Fintech focused venture capital has varied in focus over the years, but payments infrastructure is clearly in focus now.
getty
The half-year funding data is out, and it describes a venture market that has made up its mind. Global fintech funding reached $28.6 billion in the first half of 2026, up 23% from a year earlier, while deal count fell 26% to 1,605. Dollars up, deals down: the checks got bigger and fewer, which is what conviction looks like on a spreadsheet. Look at where the big checks went and the conviction has one theme. Ant International raised $1.2 billion for cross-border payments. Ramp raised $750 million at $44 billion. Kalshi took $1 billion at $22 billion and was in $40 billion talks within weeks. Rain, which builds stablecoin card infrastructure, raised $250 million. Augustus raised $180 million to build a clearing bank for stablecoins. Mercury took $200 million at $5.2 billion.
Consumer fintech, lending startups, insurtech: scraps. The 2026 venture cycle has concentrated into payments and market plumbing, most of it stablecoin-adjacent, and the last cycle that concentrated this hard into one layer of the stack ended badly enough that the comparison deserves to be made carefully rather than avoided.
The geography and framing of the half confirm the concentration. The US took $15 billion, more than half the global total, and Crunchbase’s own headline told the story straight: investors concentrating their bets on AI and financial infrastructure. Even Ramp’s round was marketed as an AI story as much as a fintech one, with TechCrunch noting investors hungry for fintechs that carry one. The rounds that used to go to consumer apps now go to whatever sits closest to money movement itself.
Down-stack, the checks keep landing on the same shelf. Augustus, a federally chartered clearing bank for stablecoins settling across Swift, ACH, SEPA and on-chain rails, raised its $180 million with QED aboard; FXC Intelligence’s funding tracker adds KAST’s $80 million Series A and ARQ’s $70 million in a single spring survey, noting the cap tables now feature Goldman Sachs, Citi and Visa. The full-year 2025 tally, counting Circle’s IPO and Figure’s, ran past $2.4 billion into the sector. The Block’s researchers noted the deal count had passed the 2021 peak while the prior cycle was still being written down.
Circle’s chart is the compressed version of the whole argument. The sector flagship priced at $31 in June 2025 and closed at $263.45 eighteen days later, an eight-fold public mania that outdid anything the private market managed. It has since given back 76% of the peak. Nothing about Circle’s business broke; expectations did, and the correction transferred billions from late enthusiasm to early liquidity. Every private mark being set this year at Ramp, Rain or Kalshi will eventually meet its own version of that chart, and the meeting is scheduled for whenever the exits open.
The thesis is regulatory, and it is already priced
The concentration has a clean logic. The GENIUS Act made the rails legal, so capital is racing to own picks and shovels before the banks finish building their own. The numbers trace the stampede: stablecoin startups raised $537 million in 2025 through September, more than five times all of 2024, and The Block found payments and stablecoin issuers taking 7.5% of all venture deals as early as the first quarter of 2025. Andreessen Horowitz raised $2.2 billion for its fifth crypto fund in May and named stablecoins first among its targets.
The strategics validate and exit the same assets on a loop. Visa invested in BVNK in May 2025; Citi Ventures followed in October; Mastercard then agreed to buy the company for up to $1.8 billion in March. Stripe closed Bridge at $1.1 billion. Coinbase completed Deribit at $2.9 billion. One growth investor, ICONIQ, led both Ramp and Rain this half, anchoring the corporate-spend layer and the stablecoin-card layer of the same thesis. When venture, corporate venture and acquirers are all long the same shelf, the shelf is priced for perfection.
To be clear about what the bulls have right: the underlying volumes are real. Mastercard’s own BVNK announcement cited stablecoin payment use cases reaching at least $350 billion in 2025 volume, Simon Taylor’s State of Fintech declared stablecoins had found product-market fit as the cross-border rail, and the M&A tape keeps clearing at premium prices, MoonPay buying Sodot for institutional plumbing, Coinbase digesting Deribit. This is a real platform shift. So was buy-now-pay-later, which is precisely why the funding pattern, and never the underlying trend, is what deserves the scrutiny.
The 2021 rhyme, stated precisely
The comparison is not that 2026 spends like 2021; it spends at a fifth of the pace. Fintech’s 2021 blowoff hit $131.5 billion across nearly 5,000 deals. The rhyme lies elsewhere: 2021 concentrated capital into one consumer layer, checkout credit and neobanking, on the theory that distribution was destiny. Klarna marks the whole arc: $45.6 billion in June 2021, $6.7 billion thirteen months later, an IPO at about $15 billion in 2025, and a market cap around $7 billion today, below half its listing price and back near the down-round mark. The layer got funded to saturation, the winners survived smaller, and the LPs paid the tuition.
The 2026 version funds issuance and orchestration: Bridge, BVNK, M0, Brale, Agora, Rain, Zerohash, Paxos, Anchorage, Fireblocks and a lengthening tail. One industry directory now counts 38 accredited EU token issuers and a big six of stablecoin card enablers, and notes the regulatory-arbitrage era is over. There are, comfortably, more stablecoin infrastructure companies than there are stablecoins with meaningful distribution. Every one of them pitches the same three customers: the banks, the networks and the platforms, which are, awkwardly, all building or buying their own.
The public market is grading the sector on a curve nobody likes
The listed comps carry the warning label. Circle, the sector’s flagship listing, priced at $31 in June 2025, peaked at a closing $263 within weeks, and trades near $63 today, down 76% from that peak even while remaining a double from the IPO. Klarna sits below its offer price. The pattern from 2021 repeats in miniature: private marks compound annually while public comps compress quarterly, and the gap between the two is exactly where the last cycle’s pain concentrated. Ramp at $44 billion is priced at nearly three times Circle’s entire market value, on the strength of software multiples applied to payments flow. Perhaps deservedly. The public market has not yet been asked, and the one adjacent listing it was asked about, Circle, drew a Morgan Stanley downgrade on USDC contraction within a year of its euphoric debut. Private investors are underwriting stablecoin volumes doubling annually and margins holding simultaneously. History’s suggestion is to pick one.
How infrastructure cycles end, and who compounds through them
Overfunded layers resolve the same way every time: consolidation at cost, quiet shutdowns, and a handful of distribution owners absorbing the value the layer created. BVNK’s sale to Mastercard is the good outcome and the template, infrastructure exiting to the strategic that owns the customer. The bad outcomes will not announce themselves; they will be bridge rounds that do not close in 2027. The compounding outcomes belong to whoever owns the relationship the plumbing serves: Ramp owns the CFO’s login, Kalshi owns the order flow, Ant owns the merchants. When a layer is overfunded, value migrates to whoever touches the customer, because everyone else is bidding against nine clones for the same integration.
The honest summary of the half: $28.6 billion of funding, concentrated into fewer hands than any half-year in recent memory, chasing a genuine platform shift whose rails are legal, growing and, at the infrastructure layer, drastically oversupplied. Dollars up and deals down is what conviction looks like. It is also what crowding looks like. The two are indistinguishable until exit, and the first exits, Circle 76% off its peak, Klarna halved from listing, have already started grading the class. The cycle picked its winner. The winner’s winners are still to be picked, and history’s suggestion is that most of them will turn out to be the customers. Watch three tells through 2027: whether Mastercard closes BVNK on schedule and at price, whether a second stablecoin-infrastructure name reaches a public listing at all, and whether the bridge-round graveyard starts filling quietly. The 2021 cohort’s lesson for operators, meanwhile, was survivable: the companies that lived cut burn early and attached themselves to real distribution before the music stopped. The music is loud right now. Consider that a timestamp.
