The National Assembly passed it late. A budget amendment, buried in the fiscal annex, taxing "certain stablecoin transactions." That is the exact phrase. Certain. Undefined. Unpriced. No rate disclosed in the text surfaced to the public. No effective date. No scoping. Meanwhile the exit-tax proposal lives to fight another day, and the self-hosted wallet reporting mandate sits in committee limbo. Three proposals, one headline, wildly different fates.
Most of the English-language coverage flattened all of this into "France taxes crypto." Wrong read. The only clause with real execution teeth is the stablecoin swap amendment — because it is the only one backed by data the tax authority can actually pull off a server. Everything else is paper. And paper does not collect revenue. Paper collects data.
I have watched this pattern before. In 2017 I sat on the 0x order-matching logic for 72 hours and found a front-running hole that looked clean on a whiteboard and catastrophic the moment real flow hit it. Same shape here. A rule that reads elegantly in a legislative annex, that produces absurd outcomes the instant you run actual capital through it. The gap between the whiteboard and the order book is where fortunes die. France just drew a line on a whiteboard.
Here is the regime that exists today. France taxes crypto assets under Article 150 VH bis — "actifs numériques." The taxable trigger is narrow and, until now, clean: a disposal counts when you exit into fiat. Crypto-to-crypto swaps defer. You move BTC into ETH, nothing realizes. You move BTC into USDC, nothing realizes either — under current law that is a same-category exchange, a deferral event. Flat 30% single rate, the PFU, on the gain when fiat finally leaves the system. That is the baseline everyone keeps forgetting to mention.
Now layer the EU framework on top. MiCA — Regulation 2023/1114 — reclassified stablecoins as EMTs and ARTs. Electronic money tokens. Asset-referenced tokens. Payment instruments, in regulatory language. Not securities. That matters enormously for the tax question, because MiCA tells you a stablecoin is a means of payment while the French amendment is quietly telling the tax code it is a fiat-equivalent exit. Two regulators. Two philosophies. One asset caught in the middle.
Then the data plumbing. DAC8 — the EU tax administrative cooperation directive — applies from January 1, 2026, with first reporting in 2027. The OECD CARF framework starts first exchanges in 2027. France is legislating in this window for a reason. It is pre-building domestic law to catch the data pipe when it opens. This is not a spontaneous fiscal grab. It is preparation. And preparation with a deadline behaves very differently from a random tax raid — it has a schedule, and schedules get executed.
So let me do what I actually do and audit the execution mechanics, because that is where this event lives or dies.
Stablecoin swaps executed on a centralized exchange or a MiCA-licensed CASP leave a centralized record. DGFiP, the French tax authority, can request it. Batch-verifiable. High executability. Self-hosted wallet reporting has no third-party intermediary — it depends entirely on taxpayer self-declaration plus on-chain clustering analysis. Low executability. Enormous enforcement cost. These two provisions were reported side by side, and that adjacency is the single most misleading thing about the entire story. It makes readers believe both are equally close to landing. They are not. One rides licensed intermediaries that already file reports. The other is aspirational compliance dressed as law.
Here is the part the headlines missed entirely. The amendment is not a new tax. It is a rewrite of what counts as a taxable trigger event. Under the current code, crypto-to-stablecoin is not a disposal. The amendment makes it one. That is not a parameter tweak. That is a reclassification of stablecoins as fiat-equivalents for tax purposes — the precise opposite of how MiCA frames them. You are not adjusting a rate. You are changing what a stablecoin is in the eyes of the state.
And the moment you do that, you hit the killer detail: stablecoin is not stablecoin. USDT into USDC. EURC into USDC. Economically neutral rotations. No gain, no loss, no substance. Tax them and you have manufactured a taxable event with zero economic content. The taxpayer is now obligated to maintain multi-currency cost-basis ledgers across assets that all trade at a dollar or a euro. Compliance cost exceeds the tax itself. I have audited enough protocols to know exactly how this ends: people stop doing the neutral thing. The rotation stops. The market freezes at the seams.
Run the computational burden. FIFO or LIFO election. Cross-wallet aggregation. DeFi interaction characterization — is providing liquidity a disposal? Is a liquidation a disposal? Cost basis on the same token living across six chains at six addresses. This is a tax on computation. The nominal rate is irrelevant when the accounting overhead dwarfs it.
