PRIM3 Brief #25: September's Crypto VC Rebound Was a Fintech Round

On October 1, the September tallies landed. Crypto venture funding came to $1.269 billion across 61 disclosed deals, up 71.1% from August's $742 million (TokenPost, WuBlockchain). Most of the commentary called it a rebound.

We'd call it something else. Look at who raised and who wrote the checks, and the month reads less like crypto venture recovering than like fintech venture discovering stablecoins.

Stablecoin payments funding was the single largest theme of the month. Four companies that move dollars over stablecoin rails (Félix, Jeeves, HIFI and Latitude) raised about $382 million between them, roughly 30% of the total. Their lead investors were a16z, CoinFund, Left Lane Capital and Oak HC/FT, and the cap tables behind them include GIC, Vista, AllianceBernstein and General Catalyst's credit fund. None of the four has a token. All four were pitched on volume.

Key takeaways

• September 2026's crypto VC total of $1.269B was 79.2% below September 2025. The nine largest rounds took about $928M, or 73%, which leaves the other 52 deals averaging roughly $6.6M.
• The stablecoin payments rounds were priced like fintech rounds: processed volume, annualized stablecoin flow, enterprise customers and corridor coverage. Token design didn't come up.
• For crypto-native founders building payments, the comparable set has changed. You're now benchmarked against Jeeves and Félix, not against a DeFi protocol's TVL.

Who Actually Raised in September 2026

September's money concentrated in three groups, and only one of them is what most founders picture when they say "crypto VC."

The first was prediction markets. Polymarket's $300 million from 1789 Capital, part of a raise reported at up to $1 billion at a $21 billion valuation (CoinDesk), was 24% of the month on its own. We've written about that category twice already; it's a market of one or two winners and doesn't say much about the broader tape.

Market-infrastructure incumbents came next. Nasdaq put $100 million into Kraken's parent Payward, also at a $21 billion valuation, to push tokenized Nasdaq-listed equities onto rails that don't close (CoinDesk). Kaiko's $57 million round was led by S&P Global. Those are strategic checks — incumbents buying a seat at their own future — not venture bets.

The third group is the interesting one:

• Félix, the WhatsApp-based remittance app for Latin America, raised $200 million: $87 million of equity led by a16z and a $113 million credit line from General Catalyst's Customer Value Fund. It says it has processed more than $8 billion for over 6 million users across 11 countries (Crypto Briefing).
• Jeeves, the expense-management fintech, closed a $110 million Series C-1 led by CoinFund with Coinbase Ventures, a16z, GIC, Vista and AllianceBernstein. Of its $5 billion platform volume, $1.5 billion is annualized stablecoin flow, and that's the fastest-growing segment (FinTech Futures).
• HIFI raised a $37 million Series A led by Left Lane Capital. It runs stablecoin payment and tokenized-asset APIs at more than $7 billion annualized volume across 87 countries, and it took part in DTCC's July tokenized-securities production trades (The Block).
• Latitude, a 15-person team built by Stripe, Uber and Coinbase alumni, raised a $35 million Series A led by Oak HC/FT to turn stablecoins into local-currency payouts through a single API (Crypto Briefing).

Put those three groups together and they account for roughly two-thirds of the month. The rest of the market, the seed and Series A protocol teams most of our founders sit next to, shared what was left. That's the same barbell we described in Brief #12, only now the heavy end is fintech rather than crypto-native.

Why "Crypto VC Is Back" Is the Wrong Read

The 71% month-over-month jump says more about August than about September. August was a $742 million month, the weakest of the year, and three or four large checks are enough to move a base that small. Year over year, September is still down 79.2% from $6.096 billion, and deal count is down 35.1%.

What's more instructive is the composition. The stablecoin payments companies that raised aren't crypto companies in the way the industry usually means it. They're fintechs whose cost of moving a dollar dropped because the dollar now moves on a stablecoin. Jeeves sells corporate cards and payouts to BMW, Lululemon and Nubank. Félix competes with Western Union and Remitly. Their customers don't know or care what chain settles the transfer.

So the investors who led those rounds underwrote them the way they'd underwrite any payments business. How much volume, at what take rate, with what loss rate, through which licences, in which corridors. Jeeves CEO Dileep Thazhmon said revenue grew "more than 50% in a few months" and credited stablecoins and AI for the inflection. That's a fintech growth story, told to fintech investors.

And even the headline needs the usual caveat. More than half of Félix's $200 million is credit, not equity. We made the broader point in Brief #19: funding totals now mix debt, strategic investments and equity into one number, and founders benchmarking against that number are benchmarking against something that doesn't describe their own raise.

What the Payments Data Actually Says

Stablecoin payments are real, but they're a fraction of the volume number people quote. McKinsey and Artemis put reported stablecoin transaction volume at about $35 trillion a year. Of that, they estimate only $390 billion in 2025 was genuine payments, and $226 billion of it was business-to-business. Consumer-to-consumer transfers came to $77 billion (McKinsey).

Those two figures map almost exactly onto September's cap tables. Jeeves, HIFI and Latitude sell to businesses, which is where 58% of real stablecoin payment volume already sits. Félix is the consumer remittance bet, and it raised the most because remittance is the one consumer corridor where the cost saving is large enough for users to switch.

