A Startup Is Rewriting the Rules of Payment Fraud Detection
Stripe Radar has long been the default choice for companies that need fraud detection baked into their payment stack. But Sardine, a fraud and compliance platform built by veterans of Coinbase and PayPal, is winning deals that Radar used to consider locked in – and the reasons why are becoming harder for Stripe to ignore.
Why Sardine Is Gaining Ground
Sardine was founded with a specific thesis: that device intelligence, behavior biometrics, and compliance automation should live in the same platform, not be stitched together from three different vendors. Most fraud tools built inside payment processors treat fraud as a byproduct problem – something to manage at the transaction layer. Sardine treats it as a primary product problem, building detection logic that runs before a payment is even initiated.
That distinction matters more than it sounds. A growing number of fintech companies and neobanks are discovering that card fraud is only part of their exposure. Account takeover, synthetic identity fraud, and money mule networks operate at the account-creation and onboarding stage, long before Stripe Radar ever sees a transaction. Sardine’s architecture captures behavioral signals – how a user moves a mouse, how fast they type, whether a device has been spoofed – and uses those signals to score risk at onboarding, not just checkout.
Stripe Radar is a strong product within its lane. It excels at card-not-present fraud detection for merchants already using Stripe’s payment infrastructure, and its machine learning models benefit from the scale of Stripe’s transaction network. But that strength is also a structural ceiling. Radar’s signals are largely transaction-based and Stripe-native, which means companies using multiple payment processors, or companies operating outside standard e-commerce, don’t get full coverage. Sardine is processor-agnostic, which is a straightforward advantage for any company not fully committed to the Stripe ecosystem.
The competitive pressure is particularly visible in crypto, fintech lending, and buy-now-pay-later verticals. These sectors deal with fraud patterns that don’t map cleanly to traditional card fraud models. Sardine’s founding team built their understanding of these attack vectors at Coinbase and Revolut, and that institutional knowledge shows up in how their models are tuned. For a lending platform trying to identify synthetic identities before disbursement, that context is the product.
Where Sardine Is Actually Winning
The clearest indicator of Sardine’s momentum is the type of customer it is attracting. Rather than competing for small merchants – Stripe’s bread and butter – Sardine is targeting mid-market and enterprise fintech companies that have outgrown entry-level fraud tooling. These are companies with dedicated risk teams, specific compliance requirements, and fraud loss numbers large enough to justify a purpose-built vendor relationship over a plug-in solution.
Sardine’s compliance features are a significant part of that pitch. The platform offers built-in Bank Secrecy Act compliance workflows, sanctions screening, and transaction monitoring – capabilities that most fraud tools don’t include, and that fintech companies are legally required to have. Bundling fraud detection with compliance automation reduces vendor count, which is a real operational win for risk teams managing multiple tools, contracts, and data integrations.
The behavior biometrics layer is where Sardine pulls furthest from the pack. By analyzing device telemetry and behavioral patterns in real time, Sardine can flag high-risk sessions that look legitimate on paper – a new account with a valid email, a real phone number, a matching address – but whose behavioral fingerprint matches known fraud patterns. That kind of pre-transaction intelligence is difficult to replicate purely from payment data, and it creates a detection gap that Stripe Radar doesn’t currently fill.
Sardine also benefits from a structural trend in fintech: the market is maturing past the “move fast and patch fraud later” mentality. Regulatory scrutiny of fintech companies has tightened, fraud losses are increasingly visible to investors and boards, and the cost of a major fraud incident now includes reputational damage, not just charge-backs. That environment makes a purpose-built fraud and compliance platform easier to justify at the procurement level than it was three years ago.
Stripe, for its part, is not standing still. Radar has received consistent updates, and Stripe continues to expand its fraud-adjacent features. But Stripe’s core incentive is to keep merchants processing on Stripe – fraud detection is a retention tool, not a standalone product line. That difference in business model shapes everything from pricing to feature prioritization, and it gives Sardine a genuine strategic opening rather than just a tactical one.
The Limits of Sardine’s Challenge
Sardine’s advantages come with real constraints. Stripe’s distribution is enormous, and for the majority of businesses – online retailers, SaaS companies, straightforward e-commerce – Radar is good enough and the switching cost is near zero since it’s already built into their payment flow. Sardine has to work harder to reach those customers, and in many cases, it probably shouldn’t bother trying. The wedge is in specialized, high-risk, and heavily regulated categories, not the mainstream merchant market.
The more interesting question is what happens as Sardine’s customer base grows and its behavioral data set deepens. Fraud detection models improve with volume and diversity of signals. Stripe’s advantage has always been the scale of its network – it sees billions of transactions and can detect fraud patterns across its entire merchant base. If Sardine can accumulate enough cross-client behavioral data to build comparable network effects, the gap closes. But it needs more runway, more customers, and a few high-profile case studies before that network argument becomes credible against Stripe’s actual numbers.
