Why Coinbase's Staking Business Is Structurally Getting Smaller

ETH Staking Mechanics, Yield Compression, and the Math Problem That Consensus Has Partially Priced

Written by Siddhant Shah on

Thesis

Blockchain rewards — primarily ETH staking — generated $677.4M in FY2025. That makes staking the fourth-largest revenue line in the company. Coinbase earns a commission of approx. 29-35% on gross staking rewards, making this high-margin recurring revenue with no direct customer acquisition cost.

The problem is that unlike most of the other flywheel layers, staking revenue is being compressed by the success of the very network it serves. And the math is working against Coinbase regardless of what management does.

How the Revenue Works

Coinbase operates as a staking-as-a-service provider. Users deposit ETH with Coinbase, which stakes it on their behalf across validators Coinbase operates on the Ethereum network. Ethereum allocates block rewards proportionally to stake weight — if Coinbase controls 10%10\% of total staked ETH, its validators receive approximately 10%10\% of total block rewards. Coinbase retains approx. 2929-35%35\% of gross rewards and distributes the rest to users. The user-facing yield is approximately 1.89%1.89\% net as of early 2026.

This structure works under one crucial assumption: the total staked ETH supply, and thus the validator set, remains relatively stable. When it grows — which it has done every quarter since Ethereum’s transition to proof-of-stake in September 2022 — the per-validator share of a fixed block reward pool shrinks.

The Yield Compression Dynamic

Ethereum’s block reward structure is essentially fixed in aggregate at any given issuance rate. More validators means the same aggregate reward pool divided among more participants. Per-validator yields decline.

In early 2024, ETH staking APY was approx. 5%5\%+ for validators. By early 2026, it’s fallen to 3.53.5-4.2%4.2\% APY. The decline isn’t accidental — it’s the predictable output of staking participation growing from approx. 20%20\% of total ETH supply to more than 30%30\%, with 3636M+ ETH currently staked.

The structural compression rate is approximately 2020-3030 basis points per year. This is not a cycle-driven phenomenon that reverses when sentiment improves. It’s a network-level mechanic: more validators implies lower yields per validator. The only mechanisms that could reverse it are mass unstaking (no precedent at scale since The Merge) or an Ethereum protocol decision to increase issuance rates (which the development community has historically opposed on monetary policy grounds).

Neither is likely.

FY2025 Actuals and the Q4 Signal

FY2025 blockchain rewards revenue was $677.4677.4M — a modest decline from FY2024 despite growth in on-platform ETH. The quarterly trajectory is more informative. Q4 2025 staking revenue was $152152M, down 18%18\% sequentially from Q3 2025. That magnitude of quarter-over-quarter compression, in a quarter where ETH prices were relatively stable, is consistent with yield compression driving a meaningful portion of the decline — not asset price alone.

Management’s Q1 2026 guidance gives $550550-630630M for total subscription and services revenue, which includes staking alongside USDC, custodial fees, and other recurring streams. Isolating staking within that range isn’t possible without additional disclosure. But the implied trajectory is flat to slightly down year-over-year.

The Consensus Estimate and the Downside Risk

The consensus FY2026E for blockchain rewards I’ve seen is $595.2595.2M — a 12%12\% year-over-year decline from FY2025. The methodology presumably applies a lower yield assumption to a roughly stable or slightly higher on-platform ETH balance.

The risk is the compound scenario: yield compression at 2020-3030bps combined with flat or declining ETH prices. Every dollar of ETH appreciation partially offsets yield compression by making the same percentage commission apply to a larger dollar-denominated base. In the scenario where ETH stays at early 2026 levels and validator yields fall another 3030bps, the dollar impact on staking revenue is approximately $2020-3030M below consensus.

cbETH and the Liquid Staking Market

Coinbase’s liquid staking token — cbETH — allows users to receive staking rewards while maintaining token liquidity, without locking ETH for the validation period. cbETH TVL as of early 2026 is approximately $543543M, a modest figure relative to Lido, which commands a majority of the liquid staking market.

The strategic question is whether cbETH adoption can grow as a share of on-platform ETH staking. If it does, Coinbase retains better user engagement and stickiness. If Lido or other competitors continue to capture flow, staking revenue becomes increasingly dependent on raw on-platform balance rather than product differentiation.

What to Watch

Monitor total ETH staking participation rate monthly. If participation crosses 35%35\% of total ETH supply, the yield compression rate is likely to accelerate. The secondary indicator is Coinbase’s quarterly staking revenue relative to ETH price: if revenue declines in a quarter where ETH prices are flat or up, yield compression is outrunning asset appreciation — structural deterioration, not a market cycle effect.

The consensus FY2026E of $595.2595.2M should be treated as the midpoint of a range with meaningful downside, not a conservative estimate.

My model estimates $249.8249.8M — substantially below consensus — driven by formula: 2.72.7M ETH in custody × $2,5002,500 average ETH price × 3.7%3.7\% net user yield. The gap versus consensus reflects the compounding effect of yield compression applied to the full balance, which most consensus estimates do not fully capture.