Ethereum Validator Performance Report Q2 2026
TL;DR: This report covers our Ethereum validator performance metrics for Q2 2026. Note that while some Coinbase staking customers choose to delegate to our partner staking providers, the data in this report applies solely to Coinbase validators.

At a Glance
Reliability: 99.97% uptime vs. 99.76% network - +21 bps; 0 slashing or double-signing events since inception.
Rewards: 2.87% APY vs. 2.80% network - +7 bps, for Q2 2026.
Scale & distribution: 4.76M ETH staked to Coinbase validators - 12.16% of total staked ETH, both averaged over Q2; 4 countries and 2 cloud providers, each region spanning multiple availability zones. Coinbase is committed to never exceeding 30% network penetration, measured against Coinbase-operated stake.
Client & relay diversity: 2 consensus clients (Lighthouse 72%, Prysm 28% of our fleet), 3 execution clients (Nethermind 39%, Reth 33%, Geth 28% of our fleet), and 7 MEV relays, with OFAC screening available as an option.
Security-first: double-signing protection, and a validator orchestration system that moves validators between data centers without downtime.
Our Staking Philosophy
TL;DR: Protect the capital, protect the network. On Ethereum, they're the same job.
When a single client runs a large share of validators, one bug can penalize those validators together and weaken the chain at the same time. There's no tradeoff between safety and performance here - they rise and fall together, so we build for both.
That principle also defines what we won't do. We don't pursue strategies that concentrate risk, compromise network integrity, or chase short-term gains at the expense of long-term sustainability. Instead, we optimize for repeatability: performance that holds up across quarters, clients, and market conditions.
Our Q2 results bear that out. We ran 99.97% uptime against a network average of 99.76%, with zero slashing or double-signing events, spread across 2 consensus clients, 3 execution clients, and 7 relays. For institutions and ETF issuers, that track record matters as much as access - which is why we report it against the network, every quarter.
Rewards
TL;DR: In Q2 2026, our institutional validator set delivered 2.87% APY against a network average of 2.80%, or +7 bps.
Staking rewards on Ethereum come from two sources: consensus layer issuance, which pays validators for attesting and proposing, and execution layer rewards, which are the priority fees and MEV captured in the blocks our validators propose. Consensus layer issuance is protocol-determined and similar for every well-run validator. Execution layer rewards are variable, and they are where relay strategy and validator configuration change the outcome.
We report APY as a simple average across our institutional validator set, the Coinbase-operated validators serving our institutional clients, compared against the network average over the same window.

Uptime and Participation Rate
TL;DR: In Q2 2026, Coinbase validators outperformed the network average in all dimensions.
Participation rate indicates how well validators perform their consensus duties. It is measured by the percentage of assigned attestations that a validator successfully signs, submits, and gets included in a block. Participation rate is a reliable indicator of validator uptime; the two are interchangeable in our definitions.
Attestations are one of three consensus duties we track. The other two:
Proposing blocks: Signing and submitting blocks produced by our MEV relays
Participating in sync committees: Producing additional signatures for blocks to enable light clients to sync quickly and trustlessly
In Q2, Coinbase validators outperformed the network average on all three duties: +21 bps on attestations, +21 bps on block proposals, and +37 bps on sync committees. Sync committee participation trailed the network in Q1 and moved ahead of it this quarter.

Validator Distribution & Disaster Recovery
TL;DR: We run validators in 4 countries across 2 cloud providers, each region spanning multiple availability zones.
To help maintain a truly distributed and decentralized Ethereum blockchain, we distribute our validators across several regions, each with multiple availability zones. This also helps compartmentalize potential outages and accommodates customers who require validators to be operated in specific regions for regulatory purposes. Our validators operate in:

