Introduction
Stripe is a payments and financial infrastructure company that lets businesses accept payments, manage billing, and build financial products through a set of developer-facing APIs. Founded in 2010 by brothers Patrick and John Collison, it has grown from a small startup into infrastructure used by businesses ranging from single-founder companies to large public corporations.
Sitting underneath much of that infrastructure is Radar, Stripe’s machine-learning fraud-detection system, which screens transactions in real time. For readers tracking applied AI in finance, Stripe is a useful case study in how a payments company embeds statistical models directly into a process — checkout — that most shoppers never consciously register as an AI decision point.
What the company does
Stripe builds software and APIs that let businesses accept and manage online and in-person payments, issue invoices, run subscription billing, and move money to sellers and platforms. Its product suite includes Payments, Billing, Connect (for platforms and marketplaces), Issuing (for card programs), Terminal (for in-person payments), Tax, and Atlas, a tool for incorporating a company. Radar, its fraud-prevention product, sits underneath most of these flows, scoring transactions using models trained on data drawn from across Stripe’s network of businesses.
Who it serves
Stripe’s customers range from individual developers and early-stage startups — many incorporated through Stripe Atlas — to some of the largest companies in the world. In its 2025 annual letter, Stripe said its infrastructure powers more than 5 million businesses directly or through platforms, and that it is used by roughly 90% of the companies in the Dow Jones Industrial Average and 80% of the Nasdaq 100. That range, from micro-business to blue-chip enterprise, is unusual for a single payments provider.
Company background
Patrick and John Collison, two brothers who grew up in Ireland, founded Stripe in Palo Alto, California, in 2010. The company raised a $2 million seed round in 2011 from investors including Peter Thiel, Elon Musk, Sequoia Capital, SV Angel, and Andreessen Horowitz, and launched publicly in September 2011 after a private beta period. Stripe is dual-headquartered in South San Francisco and Dublin, Ireland, and has remained privately held rather than pursuing a traditional IPO. It has instead used a series of employee tender offers to provide liquidity: one in February 2025 valued the company at $91.5 billion, and a subsequent offer described in its 2025 annual letter valued it at $159 billion. Co-founder John Collison said at the January 2026 World Economic Forum in Davos that Stripe was “still not in any rush” to go public.
Product and AI capabilities
The clearest expression of applied AI at Stripe is Radar, which evaluates each transaction using machine-learning models trained on patterns observed across Stripe’s processing network rather than any single merchant’s history alone. Because Stripe sits inside the payment flow of millions of businesses, its models can draw on a broader pool of fraud signals than an individual retailer’s fraud team typically has access to on its own. Beyond Radar, Stripe has described using machine learning and, more recently, AI agents in billing-anomaly detection and revenue recovery — intelligently retrying failed card payments — and has pointed to emerging “agentic commerce” and stablecoin payment flows as areas of investment in its 2025 annual letter.
Key developments
Stripe was founded in 2010 and opened to the public in September 2011 following an extensive private beta. In 2011 it raised a $2 million seed round from Peter Thiel, Elon Musk, Sequoia Capital, and others. In July 2024, Sequoia Capital offered limited partners a chance to sell shares in Stripe at a reported $70 billion valuation, part of a broader pattern of secondary liquidity rather than a public offering. A February 2025 employee tender offer valued Stripe at $91.5 billion, and a follow-up tender described in Stripe’s 2025 annual letter valued the company at $159 billion. As of early 2026, Stripe remains privately held with no announced IPO date.
Why it matters
Stripe is a useful reference point because its fraud-detection systems operate at a scale most individual businesses could never replicate on their own — pooling signal across a large, varied set of merchants rather than relying on a single storefront’s transaction history. That network effect is one of the more concrete examples of how machine learning changes a routine process, checkout, without most shoppers ever noticing. Stripe’s decision to stay private and fund liquidity through tender offers instead of an IPO is also a useful data point for readers thinking about how large fintech infrastructure companies choose to grow and disclose information over time.
Sector context
Stripe sits within Brel’s AI in finance coverage as an infrastructure layer rather than a pure lending or fraud vendor — closer in function to identity and payments networks such as Plaid than to underwriting specialists. Readers comparing how AI is applied across the sector may find it useful to contrast Stripe’s transaction-level fraud scoring with identity-first approaches or credit-decisioning tools, which apply machine learning earlier in the customer relationship.
Sources and references
This profile draws on Stripe’s own newsroom communications alongside independent reporting and reference sources, listed below.
- Stripe Newsroom — “Stripe publishes 2025 annual letter and announces tender offer” (stripe.com)
- Wikipedia — “Stripe, Inc.” (cross-checked against company statements)
- Bloomberg — reporting on John Collison’s January 2026 Davos comments