The KPMG Pulse of Fintech covering January to June landed on 26 August, measuring to a 30 June cutoff. I read $103.1 billion of global fintech investment over 2,100 transactions in it. The half before it holds $72.2 billion over 2,501 transactions. Value climbed 42.8% and 401 transactions disappeared.

Ten transactions of a billion dollars or more sat inside that figure. The Worldpay transaction alone accounted for $24.3 billion, announced in January 2026, with the TSYS issuer business changing hands at $13.5 billion in the same month. Clearwater Analytics went private in an $8.4 billion buyout, OneStream at $6.4 billion, Brex at $5.15 billion, Saxo Bank at $1.2 billion, and Kpler took a growth investment above a billion.

I add those seven named deals to roughly $60 billion. The other 2,093 deals share the remaining $43 billion or so between them. I would rather work from that split than from the headline.

I now run one calculation before I say anything about a recovery. I take each segment's reported value, divide it by the reported deal count, and read the average. Everything below is my own arithmetic on KPMG's published figures rather than an average KPMG published.

Payments drew $44.2 billion from 168 deals, averaging $263 million per transaction. Americas M&A took $64.7 billion across 202 transactions, averaging $320 million. Digital assets and crypto took $11.1 billion across 467 deals, averaging $23.8 million. AI-focused fintechs in the Americas took $21.4 billion across roughly 800 deals, averaging about $27 million. Wealthtech took $220 million globally across 32 deals, averaging $6.9 million.

I measure a thirty eight times spread between the top and the bottom of that list. Payments is not the same asset class as wealthtech and it never was, but a thirty eight times gap in average cheque tells me the two are not even in the same market for capital any more. Strategic acquirers and financial sponsors are consolidating one of them. Small rounds fund the other, and most buyers ignore it.

One pair of numbers settles the question for me. Payments absorbed $44.2 billion, or 42.9% of the global total, out of 168 deals, or 8% of the count.

Nothing else in the sector comes close to that. All 467 digital-asset deals together drew $11.1 billion, a quarter of the payments figure from nearly three times the deals. Inside the United States, crypto and digital assets drew $5.9 billion over 187 deals.

Payments carries three things the other categories lack: a transaction base, an authorisation, and a seat in somebody else's settlement chain. I read the money as paying for installed rails. I argued elsewhere this year that the largest crypto buyers bought a register or a licence rather than a codebase, and I see the payments figures making that argument at five times the scale.

The Americas took $86.9 billion across 1,120 deals, and the United States alone took $80.8 billion across 933. EMEA took $11.3 billion across 626 deals. Asia-Pacific took $4.6 billion across 350.

I run the same division on those four lines. The United States averages $86.6 million per deal. EMEA averages $18 million and Asia-Pacific averages $13 million. So American buyers write cheques five to seven times the size of everyone else, on a deal count barely larger than EMEA's.

I treat that as the most useful single fact for any owner selling a licensed entity outside the United States. Your buyer pool is deep in deal count and shallow in cheque size, and the capital that would pay a strategic premium mostly sits on the other side of the Atlantic. So I plan for a cross-border process, with an ownership change in front of a supervisor the buyer has never met, rather than a local auction. I use N5Deal's AI-driven matching and its deal search alerts on exactly that shape of problem, where the right buyer exists but sits outside my address book and outside my country.

I had not connected the mechanism underneath all of this until I lined two figures up side by side.

Fintech IPO exits produced $5.3 billion from 10 transactions in the six months to June 2026. The same period of 2025 produced $64.9 billion. I read that as a 92% collapse in the value of the public exit route inside twelve months.

Once the public window shuts, a sponsor needing liquidity sells to another sponsor or to a strategic buyer. Take-privates run the other way, since a listed company trading under what a sponsor will pay turns into a target. I see Clearwater Analytics at $8.4 billion and OneStream at $6.4 billion as that trade running both directions inside one half.

So I would not read the rise in M&A value as a vote of confidence in fintech. Part of it is exit traffic re-routed off the stock market. I keep the two readings apart, because they predict different things next.

