Stripe Radar Built the Standard. Sardine Is Rewriting It.
Stripe Radar has long been the default fraud prevention layer for fintech builders – baked into Stripe’s payment infrastructure, easy to deploy, and good enough for most use cases. For years, “good enough” held. But a growing number of fintechs are discovering that Radar’s strength is also its limitation: it was designed for payment fraud, not for the specific, layered risk profiles that define modern financial products. Account takeovers, synthetic identity fraud, money mule networks – these are not edge cases anymore. They are the operating environment.
Sardine, a fraud and compliance platform built by former Coinbase and PayPal executives, is building directly into that gap.
The company has positioned itself not as a bolt-on tool but as a full behavioral intelligence layer – one that reads device signals, behavioral biometrics, and transaction patterns simultaneously, at the moment of account creation, not just at checkout. That distinction matters more than it might initially appear. Radar catches fraud at the point of payment. Sardine is designed to catch the conditions that enable fraud before a single transaction processes.
Why Behavioral Biometrics Changes the Fraud Stack
Stripe Radar operates primarily on transaction data – velocity checks, card testing patterns, geographic mismatches. It is a reactive layer, well-calibrated to flag anomalies in payment streams. Sardine’s approach starts earlier and runs deeper, pulling in signals like how a user holds their phone, how they type, how they navigate an onboarding flow. Behavioral biometrics of this kind can surface synthetic identity fraud that looks completely clean at the transaction level because the fraudster has built a coherent financial profile over months.
This is where fintech companies – crypto exchanges, neobanks, buy-now-pay-later platforms, money transfer apps – face problems that Stripe Radar was never designed to solve. These products onboard users, extend credit, and move money in ways that create long windows of exposure before any traditional fraud signal appears. A synthetic identity might pass KYC, open an account, build a small transaction history, and then extract credit or bonuses at scale. Radar sees a clean payment history. Sardine is watching the behavioral fingerprint the entire time.
The architecture also speaks to compliance. Sardine combines its behavioral layer with built-in AML monitoring and SAR filing support, which means compliance teams are not stitching together a fraud tool and a separate compliance stack. For fintechs that operate under Bank Secrecy Act obligations – which increasingly means almost all of them – that integration removes a meaningful operational cost. Stripe offers compliance tools through its ecosystem, but Radar itself is not a compliance product. Sardine is selling a single surface where both live.
The Fintech Migration and What It Signals
Stripe’s core infrastructure lock-in has always been its greatest competitive moat. Developers build on Stripe early, and the switching cost compounds over time – webhooks, reconciliation logic, API dependencies, the whole stack. Fraud detection sitting inside that same infrastructure felt natural. But fraud detection is increasingly where fintechs are willing to make a deliberate switch, because the cost of a bad fraud stack is not an engineering headache – it’s regulatory action, chargeback ratios that threaten card network standing, and in some cases, losing a banking partner entirely.
That calculus is creating a category of fintechs that continue to use Stripe for payments while routing fraud and compliance work to Sardine. The two products are not technically incompatible, and Sardine has reportedly built its onboarding to accommodate exactly this kind of hybrid stack. For Stripe, this split is worth watching. Radar is not just a product – it is part of the argument that Stripe is the only infrastructure layer a fintech needs. Every company that adds Sardine next to Stripe is, in a small way, rejecting that argument. The pattern Sardine is building is similar to what other vertical AI layers are pulling off in adjacent categories – Mercor’s AI hiring layer has done something comparable to ATS incumbents by solving a specific workflow problem better than the platform player.
Sardine raised a $51.5 million Series B in 2022 led by Andreessen Horowitz, with participation from Google Ventures and a roster of fintech-specific investors. Its client base includes crypto platforms and challenger banks, precisely the verticals where fraud complexity is highest and where Stripe Radar’s coverage is thinnest. The company has been quiet about revenue figures and customer counts, but the investor lineup and the category focus suggest it is not chasing Stripe head-on – it is winning the accounts where Stripe’s tooling breaks down first.
The Fracture Point
Stripe will not cede this ground quietly. The company has the resources to deepen Radar’s behavioral capabilities, acquire in the fraud space, or bundle compliance tooling in ways that make Sardine’s standalone pitch harder to land. But product roadmaps move slower than fraud does, and Sardine has spent years training its models on the exact transaction environments – crypto onboarding, peer-to-peer transfers, credit underwriting for thin-file users – where the next wave of fraud is already happening. The longer it takes Stripe to close that gap, the more institutional trust Sardine accumulates with the fintechs that are making infrastructure decisions right now.
Frequently Asked Questions
What does Sardine do differently from Stripe Radar?
Sardine uses behavioral biometrics and device signals to detect fraud before transactions occur, while Stripe Radar focuses on payment-level anomaly detection.
Is Sardine a direct competitor to Stripe?
Not entirely – many fintechs use Sardine alongside Stripe for payments, routing fraud detection and compliance to Sardine while keeping Stripe for infrastructure.
