for ordinary users on protocols such as Uniswap will decrease significantly, and the protocols’ own liquidity efficiency will also improve.

### Impact on validator economics

Solana’s Alpenglow reduces validator operating costs by eliminating on-chain voting, lowering the participation threshold especially for smaller independent validators. The launch of the Firedancer client further enhances network resilience: if the Agave client has vulnerabilities, Firedancer can keep the network running—an essential step in building long-term trust.

Ethereum’s Glamsterdam, through ePBS, helps independent validators remain competitive in the face of large staking pools. In addition, the FOCIL mechanism (fork choice with inclusion list) planned in the Hegotá upgrade will further strengthen the network’s resistance to censorship.

### Impact on public-chain competition narratives

Before 2026, the narrative competition between Ethereum and Solana mainly revolved around a binary framework of “decentralization vs high performance.” These two upgrades are breaking that simple division. Ethereum improves execution efficiency through Glamsterdam, while Solana improves network security and validator decentralization through Alpenglow. The strategic directions of the two public chains are converging to a certain extent—they are both moving into each other’s strong areas.

## Conclusion

In May 2026, two leading public chains almost simultaneously reached a similar historical turning point. Alpenglow and Glamsterdam are two answers to the same era-defining question: one pushes from the consensus layer toward reconstruction, the other refines with great precision in the execution layer. The future of public chains depends on the thickness of engineering capability, not on the skill of narrative packaging.

For those focused on the long-term development of the crypto industry, what is more worth observing than short-term price fluctuations is the direction and cadence of evolution of these underlying infrastructures. Every reconstruction of a consensus mechanism and every leap in execution efficiency is accumulating certainty as blockchain moves from a “verifiable ledger” toward a “globally operable economic infrastructure.” And when two different technical routes race within the same time window, the industry’s engineering level is being pushed to unprecedented heights.