Private equity held around $3.7 trillion of dry powder entering 2026 on Preqin's count, with Bain putting the wider private capital figure near $3.9 trillion. Buyout strategies hold about $1.1 trillion of it, roughly 30%. Venture holds $580 billion, growth equity $420 billion, private debt $440 billion, infrastructure $370 billion and real estate $400 billion.

I watch the clock rather than the pile, because the clock is what changes behaviour. Current buyout funds take about 5.5 years to deploy 90% of committed capital, against 4.5 years for 2010 to 2015 vintages. I read a year of extra slippage on a trillion dollars of buyout powder as pressure to write large cheques into assets with predictable cash flows.

Payments infrastructure answers that pressure, and I use entry multiples to see who can play. Large-cap deals above $250 million enter above 11 times EBITDA. Lower-mid-market deals between $25 million and $100 million enter at 6 to 8 times. Most licensed fintech assets I look at sit in the gap between those two brackets. Owners rarely price themselves into the right bracket at the first attempt.

Fintech venture deployed $23.7 billion in the first half of 2026 on PitchBook's count, with $13.3 billion of it in the second quarter across 461 deals. I would call that a strong half by any recent standard.

The valuations moved harder than the volumes did. Median venture-growth valuations reached $1.43 billion in 2026 against $698.8 million in 2025. Early-stage medians reached $30.4 million against $16 million. I make that a doubling in both brackets.

Doubling medians alongside falling counts tells me the market pays far more for what it picks and passes on the rest. For a founder starting something new I would point at the Fintech Builder route rather than the M&A route. Licensing, structuring, operating model and go-to-market get assembled deliberately, because capital for a half-built fintech has turned expensive and scarce at once.

Own payments infrastructure with volume running through it and you sit in the one category where capital chases assets rather than the reverse. I would treat $263 million per deal as the environment you are selling into.

Own a licensed entity outside the United States and your realistic buyer is American or British, running a cross-border process. Expect to be shown the EMEA average of $18 million per deal as your benchmark. I would refuse that benchmark if the asset carries rails or a client book.

Own a wealthtech or a mid-stage platform with no licence and no transaction base, and $220 million across 32 deals worldwide is your entire addressable market. I would plan for a long process, or for a different exit altogether.

I hit an inconsistency on the identity of the Worldpay acquirer between two reports of the same KPMG dataset. One names Global Payments as acquirer of Worldpay with the TSYS issuer business going to FIS, and the other names FIS as acquirer of Worldpay with TSYS as a divestiture. Both agree on $24.3 billion and on January 2026. I flag the disagreement instead of picking a side, because I have not read the transaction documents and the two accounts cannot both hold.

I never got hold of the deal-count denominator behind the Americas AI figure. The $21.4 billion is reported against roughly 800 deals, and my $27 million average depends on that count being the AI subset rather than a wider grouping. If the count is wrong my average is wrong with it.

I found nothing that splits the KPMG payments figure between M&A, private equity and venture. The $44.2 billion mixes acquisition value with investment, and the Worldpay and TSYS transactions alone could account for most of it. That would change the average cheque for the rest of payments dramatically, and I cannot produce the number that would settle it.

My dry powder figures come from reports dated early 2026, and the market has run two more quarters since. I treat $3.7 trillion as a floor rather than a current reading. No mid-year update I trust has turned up.

One thing still nags at me. Wealthtech drew $220 million worldwide from 32 deals in a half where one payments transaction drew $24.3 billion. I have yet to find anyone in this sector writing about what happens to a category the capital simply leaves.

What follows from these numbers depends on the specific asset, and this article does not attempt to advise on one. Read it as market commentary. Nothing here advises on law, regulation, tax or investment. Reported deal values and counts vary between data providers according to what each provider includes, average deal sizes calculated here are arithmetic on published aggregates and not observed transaction prices, and dry powder and multiple figures predate the period they are used to explain. Before anyone prices, markets or buys a fintech asset off market data, an M&A adviser active in that segment should test the data, a corporate lawyer qualified where the entity sits should review the structure, and a consultant who has dealt with the supervisor involved should confirm the regulatory path.