MiCA's deadline was sold as a doorway to all of Europe. The arithmetic says it is a turnstile.

On 26 June 2026, Binance, the largest crypto exchange in the world, told customers in France, Italy, Spain and Poland that it would suspend most of its services on 1 July. From that date it stops accepting new sign-ups, deposits and most spot trades for EU residents, though existing funds stay accessible and withdrawals keep working. Two days earlier it had withdrawn its license application in Greece, telling the Hellenic Capital Market Commission (HCMCHCMCThe Hellenic Capital Market Commission, Greece's financial regulator, which was reviewing Binance's license application before the firm withdrew it.) it acted 'after careful consideration of the status and the timeline of the process.'

According to reporting by the Financial Times and others, Greek regulators had raised two concerns before that withdrawalwithdrawalWhen a firm pulls its own application before a decision, often under pressure; it counts as a failure but it is the firm leaving, not the regulator refusing.: the strength of Binance's anti-money-laundering controls, and whether its founder and former chief executive Changpeng Zhao, who still owns most of the company, could satisfy MiCAMiCAThe EU's Markets in Crypto-Assets regulation, a single rulebook that replaces the patchwork of national crypto rules across all 27 member states.'s 'fit and properfit and properA regulator's judgment about whether the people running a firm, and its controls, are honest, competent and suitable to be trusted with customers' money.' test. We flag this detail as load-bearing and reported rather than independently confirmed; the HCMC itself declined to comment. Binance says it is not leaving Europe for good and intends to seek a license in France.

The trigger is a deadline. On 1 July 2026 the transitional periodtransitional periodA grace window during which firms could keep operating under old national rules while applying for the new MiCA license; it ends 1 July 2026. under the EU's Markets in Crypto-Assets regulation, known as MiCA, runs out. The European Securities and Markets Authority (ESMAESMAThe European Securities and Markets Authority, the EU body that coordinates national financial regulators and oversees how MiCA is applied.) has set this out plainly: after that date a firm is either authorized under MiCA or it is breaking EU law, with no intermediate status and no extension mechanism in the rules. The financial stakes are real. Operating without authorization can draw administrative fines of up to EUR 5 million or 5 percent of annual turnoverannual turnoverA company's total yearly revenue; MiCA ties some maximum fines to a percentage of it so that penalties scale with a firm's size. for a company, and other categories of breach, such as stablecoin or market-abuse violations, carry even higher caps.

This piece asks one narrow question, and only one: of the crypto firms that already held a national registration in the EU, what share actually cleared MiCA's bar by 1 July, and does that pass rate show a market being actively gatekept or one thinning itself out?

Of an estimated 1,200 firms that previously held national VASP registrations across the bloc, roughly 210 had converted to a full CASP license as the deadline approached, a conversion rateconversion rateHere, the share of firms that already held an old national registration and successfully obtained the new MiCA license. of about 17 percent. Measured against the wider universe of crypto firms operating in Europe, counted at about 2,747 in 2024 and sometimes rounded to 3,000, the licensed survivors are closer to 7 or 8 percent.

A note on honesty before we go further, because these counts are softer than they look. Depending on the weekly snapshot, and on whether you count full authorizations only or also include credit institutions and pending notifications, the running total of authorized CASPs lands somewhere between roughly 200 and 245 in late June. ESMA's narrower register read 204 on 18 June; a widely used tracker built on the same register read 244 on 26 June. We use about 210 as a central figure. The denominators are softer still: the roughly 1,200 legacy base traces to crypto trade press rather than a regulator, and the wider universe of about 2,750 comes from a 2024 count set against 2026 licensees. We keep the two ratios strictly separate so they are never blurred into one flattering or one damning number.

And one more caution that turns out to matter more than any of these: 17 percent is a snapshot of a process still running. Authorizations are being issued week by week, and firms sitting in the pending pile will convert after the deadline. So 17 percent is best read as a floor, not a settled verdict.

The closest precedent for a hard, judgment-based crypto gatekeeper is the UK's Financial Conduct Authority (FCAFCAThe UK's Financial Conduct Authority, the regulator that has run a strict crypto anti-money-laundering registration regime since 2020.) anti-money-laundering registration regime, which has run since January 2020. The FCA received 359 applications from crypto firms and registered just 44 of them. Counting refusals, withdrawals and rejections together, the regulator's own framing is that more than 87 percent of applications failed to meet its standards. That 44-of-359 works out to a 12 percent approval rate, close enough to MiCA's 17 percent that the two look like the same gate.