The other thing the data says is that distribution came from outside crypto. Félix lives inside WhatsApp. HIFI's September partnership extends stablecoin settlement to Visa's card network. Jeeves rides on corporate card programmes that finance teams already use. Not one of these companies — not even the consumer one — asked a customer to download a wallet first.

Total stablecoin supply sits around $307 billion as of early October 2026 (DefiLlama), and we covered the issuer concentration underneath it in Brief #10. The supply side is consolidating into a few issuers. The value in the September rounds sat further downstream, at the edge where a stablecoin turns into a payroll run, a supplier invoice or pesos in a bank account in Oaxaca.

What This Means for Founders Building Payments

If you're building anything that moves stablecoins for someone else, your next raise will be priced against Jeeves and Félix, whether or not you think of them as peers.

That changes the diligence in specific ways. These are the questions we'd expect a payments-literate lead to ask after September, and the ones we now ask ourselves:

1. What's your volume, and how much of it is yours? Report processed volume, annualized run rate and the share that's genuinely customer-initiated. Internal treasury movements and wash flows get stripped out in the first meeting.
2. What's your take rate, and where does it come from? FX spread, a per-transaction fee, float or card interchange. Each one has a different ceiling, and the GENIUS Act already rules out paying holders interest on payment stablecoins, which closes off one of them.
3. Who owns the last mile? The stablecoin leg is cheap — the off-ramp into local currency is where licences, banking partners and failure rates live. Latitude raised a Series A on that problem alone.
4. Which licences and corridors do you hold today? Not which ones you plan to apply for. HIFI pitched 87 countries. A pre-seed team doesn't need that, but it does need one corridor that works end to end.
5. Why does this need a token? If the honest answer is "to fund the company," expect a payments investor to pass. A token can be justified by a real network function: decentralized settlement among operators, staking for liquidity providers, or incentives that a cap table can't align. It can't be justified as the revenue model.

None of this means crypto-native teams can't win payments. It means the bar now looks like a fintech bar, and founders who show up with TVL charts and a token unlock schedule will be talking to the wrong side of the table.

A Portfolio Parallel: Oobit

Oobit (a PRIM3 portfolio company and Tether-backed payments app) has been building toward this exact comparison for a while. The consumer app lets people spend stablecoins at more than 150 million Visa-accepting merchants, and in May 2026 it added Arbitrum alongside Solana to cut settlement costs to pennies (Invezz).

But the more relevant move came the same month. Oobit Business launched on May 17 with corporate cards, vendor payments, payroll and contractor settlement, and transfers between stablecoin treasuries and bank accounts, with 24-hour onboarding (Stablecoin Insider). That's the same product surface Jeeves just raised $110 million to expand.

We point to it because the competitive set is now explicit. A crypto-native payments company with a strong consumer base and a crypto-native investor list is being compared, product line for product line, with a fintech that brought stablecoins in as plumbing. The metrics that close that comparison are B2B volume, corridor coverage and unit economics. Teams in our portfolio that can report those cleanly will raise in this market. Teams that can't will find the September investors don't take the meeting.

Kima Network, another portfolio company working on cross-chain settlement without bridges, sits one layer lower in the same stack. Settlement infrastructure becomes more valuable as the payment apps above it multiply, though only if it plugs into the fiat edge those apps need.

Where PRIM3 Is Placing Bets

PRIM3 has worked with 80+ portfolio and advisory companies and supported 40+ token launches. The pattern in our own pipeline matches September's tape: the payments teams that get traction look like fintechs with a crypto cost advantage, not crypto protocols with a payments feature.

So when we look at stablecoin payments now, three things carry most of the weight. The first is a working corridor with real volume, however small, because that proves the off-ramp works. Next comes a distribution channel the customer already uses (a messaging app, a card network, an ERP, a payroll system), which matters more than the chain. The last is a take rate that survives once stablecoin interest is off the table and FX spreads compress.

We're less interested in the next general-purpose payment chain or wallet with a points programme. The September data suggests the money isn't either.

Frankly, that's a healthier market than the one we had in 2021. Payments companies are being priced on whether people use them.

That's where PRIM3 is placing capital this quarter. More on our thesis at prim3.vc.

FAQ

How much did crypto startups raise in September 2026? Crypto VC funding reached about $1.269 billion across 61 disclosed deals in September 2026, up 71.1% from August's $742 million but 79.2% below September 2025, according to WuBlockchain and TokenPost. Four stablecoin payments companies (Félix, Jeeves, HIFI and Latitude) raised about $382 million of that total, and Polymarket's $300 million round accounted for another 24%.

What do investors look for in a stablecoin payments startup? The September 2026 rounds were priced on fintech metrics: processed or annualized volume, take rate, customer quality, corridor and licence coverage, and the cost of the last-mile conversion into local currency. Jeeves reported $1.5 billion of annualized stablecoin volume, HIFI more than $7 billion annualized and Félix more than $8 billion processed. Token design was not part of any of these raises.

Do stablecoin payment startups need a token? None of the largest stablecoin payments rounds in September 2026 involved a native token. A token can still make sense for a payments network that needs decentralized settlement or incentive alignment among operators, but PRIM3's view is that it should be justified by a specific network function, not used as the revenue model or the fundraising instrument.