Meta has stopped accepting cards from another issuer. This week, some advertisers paying for @Meta Ads with cards from neobank @slashapp lost access to their ad accounts: Meta stopped accepting part of Slash's credit BINs. Slash confirmed the incident in a message to clients and says 95% of affected users are already back up, with a working fix being a debit BIN or paying invoices via ACH. Meta hasn't given a reason. But the scale is notable: according to Slash's CEO, almost 2% of all Meta ads are purchased with Slash cards, and about $4B a year runs through the company's credit cards. The more interesting part: Slash itself says this isn't a one-off but a consequence of changes Meta is rolling out gradually, and expects other providers to be affected too (no issues at other providers spotted so far). That lines up with what we saw in April, when Meta moved large advertisers off cards and onto invoicing and direct debit. The takeaway for fintech is simple: if your card product depends on a single merchant, the risk doesn't live with you - it lives in their policy. And large B2B billing is leaving card rails for direct ones - ACH, debit, and next, stablecoins.

How many stablecoin cross-border B2B companies are actually funded? I reviewed 589 companies from the @VerdaVentures Stablescape list and checked each one individually for confirmed funding rounds across all available sources. I also removed projects that had been misclassified into the category, which left 568 real ones. Result: 193 of 568 ( 34% ) have a confirmed funding round. 375 ( 66% ) show no public evidence of outside capital. Of the 193 funded: - 159 disclosed the amount - $4b in total, median $7m per company - 34 raised money but did not disclose the size What else the check revealed: 1/ Capital is concentrated at the top. 64 companies at Series A or later account for 76% of the $4b disclosed. The long tail of ~ 375 companies is either bootstrapped, pre-launch, or already dead. 2/ Market maps overstate the sector. "589 companies building stablecoin rails for cross-border payments" sounds like a crowded market. The real investable universe is under 200 companies, and the one with meaningful traction is smaller still. 3/ Africa is large by headcount, thin by capital. 90 companies, 51 of them in Nigeria - the third-largest region by count after North America and Europe, but with the lowest funded share among major regions ( 30% ) and the smallest median round ( $3m vs $9.5m in North America ). What this means for investors The category has the structure of an early-stage sector. Of 568 companies, 194 were founded in 2025-2026, and 2025 was a record year for rounds - 79 vs 40 in 2024. At the same time, capital is already concentrated: 64 companies at Series A or later hold 76% of disclosed funding. The market is young in composition and mature at the top - the classic picture before a wave of seed deals, when the infrastructure leaders are set and the next layer has not yet been selected. 375 companies without public funding are not 375 opportunities. Most of them will never get past a landing page. But this tail contains the 2024-2026 cohort that has not yet raised institutional money: 217 teams. For a pre-seed/seed fund, that is the pipeline - and it should be screened on traction, not on the existence of a website. The practical takeaway: deal competition at Series A in this category is already high - 60+ companies and large checks. At pre-seed/seed it is low, because two-thirds of the market has no investor at all. If the fund's thesis is stablecoin rails for cross-border settlement, the entry window right now is not at the top but in the tail, and the filter should be onchain volume, not a position on a market map. Caveat: public sources only, so quiet rounds are not visible - 34% should be read as a lower bound.

Yesterday: @mural_pay in my “underrated” tier. Today: $15M Series A led by @tether. My choice was the right one. The thesis was simple - $2.1B processed in 128 countries on ~$13M of funding, with @opera and @deel as live customers, not logos. Tether leading isn’t about returns, it’s about owning the rails where businesses settle in USDT. Next test: does this show up as on-chain USDT settlement, or stay in fiat corridors?

In December 2025 @ARKInvest put my Crypto Cards dashboard into its DAT Market Snapshot. @ether_fi was the lead name on that slide, with $56M of card spend that month. Ten months later, pulling EtherFi Cash on its own: $1B in cumulative card spend, crossed on 17 October. $815M of it came after that slide. The first $500M took 18 months from launch. The second $500M took four and a half. Monthly volume went from $56M in December to $129M in September, with October running at ~$140M. 12.7M transactions, 117.6K cards since launch.

Eight weeks after the chart: Circle now trails both networks. Since 19 Aug: @Visa +10.1%, @Mastercard +7.6%, @circle +4.6%. S&P flat. Visa at a 1Y high. This week shows how Circle trades. Oct 15: +9% with no news. Oct 16: Circle announces USDC payments into 40 emerging markets - stock closes -5.5%. The stock trades on rates and regulation, not on operating news.

Polygon is already the main rail here, not the future one. USDC deposits into @tazapay wallets, Jun-Oct 2026: $191M → $364M → $474M → $771M, and $430M in the first two weeks of October. @0xPolygon carried 61% of September’s flow, Ethereum the rest. https://t.co/ndhfUBlPgK

Stablecoin payments will have an advantage where the existing payment infrastructure is very slow, expensive, or both. Where services like Pix or SEPA Instant are available, users simply have no reason to switch. That’s why onboarding a non-crypto audience isn’t just a matter of packaging. You need to hide the blockchain under the hood, but that’s not enough: a product that looks like regular fintech competes with Revolut or Wise. The winners will be those who hide the complexity but demonstrate a measurable advantage - cheaper, faster, or available where there are no alternatives.

