Razorpay today announced the launch of Razorpay Vulcan, a transformer-based AI foundation model designed specifically for payments. Built using technology from NVIDIA and AWS, the model is designed to improve payment success rates, fraud detection, risk assessment and checkout experiences across India’s digital payments ecosystem.
Razorpay said the model has been trained on approximately 3 trillion data points across 4 billion payments, with around 3,000 signals analysed per transaction. The company said the model is designed as a shared intelligence layer across multiple payment functions rather than as separate models for routing, fraud, risk and checkout.
The development comes as India’s digital payments ecosystem continues to expand across UPI, cards, net banking, wallets and other payment methods. Razorpay said its internal study covering 1.5 million shoppers and more than 51,000 businesses identified recurring issues around failed transactions, payment delays and drop-offs across different markets.
Components of the model have already been deployed across Razorpay’s network for routing, fraud and risk decisions. According to the company, early deployments with customers including Blinkit, Bachat and redBus have resulted in an 8-10% improvement in payment success rates, eight times more international card fraud being detected and stopped, and five times more fraudulent or disputed transactions being identified without an increase in alerts. Razorpay also said 40% more shoppers are seeing their preferred UPI application on its Magic Checkout, helping complete an additional 1-2 lakh purchases each month.
Unlike conventional machine learning models that are generally developed for specific use cases, Razorpay said its foundation model is designed to learn patterns across the payments ecosystem and extend that understanding to multiple use cases. The model is not an LLM; instead, it is trained to interpret transaction and payment behaviour across the company’s network.
Razorpay said the model’s architecture and training data have been developed in-house. NVIDIA GPUs were used to train and run the model at scale, while AWS infrastructure, including Amazon SageMaker, supported its development, training and deployment.
The model is designed to support several payment functions, including real-time routing, network-level fraud detection, risk assessment for cash-on-delivery orders and personalised checkout recommendations.
For merchants, Razorpay said the model is intended to reduce failed transactions, payment drop-offs, fraud losses and return-to-origin orders. For consumers, the objective is to improve the reliability and predictability of digital payments.
Harshil Mathur, CEO and Founder, Razorpay, said, “India’s appetite for digital payments is real, but it isn’t universal yet – for a large part of the country, going digital still comes down to one thing: does it work every single time? That’s the customer we built this for: the one still deciding whether to trust a screen over cash in hand. An AI-led payments foundation model doesn’t just solve today’s problem and stop there. Every payment teaches the system something that makes the next payment better.”
Pahal Patangia, Head of Global Industry Business Development and Payments, NVIDIA, said, “India’s rapidly evolving digital economy is creating an opportunity to make payments more intelligent, reliable, and secure. NVIDIA’s work with Razorpay in partnership with AWS on AI payments foundation models has opened up a new frontier, turning complex payments data into real-time contextual intelligence.”
Kiran Jagannath, Head of FSI and Conglomerates, AWS India and South Asia, said, “Razorpay is reimagining payments intelligence at India scale with an AI foundation model – built on Amazon SageMaker – that consolidates billions of transaction insights into a single, continuously learning intelligence layer, replacing fragmented ML models with unified AI that delivers higher payment success rates, rapid iteration, and enterprise-grade security for mission-critical payment flows.”
Razorpay said it plans to extend the model’s use across payment authentication, routing, fraud detection and lending as it continues to develop a broader AI-based intelligence layer for digital payments.