The fintech startups most likely to matter next are not the consumer apps making headlines. They are the infrastructure companies: the ones selling payments plumbing, banking-as-a-service rails, instant account verification, and compliance tooling to banks and platforms that would rather buy than build. That is the pattern the funding data and the recent deal activity both point to.
The numbers behind that read come from the industry's own trackers. According to Plaid, the number of fintech startups in the Americas grew from 5,868 in 2018 to nearly 14,000 in 2024, even as funding swung hard in the other direction. Global fintech funding hit a record $132 billion in 2021, contracted to $75.2 billion in 2022, a 46% drop, and broader startup funding reached $314 billion in 2024, up 3% year over year. Growth in company count with a flat funding line means one thing: the easy money is gone, and only the models with real revenue are building.
What counts as a fintech startup in the first place is looser than most coverage suggests. IBM uses the term interchangeably with "fintech startup," meaning a company whose core capability is delivering fintech products and services, but the ecosystem also includes established technology firms moving into finance and traditional banks adopting the same tools. The label is a self-identifier, not a license category.
Which fintech models are actually gaining traction?
Four models are drawing the most sustained interest, and they share a common trait: they get paid by businesses, not by consumers who churn when a promo rate ends.
- Payments infrastructure. Cashless adoption keeps climbing. Plaid reports that 41% of Americans now say all their payments in a typical week are digital, and direct bank transfers cost merchants significantly less than card acceptance. Startups that make account-to-account payments fast and simple sit on that cost gap.
- Banking infrastructure and BaaS. The companies that let a software platform offer accounts or cards through a sponsor bank. This model took real damage from recent consent orders and the Synapse collapse, which separates the survivors from the tourists. Our coverage of how FDIC and Fed consent orders reshaped BaaS banking after Synapse explains who now bears the compliance burden.
- Embedded finance. Banking and payments delivered inside non-financial software. IBM notes the expansion of embedded finance as one of the industry's predicted next trends, and the mechanism is straightforward: the distribution is the platform's, the balance sheet is a bank's, the startup takes a cut of both.
- Verification and compliance tooling. Instant account verification has already displaced micro-deposits for bank linking, and fraud mitigation is a named growth area. Every new consumer fintech is also a new fraud surface, which makes the tooling layer a durable market.
Consumer-facing models have not disappeared, but the bar moved. Neobanks now have to show a path to interchange revenue and cross-sell rather than user growth alone, which is why our explainer on how Chime and other neobanks turn free checking into interchange revenue matters for judging which of them survive.
What does the funding cycle tell us about who gets backed?
The funding cycle rewards a different founder than it did in 2021. Plaid points to the acquisitions of Brigit, Bridge, MoneyLion, and TradePMR, plus confidential IPO filings from fintechs including Chime, as signs of renewed investor interest. That is exit-driven interest, not a return to growth-stage checks written on user counts.
The practical consequence: seed rounds still happen, but Series B and later capital goes to companies that can show unit economics. A fintech startup selling to banks or platforms can demonstrate revenue per contract. A consumer app selling a budgeting tool has to show retention, and retention in personal financial management is historically weak. Plaid's own Fintech Effect survey found the average consumer uses three to four different apps for payments, taxes, and banking, which cuts both ways: lots of adoption, little loyalty. Readers following this should also see How FDIC and Fed Consent Orders Reshaped BaaS Banking After Synapse.
Our analysis: the next wave of notable companies will look boring from the outside. Middleware, verification, settlement, and treasury tooling do not generate launch-day coverage, but they generate contracts with multi-year lock-in, which is what acquirers pay for.
Why does the trust gap with banks keep narrowing?
Consumer trust is the quiet variable in every model above. Plaid's survey found 87% of consumers are comfortable using national banks while 79% say they are comfortable using fintech companies. A gap of eight points is not the moat legacy institutions once assumed it was. This connects to our earlier piece, How Embedded Finance Puts Banking Inside Software Platforms Through Sponsor Banks.
Reach tells a similar story. IBM cites a Brookings Institution report finding that over 90% of Hispanic consumers use fintech, along with 88% of African Americans and 79% of Asian consumers. Fintech's distribution advantage is concentrated exactly where branch networks have historically been thinnest. There is a counterweight, though: digital financial services can exclude people with limited internet access or low digital and financial literacy, and they raise consumer-protection concerns around fraud, cybersecurity, and data privacy. Access and exposure arrive together.
Is 'fintech' even a fixed category anymore?
No, and the definitional drift is itself informative. The term is older than most people assume: per Wikipedia's financial technology entry, the earliest documented use dates to 1967, in a Boston Globe article titled "Fin-Tech New Source of Seed Money," and the word gained prominence in the early 1990s when Citicorp's John Reed used it for the Financial Services Technology Consortium. The infrastructure milestones predate the label entirely: Fedwire launched in 1918, Barclays deployed the first ATM in 1967, NASDAQ became the first digital stock exchange in 1971, and SWIFT standardized interbank messaging two years later.
Michigan Technological University draws a sharper line: to be FinTech rather than mere financial technology, an offering must go beyond what is standard in its market. EMV chips were once FinTech; now they are status quo. As Dan Green, a principal at Blackfin Group, put it in MTU's explainer: "If an innovation is truly great it will be adopted by everyone and become the way things are done." The winners in any cohort of fintech startups are, in a sense, working to stop being fintech startups. Their technology becomes plumbing, and they move on to the next edge.
Green's other observation deserves equal weight: "FinTech is more about innovation than disruption." Disruption is the most overused word attached to the sector. The companies that last tend to be variations on existing processes made faster and cheaper, built by people who worked inside financial services and know where the process breaks.
What this means for anyone tracking the sector
Judge a fintech startup by three questions, not by its funding announcement. Who pays it, and is that revenue contractual or promotional? What regulated activity does it touch, and through whose charter or license? And what happens to its unit economics when the cost of customer acquisition returns to normal?
The sector's direction is legible in the data already cited: company count up, funding flat, trust gap narrow, exits reopening. The next wave of fintech startups to watch will be the ones whose revenue does not depend on venture capital continuing to believe in the category. That has always been the test. The difference now is that the market is actually applying it.
