FF News — The Fintech News Network

Sardine Launches AI Labs to Combat Financial Crime with Frontier Intelligence

By Lauren Towner · 30 September 2026

Press Release: Sardine Launches AI Labs to Combat Financial Crime with Frontier Intelligence | Featured Image by FF News

Sardine has launched Sardine AI Labs, an applied research group dedicated to utilizing foundation models to combat financial crime. For fintech professionals, this represents a significant shift toward domain-specific AI that leverages massive proprietary datasets to solve persistent industry challenges, such as the "cold-start" problem for new card issuers and the high rate of false positives in sanctions screening.

What was announced

The launch of Sardine AI Labs marks the creation of a specialized research group focused on how artificial intelligence can interpret the "language" of financial behavior. The lab is designed to address a critical gap in the industry: while academic researchers often lack access to large-scale, real-world datasets, Sardine is utilizing its proprietary network which covers more than 6.5 billion devices, 441 million consumers, and 3.4 million businesses. This dataset encompasses 6.6 billion transactions and $1.8 trillion in total payment volume.

Sardine AI Labs is also introducing a research fellowship program, offering up to five positions to independent researchers currently enrolled in universities across the United States or Canada. Applications for the inaugural cohort are open through December 13, 2026, with fellows set to be announced in mid-January 2027. These researchers will focus on areas such as identity matching, money-laundering detection, and real-time inference performance.

Alongside the announcement, the lab published early results from a card foundation model trained on approximately one billion transactions. The model was tested against card issuers that were entirely excluded from its training data to simulate a "cold-start" scenario. The results showed a 68% improvement in fraud detection accuracy for consumer card issuers and a 41% improvement for business card issuers when compared to traditional machine-learning methods. The lab intends to continue exploring modeling for large sequences of user behavior, transfer learning, and adversarial robustness to turn AI advances into measurable financial protection.

"We are uniquely positioned to train a model purpose-built for risk because we sit on the industry’s fastest growing fraud and fincrime dataset, which spans more than 6.5B devices, 441M consumers, 3.4M businesses, 6.6 billion transactions and $1.8 trillion in payments."

Soups Ranjan, CEO and Co-founder of Sardine.

The companies involved

Sardine is an agentic risk platform specializing in fraud and financial crime prevention. The company has established itself as a central player in the fintech ecosystem by providing a unified platform for identity verification, fraud detection, and compliance. Led by Soups Ranjan, Co-founder and CEO, and Niranjan Shetty, Head of Data Science, Sardine operates at the intersection of behavioral biometrics and transactional data. The firm’s infrastructure is designed to handle the high-velocity requirements of modern finance, processing trillions of dollars in payments across various rails including card, ACH, and wire.

In February 2025, the company secured $70 million in funding to enhance the productivity of fraud and compliance teams. This capital has supported its expansion into various sectors of the financial industry, including partnerships with banking-as-a-service providers and payment processors. By maintaining a network that spans billions of devices and transactions, Sardine leverages a scale of data that is often inaccessible to individual financial institutions. This proprietary dataset serves as the foundation for its new research initiatives and its broader efforts to automate risk management through advanced intelligence.

What FF News has reported before

FF News has closely followed Sardine’s growth and its expanding network of industry partnerships. In early 2025, we reported that Sardine AI Raises $70M to Make Fraud and Compliance Teams More Productive, a funding round that set the stage for its current research initiatives. Later that year, the company expanded its reach into the banking sector as Sardine and Helix Partner to Deliver Real-Time Fraud Monitoring and Compliance Capabilities to Sponsor Banks.

In 2026, the company turned its focus toward real-time payment security, as seen when Modern Treasury and Sardine Partner to Combat Real-Time Payment Fraud. Most recently, the firm has moved into the digital asset space, where Infinite and Sardine Partner to Bridge Stablecoin Compliance and Real-Time Risk Intelligence to address compliance in stablecoin transactions.

What this means

The launch of Sardine AI Labs signals a shift in the fintech sector away from general-purpose AI toward domain-specific "neolabs." By focusing on the "cold-start" problem, Sardine is addressing a critical vulnerability for new market entrants who lack the historical data to defend against sophisticated fraud rings. This move places significant pressure on legacy fraud prevention providers that rely on static, tabular data models, which often struggle to capture the nuance of sequential behavioral patterns. However, the industry remains at a crossroads regarding explainability. As these foundation models take on high-stakes tasks like sanctions screening and anti-money laundering, the primary challenge will be satisfying regulatory requirements for transparency in automated decision-making.

Companies in this story: Sardine

People in this story: Niranjan Shetty, Soups Ranjan

More from News