Is ETH staking income taxable? Rev. Rul. 2023-14, explained
Yes — ETH staking rewards are ordinary income under IRC §61, recognized at fair market value when you gain dominion and control. Here's what Rev. Rul. 2023-14 actually says for solo validators.

The short answer is yes. ETH staking rewards are ordinary income under IRC §61, recognized in the taxable year you gain "dominion and control" over them, at their fair market value at that moment. This has been the IRS's stated position since August 2023, when it issued Revenue Ruling 2023-14.
This is educational content, not legal or tax advice. Tax law is fact-specific and changes. For positions on your own return, consult a qualified tax professional.
What the ruling actually says
Revenue Ruling 2023-14 (2023-33 I.R.B. 484, issued August 14, 2023) addressed a question the IRS had left open since Notice 2014-21 established that virtual currency is property for federal tax purposes: when staking rewards arrive, what happens to them under the tax code?
The ruling's answer has two parts.
First: staking rewards are gross income under IRC §61(a). That statute defines gross income as "all income from whatever source derived." Staking rewards qualify. They are ordinary income, not capital gains — you don't need to sell them to trigger a tax event. The moment you receive them, a tax event occurs.
Second: the "when" is dominion and control. Under IRC §451(a) and Treas. Reg. §1.451-2(a), income is includible in the taxable year in which the taxpayer actually or constructively receives it. Constructive receipt attaches when income is "credited to the taxpayer's account, set apart for the taxpayer, or otherwise made available so the taxpayer may draw upon it at any time without substantial restriction."
For PoS staking rewards, the IRS concluded that dominion-and-control attaches when the staker can actually dispose of, transfer, or use the ETH — not at some earlier accounting credit. The ruling applies to both direct staking (running your own validator) and exchange-facilitated staking. FMV is determined at the date and time dominion-and-control attaches, using methodology consistent with IRS Notice 2014-21.
What dominion and control means for a solo validator
If you run an Ethereum validator with standard 0x01 credentials, the architecture has two layers: the consensus layer (where rewards accrue to your validator balance per epoch) and the execution layer (where the ETH actually lands in a spendable wallet).
Prior to Capella (April 2023), ETH staking rewards were locked inside the beacon chain — you couldn't do anything with them. After Capella, the protocol introduced automatic sweep withdrawals: every few days, the consensus layer sweeps accumulated rewards above 32 ETH to your designated execution-layer withdrawal address. That sweep credit — when the ETH lands in the address you control — is the clearest "dominion and control" event. You can now spend it, transfer it, or exchange it without restriction.
The withdrawal-time reading (the dominant CPA default): Income is recognized when the sweep transaction credits your execution-layer withdrawal address. Under Treas. Reg. §1.451-2(a), the per-epoch validator-balance credit — before Capella enabled withdrawals, or before your specific withdrawal address was set — was subject to substantial restriction. You couldn't draw on it at any time, because withdrawal wasn't technically possible. On that reading, recognition attaches at the sweep, not at each epoch's credit. This is the conservative default in common CPA practice following Rev. Rul. 2023-14, and it is TrueStake's Phase 0 filing default.
The earning-time reading (an alternative within the ruling's scope): Each per-epoch validator-balance credit is the operative event. The argument: after Capella, a solo staker can initiate a partial withdrawal at any time by changing their withdrawal credentials — there is no substantial restriction. On that reading, dominion-and-control attaches at each epoch, not at the sweep. Rev. Rul. 2023-14 does not foreclose this reading, and some practitioners apply it.
Both readings are grounded in the same ruling and the same underlying statute. The practical difference: the withdrawal-time approach produces fewer, larger income events; the earning-time approach produces hundreds or thousands of smaller events across the year. The total ETH income figure will be similar under both policies; the difference is in timing and FMV.
