# 100 days of MiCA without an extension: OTC eats the exchange's lunch

> One hundred days after the end of the MiCA extension, 16 million trades show an OTC market growing 57% while exchanges fall 47%.
> By José Maldonado · Regulation & CBDC · 2026-10-20

On July 1, 2026, the MiCA extension ended, and one hundred days later, [Finery Markets](https://finerymarkets.com/) published the first serious tally of what has happened since then, and the result upends the intuition nearly the entire sector works with. The sample is not surveys or estimates: it is more than 16 million institutional spot trades executed on its platform between January 1 and September 30, compared with the same period in 2025.

The headline is a one-hundred-point gap. OTC volume grew 57% year on year while the twenty largest centralized exchanges lost 47% and the twenty largest decentralized exchanges lost 35%, according to the CoinGecko and DefiLlama data the report itself uses as a contrast. In January, anyone looking only at exchanges would have concluded the market was dying.

The report's reading, signed by Finery's leadership and by executives from Banking Circle, Fipto, The Vault and Sumsub, is that MiCA has not built a European retail market. It has built an institutional execution layer, licensed, with identified counterparties and with an entry cost that pushes out small operators. The euro, meanwhile, still accounts for 0.3% of the stablecoin market.

## Who is measuring, and what interests are behind it?

Finery Markets is not a think tank. It is a non-custodial crypto ECN and a trading-as-a-service platform, launched in 2019, that operates through an order book, request for quote and pricing streams. It has accumulated more than $60 billion in traded volume, serves 150 institutional clients in 41 countries and appears in the Deloitte Technology Fast 50 of 2024 and in CNBC and Statista's Top 300 fintechs of 2025. Euromoney has just given it the award for best digital markets connectivity of 2026, decided on October 16, four days after publishing the report.

The full report, with methodology, charts and footnotes, is [published](https://finerymarkets.com/blog/100-days-of-mica-after-the-grandfathering-deadline) by Finery Markets and can be downloaded without registration. That profile explains the bias and also the value. Whoever writes the document sees the real flow of OTC desks, brokers, payment providers, hedge funds and custodians, not the volume of a retail exchange. And it is signed by people with names and surnames.

Sergey Klinkov, managing director at Finery, writes the market chapter. Ida Mokhtassi, head of legal at Banking Circle, takes on the EBA's double test on reserves. Patrick Mollard, CEO and co-founder of Fipto, explains why unlicensed firms have not left. Artem Stopnevich, CEO of The Vault, covers custody. Egor Larichkin, product legal counsel at Sumsub, provides the verification data. All the firms are stakeholders in the outcome, and that is worth keeping in mind when reading their conclusions.

## The exchange loses and OTC wins

The quarterly series avoids the trap of a single good quarter. OTC grew 40% in the first quarter of 2026, 114% in the second and 28% in the third. The second quarter is compared with a very low base from 2025, which is why the jump looks bigger than it is. The third, still 28% above the market, is the figure that supports the thesis of normalization rather than exhaustion.

Indicator

9M 2026 vs. 9M 2025

OTC spot volume

+57%

Spot volume on the 20 largest CEXs

−47%

Spot volume on the 20 largest DEXs

−35%

Stablecoin volume

+76%

EURC volume

×46

ETH volume

+222%

BTC volume

−13%

The report recalls that in 2025 OTC grew 109% against 9% for centralized exchanges, and that in the first nine months of that year the gap was 138% versus 22%. The persistence of that divergence over three years points to a change in how large blocks are executed: discretion, lower market impact and capital efficiency push the institutional investor off the public book.

Klinkov draws the analogy that structures the whole document. MiFID II made compliance more expensive in equities and fixed income in 2018 and pushed out small brokers; MiCA, he argues, will do the same to crypto operators without scale. Those that survive will be infrastructure prepared for bank auditing, and the rest will become clients or partners.

## Stablecoins are already the market

Of every hundred trades that passed through the platform, 83 were in stablecoins. Volume in that segment grew 76% year on year, below the 136% of 2025 but far above the overall figure. Crypto-to-crypto and crypto-to-fiat conversions make up a small part of the flow. In practice, the institutional market no longer buys and sells crypto against money: it moves tokenized dollars and euros and uses the rest as a satellite position.

