Ireland published the roadmap for its new tax-advantaged investment accounts on Monday 31 August 2026, and crypto did not make the list. Listed shares, bonds, ETFs, other investment funds and insurance-based investment products are all in. Crypto assets, derivatives and interest-bearing cash deposits are explicitly out, on the stated grounds that they are highly complex and risky products.
Nothing was exploited to produce that outcome. No contract failed, no oracle lied, no venue froze withdrawals. A government published a list of eligible assets, and an entire category landed on the wrong side of it. The argument that follows will be the obvious one: does crypto deserve the tax break or not. That argument is the asked half. The interesting half is what the list is actually measuring.
TL;DR
- Ireland’s Savings and Investment Account, set out in a roadmap published on 31 August 2026 by Tánaiste and Finance Minister Simon Harris with Minister of State Robert Troy, launches in 2027 and excludes crypto assets as “highly complex and risky”.
- The exclusion grades the wrapper, not the asset: a regulated fund with bitcoin exposure can sit inside the account while the same bitcoin held directly cannot.
- The property that qualifies you is third-party attestation, not verifiability. Tax codes have a category for “someone else reports it for you” and no category at all for “anyone can check it themselves”.
- On-chain gaming hits the identical category error: a licence makes a server-side RNG “fair”, while a Chainlink VRF draw that a stranger can replay from public data is “unregulated”.
- Satoshie publishes odds in-contract and proves randomness with Chainlink VRF, which is verification without a wrapper. That is the forty-second unasked half of fairness.
What Ireland actually announced
The Savings and Investment Account is a retail participation product, not a crypto policy. Irish households sit on somewhere between €170bn and €197bn in bank deposits depending on the measure you use, while holding roughly 2.3% of financial assets in direct listed equity and debt against an EU average nearer 7.5%. That gap is the problem the scheme exists to close.
The mechanics are sensible. It is open to Irish tax residents aged 18 and over, with no minimum contribution, no lock-in and portability between providers on a tax-neutral basis, and it replaces the eight-year deemed disposal rule, the single most user-hostile feature of Irish fund taxation, with a tax-free threshold and a low flat rate above it. The rate and threshold arrive with the Budget on 6 October 2026. For comparison, capital gains currently run at 33% and fund exit taxes reach 41%.
So an ETF buyer inside the account gets a materially better deal than a crypto buyer outside it, who stays on 33% CGT with no threshold and no simplification. The policy has a direction, and it is not subtle.
The half nobody is asking about
Here is the part worth sitting with. “Highly complex and risky” is doing almost no work in this decision. Leveraged equity funds are risky. Long-dated corporate bonds in a rate cycle are risky. A single-country emerging markets ETF is risky. None are excluded, because risk is not the boundary being drawn.
The boundary is the wrapper. The European Commission’s September 2025 recommendation, which Ireland is following, excludes crypto but carves out tokenised versions of financial instruments that would otherwise qualify. Read that carefully, because it is the tell. A tokenised share is fine. The token is not the disqualifier. What qualifies an asset is what sits behind it: an issuer, a regulated market, and above all a provider who files the return on your behalf.
Which means the same economic exposure gets two different tax treatments depending on the container it arrives in. Hold bitcoin exposure through a regulated fund structure and you are inside the scheme. Hold the coin itself, the thing the fund is a claim on, and you are outside it. The state did not price the asset. It priced the paperwork wrapped around the asset.
Attestation is not verification
Strip out the politics and there is a real design principle underneath. The reason provider reporting is the load-bearing feature is that a tax system needs a party it can hold responsible for a number. Not a true number, a reported number, from an entity with an address and a licence to lose. That is attestation. Someone vouches for your position and accepts the consequences of vouching wrongly.
Verification is a different property entirely. Your on-chain position is not attested by anyone. It is a fact any stranger with a node can confirm without permission, without trusting you and without trusting a provider. By every epistemic standard it is stronger evidence than a broker statement, which is a claim about a database you cannot read.
And it does not qualify, because there is no category for it. The tax code can recognise “a licensed intermediary told us” and it cannot recognise “the ledger is public and you can check”. Verifiability is not scored badly here. It is not scored at all.
Where this lands for on-chain gaming
If that structure feels familiar, it should, because gaming has been living inside it for years. A licensed casino’s random number generator is considered fair because of the licence, the jurisdiction and a periodic audit certificate. All three are attestations. A test house sampled the RNG at some point in the past, in an environment the operator controlled, and signed a document. You cannot check the draw you personally played. You can check that somebody credible said the machine was fine, once.
A Chainlink VRF draw is the opposite arrangement. Satoshie’s coinflip resolves on randomWords[0] % 2. A raffle winner comes from keccak256(VRF word + prior blockhash) % ticketsMinted. The randomness request and its fulfilment are both on-chain, with a cryptographic proof verified by the contract before the result is usable. Nobody vouches for it. Nobody has to. You replay it yourself from public data, for the exact draw you played, months later, with no cooperation from us.
That is a strictly better fairness property, and it carries strictly less institutional weight, for precisely the reason Ireland just demonstrated at national scale. Regulatory categories grade containers. Verification grades artefacts. The two frameworks barely touch.
What to do with that
Not sulk about it, for a start. The pattern is legible and it is not malice. Institutions built their fairness machinery in a world where checking things yourself was impossible, so they built a chain of accountable parties instead. That machinery works reasonably well. It simply has no input port for a proof.
The mistake would be chasing the wrapper: bolting on intermediaries and certificates so verifiable systems can be graded by a rubric designed for unverifiable ones. That trade hands back the exact property that made the thing worth building.
The better move is to make the artefact so easy to check that the absence of a wrapper stops reading as an absence of anything. Publish the odds in the contract rather than a terms page. Prove the randomness on-chain rather than in an annual PDF. Let the player verify their own draw instead of trusting our summary of it.
Ireland’s savers get a genuinely better product in 2027, and almost none of them will ever independently confirm a single number inside it. Every Satoshie player can confirm every number in every game they play, today, and no budget speech will award them a category for it. Build for the check anyway. Categories move slower than proofs, but they do move.
📷 Photo by Henrique S. Ruzzon (@ruzphotos) on Unsplash