Fair market value: what it means and why the timestamp matters
Under Rev. Rul. 2023-14 and IRS Notice 2014-21, FMV is determined as of the date and time dominion-and-control attaches — not at year-end, not at the time you file, not at the time you eventually sell.
This has a concrete implication: if ETH's price was $3,200 when your sweep landed on March 15, you recognize $3,200/ETH worth of ordinary income on that date, regardless of what ETH trades for later. If you later sell the ETH for $4,000, the $800 gain above your $3,200 cost basis is a separate capital-gains event. If you sell for $2,800, the $400 loss is a capital loss. The income event at receipt is separate from any subsequent gain or loss on disposition.
This is the same treatment as any other property received as compensation. If your employer pays you a bonus in restricted stock, you recognize ordinary income at FMV when the restrictions lapse. Staking rewards work the same way — property received with no strings attached triggers ordinary income at FMV at receipt.
The FMV methodology matters for audit defensibility. "Close of day" prices and "daily average" prices are not the same thing, and neither is necessarily the "date and time" of receipt the ruling requires. A per-event timestamped price — pulled from a liquid market data source at the actual hour of the sweep — is the most defensible approach. Document the source and apply it consistently across every recognition event in the year.
Why the validator pubkey matters (and wallet scans miss it)
Most crypto-tax tools work by scanning wallet addresses for incoming transactions. For post-Capella solo validators with 0x01 credentials, that approach captures the sweep withdrawals — but it misses two categories of income that can be material.
MEV-Boost block proposals pay the fee recipient, which may be a different address from your withdrawal address. If you've configured a separate fee-recipient address or use a SaaS-managed validator service that routes MEV differently, a wallet scan of your withdrawal address misses every MEV tip. At current MEV levels, a single block proposal can produce a tip larger than a month of attestation rewards. That's not rounding error — it's a missed income event that doesn't appear anywhere in a wallet scan.
Pre-Capella accrual — every epoch from your validator activation through April 18, 2023 — never appeared in any execution-layer wallet. Rewards accumulated inside the beacon chain and were paid out as a lump-sum sweep after Capella. A wallet scan sees one large deposit and applies the FMV of the sweep date to the entire amount, collapsing two or more years of staking income into a single day's FMV. Under Rev. Rul. 2023-14's dominion-and-control standard, that methodology is hard to defend: the IRS expects FMV at the time of receipt for each income event, not a single lump-sum FMV applied retroactively.
The complete data source for validator income is the beacon API, keyed by validator pubkey, not the execution-layer withdrawal address. Starting from the pubkey, an ingestion pipeline can reconstruct the full reward history — attestation rewards, sync-committee rewards, block-proposal rewards, MEV tips — by event type, by epoch, with wei-level precision. That is the starting point for an audit-defensible record.
The open question: EIP-7002 and 0x02 credentials
The Pectra upgrade (May 2025) introduced 0x02 credentials, which change how withdrawals work. Validators with 0x02 credentials are not subject to automatic sweeps — their balances grow inside the beacon chain until the operator explicitly triggers a withdrawal using a transaction on the execution layer (EIP-7002).
This creates a timing question Rev. Rul. 2023-14 does not reach: does dominion-and-control attach when the staker signs and submits the EIP-7002 trigger transaction, or when the ETH actually credits the execution-layer address?
Two competing readings exist:
- Trigger-transaction timing: The act of signing the trigger transaction demonstrates the staker's unilateral control — they have the ability to extract the ETH whenever they choose. Some practitioners argue this is the operative dominion-and-control event.
- Settlement timing: Dominion-and-control requires the ETH to actually credit the EL address — the same standard applied to 0x01 sweep receipts. The trigger initiates the process but doesn't transfer the ETH; the withdrawal credit does.
Rev. Rul. 2023-14 was issued in 2023, predating Pectra by roughly two years. It does not address EIP-7002, trigger transactions, or 0x02 credentials. This is genuinely unsettled. Do not assert a definitive recognition moment for 0x02 triggered withdrawals without counsel confirmation. This question is tracked as an open item in TrueStake's research library, pending IRS guidance or counsel review.