That figure should be cross-referenced with what is happening in Europe. ESMA gave platforms three months to delist unauthorized stablecoins, and Coinbase set October 30 as the deadline to delist USDT in the Union. The piece on [Coinbase and the October 30 deadline](https://observatorioblockchain.com/en/regulation-cbdc/coinbase-plazo-retirar-usdt-europa-30-octubre/) describes the same movement that appears here as a flow trend.

## The euro exists, but it accounts for 0.3% of the crypto market

The report's most cited chapter mixes a success and a corrective. EURC multiplied its year-on-year volume by 46 in the first nine months of 2026, but it still represents 0.12% of the platform's stablecoin volume. The supply of euro-linked stablecoins reached $848 million on September 7, 22.6% more than the $692 million in January, against $298.7 billion for the dollar and a gain of 0.05% in the same period, according to Token Terminal. The euro's share is around 0.28%.

Mollard's conclusion, from Fipto, is the most useful in the document and the least epic. There were three changes in one hundred days: the regulatory perimeter shifted toward the licensed partner, demand came from payers in Canada, Africa and Asia-Pacific who need a euro leg, and issuance finally found a usable path. Availability first, volume later. Euro-denominated trading, which includes dollar stablecoins traded against the euro, grew 24% year on year, half the rate of total OTC: euro liquidity is still routed through the dollar.

## Ethereum eats Bitcoin's lunch within the majors

Institutional flow concentrated more, not less. Bitcoin and Ethereum together went from 70% to 79% of non-stablecoin volume, and the internal composition flipped. Ethereum rose from 23% to 50% of the flow, with volume 222% higher, while Bitcoin fell from 48% to 29% and its volume retreated 13%. The rest of the assets fell from 29% to 21%.

The report itself calls for caution in reading this. A shift within the two largest assets is not a broadening of appetite for the rest of the spectrum, and 222% on a small institutional base does not justify talk of structural rotation. It remains the most striking and least interpretable figure in the 25 pages.

## The real cost is in the EBA's double test

Where the report becomes operational is on reserves. The EBA's technical regulation imposes two simultaneous tests on credit institutions that custody reserves for a stablecoin issuer. The first limits the deposit to 1.5% of the bank's total assets. The second limits the issuer's share according to the size of the entity: 25% for systemic banks, 15% for large ones and 5% for small ones. The more restrictive applies.

Mokhtassi gives an example that explains why issuers complain. With €1 billion in reserves and small banks, the 5% cap leaves €50 million per entity and forces work with at least 20 banks. If those banks have €2 billion in assets, the 1.5% test lowers the limit to €30 million and the figure rises to 34 entities. With large banks the number falls to seven and with systemic ones to four. Entering the European market becomes a problem of managing banking relationships, not of technology.

Banking Circle's proposal is that the double test not apply when reserves are held in a structure separate from the balance sheet and with available liquidity. The 2008 argument sums it up without embellishment: the size of the bank does not guarantee its safety. It is a self-interested position, because Banking Circle is a correspondent bank that wants those reserves, and it is still the most concrete criticism in the document.

## The flip side: fewer users and more fraud

Sumsub provides the figure no regulator has put in a press release. Crypto verifications of European users fell almost 13% in the quarter after July 1 compared with the previous one. Germany and Belgium contracted sharply; Lithuania, Malta and Finland gained activity. The reading is that users follow licensed platforms and the jurisdictions where those platforms have established their European base.

The second finding goes in the opposite direction and is the one that matters for risk. The proportion of European applicants rejected for fraud attempts rose almost 50%, more than double the increase recorded outside the Union. Larichkin explains it by the calendar: a licensing deadline triggers rebuilds of client onboarding, migrations to new legal entities and mass re-verifications that scammers exploit with fake re-verification requests and support impersonation.

## The next test is euro volume

MiCA has solved availability and left demand pending. In one hundred days the European market learned to issue tokenized euros, to sign licensing umbrellas and to live with the delisting of unauthorized stablecoins, as the standoff over [Binance and its European authorization](https://observatorioblockchain.com/en/regulation-cbdc/binance-europa-sin-autorizacion-mica/)shows. What it has not achieved is for that euro to be traded.

The report itself leaves two questions unanswered. The first is whether OTC flow will keep growing once centralized exchanges stop falling, because so far the gap has been measured against a contracting market. The second is whether the EBA's double test will be revised before the large issuers decide not to enter the European market. One hundred days are enough for a snapshot, not for a verdict.