Now the economics, which are worse than the mechanics. The core effect is a lock-in effect, not a revenue effect. Tax the neutral swap and rational holders do one of two things: they hold and never rotate, or they change tax residency. Velocity drops. The tax base contracts. This is the Laffer curve wearing a suit — raise the friction, shrink the flow, collect less than you projected. The projection always assumes static behavior. Behavior is never static in a market where capital moves in seconds.
The DeFi composability hit is the piece the market is sleeping on. Every core DeFi operation runs through a stablecoin as intermediate medium. Collateral posting. Liquidation. Portfolio rebalancing. Cross-protocol arbitrage. If each step is a taxable event, the marginal cost of a French resident participating in DeFi rises structurally. Not "a little more tax." A moat around the entire activity. And it lands hardest on the thinnest liquidity in the market: euro stablecoins. EURC, EURCV — these books are shallow. Any friction gets amplified. Market makers do the math and leave the tax jurisdiction. You do not tax a market maker. You evict one.
Which brings me to the contradiction nobody in Paris wants to say out loud. France was a lead negotiator on MiCA. France hosts SG Forge, the Societe Generale subsidiary that issues EURCV — one of the flagship euro stablecoins in the entire bloc. France has spent years positioning Paris to take London's crown as the EU crypto hub, using the AMF's early PSAN licensing regime as the on-ramp. And now the same state is taxing the use of the very instrument its own champion bank issues. One hand builds the stablecoin industry. The other hand taxes its circulation. That is not policy. That is two ministries at war, and the tax ministry won the drafting session.
The single most exposed company in this entire affair is the one no one is naming: Ledger. A French company whose entire brand is "not your keys, not your coins" — self-sovereignty as product. If its home country passes a self-hosted wallet reporting mandate, that company faces a domestic legitimacy problem. Not a financial one. A narrative one. Its core promise collides with its own government's disclosure regime. I built a regulatory network in DC in early 2025 — former SEC staffers, bank regulators, people who read statutory language for a living — and the transferable lesson is this: when a jurisdiction's tax code contradicts its own industrial strategy, the industry reads the tax code, not the press release. Ledger's customers will read the reporting clause.
Here is the contrarian read, and it is the one I would put money behind. The real output of these provisions is not revenue. It is a data baseline. Self-hosted wallet reporting, in a world where the state cannot effectively tax those assets anyway, produces almost no tax and an enormous amount of information. Holder profiles. An information asset. This is the classic administrative ratchet: secure disclosure first, tighten enforcement later. The tax is the excuse. The database is the prize.
And second: the three proposals are not one package, despite being reported as one. Their legislative maturity is wildly different. Watch the pipeline. National Assembly passage is not law. The Senate — the upper house, whose finance committee has an independent streak on crypto clauses — has not even weighed in. Then the CMP, the joint reconciliation committee, where texts get surgically rewritten. Then the Conseil constitutionnel, which has a history of striking budget riders on cavalier budgetaire grounds and tax-equality violations. And hanging over all of it, the government's ability to invoke Article 49.3 and shove the whole budget through on a political vote rather than a technical one. That is the real risk variable, and it has nothing to do with crypto.
My read on outcomes: the stablecoin clause probably survives in some form, because it has data behind it and the executive wants the headline. The exit tax probably gets gutted — extending exit tax to crypto requires solving valuation timing, constitutional tension with freedom of movement and property rights, and double-taxation treaty interplay, and that is a heavier legal lift than the stablecoin trigger. The self-hosted wallet reporting gets delayed or deleted, because it is unenforceable and constitutionally shaky — forcing citizens to self-declare private key control with no cross-verification brushes against privacy rights and the principle against self-incrimination.
So the tradable signal is not the vote. The tradable signal is the final text. And here is the tell: if the final text never defines "stablecoin swap," the ambiguity is not a drafting failure. The ambiguity is the tax. An undefined trigger with retroactive potential and interpretive discretion is worth more to a finance ministry than a clean rate, because it can be pointed at whatever flow it wants later.
Watch three things. The Senate finance committee markup — that is where the definition gets written or buried. Euro stablecoin on-chain flow migrating to non-French venues — measurable, and it will move before the law does. And whether DGFiP publishes guidance before the DAC8 pipe opens in 2026. If guidance lands first, they intend to collect. If silence holds, they intend to watch.
Governance isn't a meeting. It's a raid. And this one was executed at 3 a.m. in a budget annex, against an asset the same government spent three years telling the world to trust.