For 0x01 validators — the large majority of the current validator set — the withdrawal-time default (recognition at the sweep credit) remains the settled, defensible approach per Rev. Rul. 2023-14.
What about Jarrett?
The Jarrett refund gets cited frequently as evidence that staking rewards might not be taxable until sale. It does not support that conclusion.
Here is what happened: Joshua and Jessica Jarrett reported their Tezos staking rewards as ordinary income, paid the tax, and then filed for a refund arguing that staking rewards are newly created property — like a farmer's crop, not taxable until sold. The IRS issued a refund without explanation (a common tactic to moot litigation before a potentially unfavorable ruling) and the Jarretts took the money. No opinion on the merits was issued. The IRS did not concede the argument.
Rev. Rul. 2023-14 then addressed and rejected the "newly created property" theory directly. The dominion-and-control standard applies. The Jarrett argument is affirmatively foreclosed under current IRS guidance. Filing on the basis that staking rewards are not income at receipt is not a defensible position under current law as stated by the IRS. Further Jarrett-related litigation is ongoing, but as of this writing there is no published federal court opinion on the merits, and Rev. Rul. 2023-14 is the controlling IRS authority.
What audit-defensible recognition actually requires
Rev. Rul. 2023-14 tells you what the rule is. Applying it defensibly requires the data to match.
Per-event records, not an annual tally. Each recognition event needs its own income record: the timestamp, the ETH amount in wei, the event type (sweep withdrawal, MEV tip, sync-committee reward), the FMV at that moment, and the USD income figure. An annual total without underlying event records is not defensible.
FMV at the event timestamp. The ruling requires FMV at the date and time of receipt. Document your price source, the price applied, and the methodology — and apply it consistently across every event in the year.
Reconciliation to on-chain facts. Every income figure should trace to a specific on-chain receipt or beacon API response, with a reconciliation confirming they match. A number you can explain and source is a number a CPA can defend. A number that came from "I added up my wallet receipts" is not.
Validator-pubkey-level tracking, not wallet-level. As described above: MEV and pre-Capella accrual are invisible to wallet scans. Complete reporting requires starting from the beacon API.
Recognition-policy documentation. Whether you use the withdrawal-time or earning-time approach, document it and apply it consistently throughout the year. Switching approaches mid-year or mid-period is itself an audit risk.
The practical consequence of 1099-DA is that exchanges reporting your staking income to the IRS are also creating a matching record. If your return doesn't account for the same amounts the broker reported, that discrepancy gets flagged. Even if you don't receive a 1099-DA (solo validators don't — see our policy primer), the infrastructure for data-matching is maturing. Underdocumented returns get harder to defend every year.
Sources: IRS Revenue Ruling 2023-14, 2023-33 I.R.B. 484 (Aug. 14, 2023) — primary authority on staking-reward income recognition; IRS Notice 2014-21 (Mar. 2014) — virtual currency as property; IRC §61(a) — gross income definition; IRC §451(a) and Treas. Reg. §1.451-2(a) — actual and constructive receipt standards. EIP-7002 / 0x02 recognition timing is an open question not addressed by Rev. Rul. 2023-14 — treatment above is descriptive of competing readings, not a filing recommendation. This article is educational, not legal or tax advice — consult a qualified tax professional for positions specific to your return.
Citations
- [1]IRS Rev. Rul. 2023-14· Staking rewards are ordinary income at dominion and control
- [2]IRS Notice 2014-21· Virtual currency treated as property for federal tax purposes
- [3]IRC §61(a)· Gross income means all income from whatever source derived
- [4]IRC §451(a)· General rule: items of gross income included in the taxable year received
- [5]Treas. Reg. §1.451-2(a)· Constructive receipt: income credited to account or set apart without substantial restriction