Ireland 🇮🇪
Japan 🇯🇵
Netherlands 🇳🇱
Singapore 🇸🇬
Additionally, we distribute our validators between 2 cloud providers: AWS and GCP, which provides diversity, helps further compartmentalize outages, and mitigates the risk of a technical issue impacting our ability to run workloads on a given cloud provider. This distribution strategy also helps mitigate risk in the event of a cloud provider curtailing the hosting of blockchain validators.
If a prolonged issue with a cloud provider or given region were to occur, our validator orchestration system enables us to safely move our validators between data centers. This system has yet to be required in response to an outage. However, it has a proven track record when used for routine validator migrations, either at the request of customers or for scheduled maintenance.
Client Diversity
TL;DR: We run 2 consensus clients and 3 execution clients, so no single client defect reaches our entire fleet.
Every Ethereum validator runs two pieces of software: a consensus layer (CL) client, which handles attestations, block proposals, and sync committee duties, and an execution layer (EL) client, which builds and validates the transactions inside each block. A defect in either one affects every validator running it at the same time. Running a mix of clients at both layers is how we bound that risk, and it matters at two levels.
For our customers, diversity limits the impact radius and reach of any single defect. For the network, client concentration is a consensus problem rather than an availability one. When one client runs a large enough share of all validators, a single defect can stall finality,
Client selection is one of the few decisions a large operator makes that changes that risk for everyone, and we treat it accordingly. Our ETH validators support:
Consensus clients: Lighthouse, Prysm
Execution clients: Nethermind, Reth, Geth

Relay Diversity
Relay diversity (via MEV-Boost / Commit-Boost) reduces reliance on a single relay's uptime or policies, improving redundancy and reducing centralization risk in block building. Using multiple reputable relays (configured appropriately) can increase the likelihood that proposers receive and select the best available bid, which can improve execution layer rewards (priority fees & MEV) that are otherwise variable.
Our validators are connected to 7 unique MEV relays:
We offer OFAC screening as an option, further increasing the diversity of relays for the subset of our customers that need filtering.
Looking Ahead: The Glamsterdam Upgrade
Ethereum's next major network upgrade, Glamsterdam, is expected in the second half of 2026 and represents one of the most significant structural changes to the protocol since the Merge. It restructures how blocks are proposed and propagated and lays the groundwork for parallel transaction processing, which together support higher throughput and more predictable validation under heavier network loads. Several of its proposals are particularly noteworthy:
Enshrined proposer-builder separation (ePBS, EIP-7732): Enshrines a bidding mechanism for blocks directly in the protocol, and separates the propagation of the consensus block from the execution payload. Staging propagation in this way gives the network more time to distribute and execute blocks, improving performance under high load and easing data availability concerns associated with larger payloads and blobs. While this is a change to the dynamics of the block building market, we see this primarily as a benefit to overall network health, supporting decentralization and resilience. How the relay and builder market will evolve under this model is not yet certain, given MEV marketplace is an emergent behavior. That said, we expect to continue using extra protocol auction formats, such as MEV-boost, for most block building. Low latency drives price performance, and as such we expect these to continue to produce the highest value blocks, but are maintaining flexibility in our approach so that we continue to deliver best in class returns for our customers, and performance for the overall network no matter how the dynamic shifts.
Block-level access lists (BALs, EIP-7928): Give every block an upfront map of the accounts and storage its transactions touch, enabling parallel transaction processing, faster node synchronization, and more predictable validation as block sizes grow.
ETH transfer logs (EIP-7708): Emit a standard log for every ETH transfer, making native ETH as straightforward to track and reconcile as ERC-20 tokens and simplifying accounting, monitoring, and reporting.
Increased exit and consolidation churn (EIP-8061): Currently being considered for inclusion rather than confirmed for the upgrade. If adopted, it would raise the rate at which staked ETH can exit the active set, shortening the time required to unstake large positions during periods of high demand and improving liquidity planning and the ability to rebalance or respond to market events.
The first three proposals above are scheduled for inclusion, while the exit churn change remains under consideration, and the final scope and timing of the upgrade are still being determined. As with prior upgrades, we are actively tracking Glamsterdam across test networks and preparing our validator infrastructure so that the transition is seamless for our customers, consistent with our commitment to operating on a disciplined and resilient foundation built for the long term.
Stake with Coinbase
Coinbase Prime integrates staking with secure custody, trading, financing, and other prime services to provide a turnkey solution for institutions. Prime makes it easy for institutions to seamlessly stake their assets with just the click of a button.




