The IRS safe harbor rule dropped in November 2025, and within three months, Fidelity, Grayscale, and 21Shares all filed to add staking to their Ethereum ETFs. The market cheered. But the code behind Fidelity’s staking architecture reveals a carefully engineered trade-off between yield and liquidity—one that most investors will not see until the first slashing event hits.
Let’s be clear: Fidelity Ethereum Fund (FETH) is not a new product. It is a $903 million SEC-registered ETF that now plans to stake up to 100% of its ETH holdings. The mechanics are straightforward: the trust holds ETH, three custodians (Anchorage Digital Bank, BitGo Bank & Trust, and Fidelity Digital Assets) handle asset custody and arrange staking, and three node operators (Blockdaemon, Figment, and Galaxy) run the actual validators. The reward split is 15% fixed for the service providers, with the remaining 85% retained by the trust, used to pay management fees, and the net distributed to shareholders in cash quarterly.
On paper, this is a textbook upgrade from a passive holding vehicle to an income-generating product. The IRS safe harbor rule removed the tax uncertainty—staking no longer disqualifies the grantor trust status as long as net rewards are distributed at least quarterly. Fidelity ensures that. The yield is real: ETH staking rewards come from protocol inflation and execution fees, not from token subsidies. Based on the current staking yield of around 3% annualized, FETH’s annualized gross staking revenue would be approximately $27 million, of which $4 million goes to fees, leaving $23 million before the ETF management fee. After subtracting the 0.25% management fee on $903 million ($2.26 million), the net distributable yield is roughly 2.3%—a decent premium for institutional investors who value simplicity and regulatory comfort.
But the real story is in the engineering. The “custodian – node operator” double layer is a design pattern I recognized immediately from my audit of DeFi staking contracts in 2020. Back then, I found a reentrancy vulnerability in a reward distribution function that allowed infinite minting. The fix was simple, but the lesson was that financial logic hides in state-changing functions. Here, the state changes are not in a smart contract but in the contractual agreements between Fidelity, the custodians, and the node operators. The code is not on-chain; it is in legal documents. And that is where the risk lies.
Core: The Liquidity vs. Yield Trade-off
The most critical technical constraint is that staked ETH cannot be withdrawn immediately. Validators face an unbonding period of several days before they can exit. Fidelity’s prospectus explicitly states that the fund may delay redemptions and may pay them in cash rather than ETH. This is a direct consequence of staking: the trust cannot instantly access the full $903 million in ETH if a wave of redemptions hits. The product design mitigates this by reserving some ETH for liquidity, but the 100% staking target is only a “maximum”—in practice, the staking ratio will likely fluctuate based on market conditions. There is no minimum staking requirement, which gives the fund flexibility to adjust.
From a quantitative perspective, the trade-off is clear. An investor who self-stakes 32 ETH earns approximately 3% annualized, minus a 10-15% fee if using a liquid staking pool. FETH’s net yield of ~2.3% is lower, but the investor gains a regulated, tax-simplified vehicle with no minimum stake. The 15% fee for the staking service is actually competitive—Lido charges 10% for its liquid staking protocol, but that includes the cost of decentralized node operators. Here, the three node operators are centralized entities, but they are also the most reliable in the industry. Fidelity chose them over building its own validator infrastructure, which tells me that the fund views staking as a commodity service, not a core competency.
Gas wars are just ego masquerading as utility. In this case, the gas costs are not the issue—the real cost is the opacity of the fee distribution. The 15% fee is split among the sponsor, custodians, and node operators, but the exact allocation is not disclosed. This lack of transparency is common in traditional finance, but for a product that claims to bridge DeFi yields, it feels like a step backward.
Contrarian: The Centralization Blind Spot
The contrarian angle is that Fidelity’s staking ETF is not a decentralization win. It is a centralization injection. By funneling millions of dollars in staking power through three node operators, Fidelity increases the concentration of validator sets. Blockdaemon, Figment, and Galaxy are already among the largest validators in the Ethereum ecosystem. Adding FETH’s potentially 90,000 ETH (at 100% staking) to their portfolios further concentrates power. This is not a theoretical risk—it is a direct threat to the principle of permissionless validation.
Moreover, the custodians’ liability is limited. The prospectus warns that custodians have limited responsibility for node operator actions. If a slashing event occurs due to a node operator error, the fund may lose ETH, and the custodians are not fully liable. The maximum loss is not quantified. Code does not lie, but it often forgets to breathe. The legal code here is missing the breath of explicit loss allocation.
Takeaway: The Stress Test
The real test for Fidelity’s staking ETF will come during a network stress event. A slash, a hard fork, or a sudden yield drop will reveal whether the multi-custodian, multi-operator structure is resilient or just complex. If the first slashing event is handled smoothly, it will set a strong precedent for institutional staking. If not, it could set back the entire staking-as-a-service model. The most elegant solution is the one that removes the most complexity. Fidelity has added complexity. Now we wait to see if it pays off.