They are not, and the chart below is built to show why rather than to hide it. Put the ratios side by side and the first thing you see is that they do not share a denominatordenominatorThe bottom number in a fraction; change what you divide by and the same count of licensed firms can read as 7 percent or 17 percent or 40 percent..

The 17 percent divides licenses by the registered base. The 12 percent divides FCA approvals by FCA applicants. Those are different measurements. The genuinely comparable number, MiCA's pass rate among firms that actually applied, is the fourth bar: a derived estimate of roughly 40 percent. We build it by scaling France's application behavior (below) to the bloc, so treat it as an illustration, not a register count; depending on which legacy base you assume, the figure swings from the high 20s to the high 50s in percent. But the direction holds across every plausible version. On a like-for-like basis MiCA does not match the FCA's severity. It looks more permissive.

Here is where a careful reader should push back, and where the reporting forced a revision.

A low pass rate is a fact. 'MiCA rejected them' is a causal claim, and a different thing. A market can shrink to 17 percent two ways. One is a cull: firms apply and the regulator refuses them. The other is attrition: firms look at the cost and the scrutiny and simply walk away or never file at all. Same body count, very different cause. If most of the missing firms never applied, then 'cull' is the wrong word.

The only granular test we have comes from France. Of about 90 registered French providers without a MiCA license, roughly 40 percent had no intention of applying, another 30 percent never responded to the regulator at all, and only about 30 percent had actually applied. Seven in ten never entered the ring.

Read carefully, this disproves the pure-cull story for France. The contraction there is mostly self-selection. But it is one country, and a small base: the 'no intention' group is only about 36 firms, so we lean on it for direction, not precision, and we cannot assume the other 26 states behaved the same way.

Same body count, very different cause. If most of the missing firms never applied, then 'cull' is the wrong word.

If most of the thinning is firms leaving rather than regulators refusing, the next question is where the ones who stayed ended up. The answer is lopsided.

Germany holds roughly one in four of all CASP licenses, more than the next two countries combined. It is tempting to read that as forum shoppingforum shoppingWhen firms pick the country with the friendliest or fastest regulator to get licensed, then use passporting to serve everyone else., firms flocking to a soft touch. The evidence points the other way. Germany's regulator, BaFinBaFinGermany's federal financial regulator, which has authorized the largest single share of MiCA licenses., has a reputation for being strict and slow, and it could lean on an existing national crypto-custodycustodyThe safekeeping of customers' crypto assets by a licensed firm, similar to a bank holding deposits. licensing regime that gave serious applicants a head start. Concentration here looks like capacity and preparation, not leniency.

Where the forum-shopping worry has actually landed is Malta. In July 2025 ESMA published a fast-track peer review of a Maltese license and concluded the authorization should have been more thorough and was granted ahead of other member states. ESMA has separately argued for a single EU-level supervisor for the biggest cross-border crypto firms, precisely to stop firms from picking the easiest regulator and passporting everywhere else. So the concentration in the chart is two different stories: Germany's lead is the slow, strict kind of survival, while the lighter-touch route ESMA worries about runs through smaller fast-track states.

Put the pieces together and the headline number reorganizes itself. Yes, only about one in six already-registered firms converted, a survival rate that superficially rhymes with the FCA's brutal record. But the mechanism is almost the opposite. The FCA built a gate and turned most applicants away. MiCA built a gate that most firms looked at and chose not to approach; among those that did apply, the pass rate appears markedly higher than the FCA's.

That is the difference between a doorway and a turnstile. A doorway is the promise MiCA advertised: clear the bar once, sell across 450 million people. A turnstile is what the arithmetic actually describes: a narrow opening that most of the old crowd never bothered to push through, while the few who did mostly got to the other side. The market did shrink dramatically. It just shrank mostly by self-selectionself-selectionWhen the people who drop out of a process choose to do so themselves, rather than being pushed out; the survivors are shaped by who decided not to try., with Binance's stalled Greek application the loud exception sitting on top of a much quieter exodus.

The honest verdict, then, is conditional. If you are a firm willing to staff up on compliance and file a serious application, MiCA in mid-2026 looks more like a doorway than the FCA ever did. If you are one of the thousands of lightly capitalized registrations that populated Europe's pre-MiCA map, the deadline was a turnstile you were never going to clear, and most of you decided that before the regulator ever had to.