One more cut of the @Revolut data: people, not dollars. Weekly active counterparties tell a different story than volume. On the announcement week, $USDT addresses jumped from 26k to 90k - that is the only place the delisting shows up as a spike. The money they moved was ordinary; the number of people was not. Since then $USDC counterparties have tripled, from 12k to 37k. USDT is down to ~200 a week. Clients did not leave. They were moved, and they stayed.

What Circle is actually buying: Tazapay on-chain On 8 September 2026 Circle announced it would acquire Tazapay, a Singapore B2B cross-border payments company with a claimed $25B in annualised volume and payouts in over 100 markets. @circle_ventures had led its Series B extension six months earlier. The press release talks about rails, licences and last-mile reach. It does not say how much of Tazapay's business already runs on USDC. The blockchain does. the route The product is simple: a client funds its account with USDC, @tazapay pays the client's suppliers in local fiat. On-chain this is a fixed route. Roughly 60% of deposits arrive through a settlement contract on @0xPolygon and @ethereum; the rest comes as direct transfers to a treasury wallet from a handful of large counterparties. From there funds move through an intermediate wallet to two addresses @arkham labels as Circle deposits, are redeemed for dollars and leave the chain. Virtually every stablecoin that enters Tazapay exits through Circle. The contract is the newer piece: testing from April, production from July. $54M in July, $472M in August, $572M in the first 25 days of September, average ticket around $300K. Each call is three transfers: the client pays in, the contract forwards the round invoice amount to treasury, 2 basis points go to a fee address. These are not merchant checkouts but treasury-sized deposits from a small number of corporate clients. the number that matters The cleaner metric is what reaches @circle. Counting only transfers from the four wallets I can tie to Tazapay, monthly redemptions run $190M in June, $356M in July, $449M in August and $562M in the first 25 days of September. Polygon has been flat since July at $340 - 370M; all of September's growth is Ethereum, almost triple in a month, which means at least one large new client. Annualised, that is roughly $8B of @USDC. Set against Tazapay's own $25B, allowing that the figures measure different things, about a third of its payout business is already funded in stablecoins. In June it was under 10%. This changes the reading of the deal. Circle is not buying a channel that might one day route volume through USDC. It is buying a customer that already burns over half a billion dollars of USDC a month through Circle Mint, and the fiat rails on the far side. That is vertical integration of an existing flow, not an option on a future one. what is not there There is no mint: over four months no USDC was issued to any of the four wallets. The product is one-directional, stablecoin in, fiat out. @tazapay is an off-ramp, and Circle is buying off-ramp capacity where it has none. There are no stablecoin payouts: apart from redemptions, the treasury cluster sent out one transfer of $9.9M and two smaller ones. The "stablecoin payroll" on Tazapay's site describes the funding leg, not the disbursement leg. There is almost no on-chain revenue: 2 basis points, about $150K on $750M of Polygon volume in August and September. The margin lives in the FX spread at the off-ramp, where no explorer can see it. The chain is a cost centre; the currency conversion is the business. what changed in september The route was stable from June. On 3 September, five days before the announcement, a wallet that had previously only funded the treasury began sending directly to @circle. From 8 September $158M went through it in eight days, and on 20 September the old @0xPolygon route went quiet. In three weeks around the announcement Tazapay rewired its on-chain operations, most likely in preparation for closing: new custody or new Circle Mint accounts. The timing fits the deal; the mechanism is not visible, and I would not go further than that. A footnote: dozens of addresses mimicking Tazapay's counterparties have been sending zero-value transfers to its wallets for months. Address poisoning at $7M per transaction is a rational attack, and a reminder that behind "instant settlement" there is a person checking forty hex characters. the caveats The Polygon Circle address is effectively dedicated to Tazapay. The Ethereum one is not: over a third of its September inflows cannot be attributed, so Ethereum figures are a floor. Clients who fund Tazapay in fiat are invisible on-chain, so the stablecoin share is an estimate against a self-reported denominator. And four months of data is enough for a trend, not enough to annualise with confidence. the turn The stablecoin sandwich, fiat to token to fiat, is usually described as a bridge that disappears once both ends are on-chain. Tazapay's numbers say the opposite: all of the value is in the bread. Nobody on the receiving end wants the token, the token earns the intermediary two basis points, and the entire product is the fiat leg that stablecoins were supposed to make obsolete. Circle understood this. Buying @tazapay is an admission that $USDC growth in cross-border commerce runs on off-ramps, and that they are better owned than rented.

Spent the last week on a question that looked like a paradox: why does a country with the world's best payment rail hold two-thirds of its crypto volume in USDT? Turns out the best evidence isn't adoption data. It's the central bank. Res. 521 reclassified stablecoins as FX. IOF tax on the table. Res. 561 from Oct 1 cuts them out of eFX settlement. You don't build three rules around an experiment. My piece for @oobit 👇