European buyout and growth investor Finch Capital has published its inaugural State of AI in Financial Tech report, arguing that financial technology is emerging as a net beneficiary of the AI wave that sent the wider software industry into turmoil since the beginning of the year.
The report draws on public market data, the Anthropic Economic Index, and disclosures from leading FinTechs including Klarna, Adyen and PayPal.
Financial Technology has held up where SaaS has cracked
Since the so-called “SaaSpocalypse” began in January 2026, public market data tells a clear story. The Finch Capital FinTech Index has fallen 19 percent from its peak. A comparable basket of generic SaaS names has fallen 32 percent, a sell-off three times as severe at its worst point.
FinTech has already bounced 11 percent off its low as the market stabilises; SaaS continues to set new lows, down as much as 33 percent from peak. In other words: Financial Technology has been less hard hit by AI as other software and technology sectors.
Why Anthropic’s breakthroughs broke SaaS but not financial tech
The catalyst is no secret. The sell-off was triggered by a step-change in frontier model capability, most visibly the release from Anthropic earlier this year, that made large parts of the generic SaaS workflow stack look automatable. But the same models that are eroding SaaS moats are reinforcing the moats that matter in Financial Tech.
Three structural reasons stand out in the report:
Regulation
Financial Tech sits behind years of licensing, supervisory engagement and compliance investment under regimes such as DORA, PSD3, MiFID II, AMLR, MICA and the EU AI Act.
AI can accelerate the work of compliance – KYC refresh, transaction monitoring, SAR drafting, trade surveillance, but it cannot manufacture the licences, audit history or regulator relationships that make a Financial Tech business defensible in the first place.
Proprietary data and approved networks
FinTech moats are built on transaction data, fraud signals, credit performance and access to scheme- and bank-approved rails. None of this can be vibe coded.
Human judgement
Anthropic’s own Economic Index shows Claude is 94 percent as capable on Business and Financial tasks as it is on Computer and Math tasks. Yet only 3 percent of enterprise AI usage today is in Business and Financial work, compared with 52 percent in coding.
AI is replacing workflows far faster than it is replacing the regulated, judgement-heavy decisions at the core of finance – exactly where Financial Tech operates.
A productivity story
Aman Ghei, partner at Finch Capital, says: “The consensus narrative is that AI breaks software, and that Financial Tech is just a regulated flavour of software. We think the consensus is wrong. The same Anthropic release that triggered the SaaSpocalypse is quietly making financial tech more valuable, not less.
“AI is brilliant at automating workflows, but it cannot conjure a banking licence, replicate a decade of transaction data, or take regulatory responsibility for a credit decision. As generic SaaS moats erode, fintech’s moats – regulation, proprietary data, approved networks, human judgement – are being reinforced and this is particularly true in Europe.”
The report also pushes back on the idea that AI is already showing up in revenue. FinTech and SaaS revenue growth have converged at around 13 percent year-on-year.
Where AI is showing up is on the cost side: median revenue per employee across the FinTech Index jumped 9.8 percent in a single quarter, while EBITDA margin moved just 0.6 percentage points – labour productivity is accelerating, but the savings are migrating into other line items, including a roughly 4x rise in frontier model inference cost in under a year.
Company disclosures point to 30 to 60 percent efficiency gains in support, fraud, onboarding and collections. Klarna alone now reports AI is handling around 66 percent of its customer support queries.
By 2030, Finch Capital estimates AI will cut 15 to 50 percent of operating costs across FinTech verticals, but who keeps those savings will vary sharply by sub-sector. Insurance and RegTech are positioned to retain them in margin; Payments will see them competed away in price.
Ghei says: “AI is already delivering efficiency gains on core financial technology cost drivers and accelerating labour productivity. However, the cost of running frontier models is also ramping up significantly and might be absorbing some of the savings in the short term.
“We are seeing business models transforming for financial technology companies and those with deep moats and high operating leverage will emerge as the strongest beneficiaries of AI.”
Where the category winners sit
The report identifies sub-sectors that combine deep compliance moats with high AI-automation potential – what Finch Capital calls the “regulated AI edge”.
Trade surveillance, KYC, fraud detection and insurance underwriting score highest. Lower-moat areas such as personal finance apps, lead-generation marketplaces and basic accounting tools are most exposed to AI commoditisation.