🇮🇳印度·创业公司·2026年8月21日·来源： Fintech News Singapore

Razorpay Launches AI Model for Payments Built With NVIDIA and AWS

内容仅提供英文版本

Razorpay, an Indian fintech major, has launched Vulcan, an AI foundation model designed to enhance payment success rates and bolster fraud detection across its network. The model was trained on nearly three trillion data points from four billion payments, leveraging NVIDIA GPUs for processing and Amazon Web Services (AWS) for development and deployment. Vulcan consolidates various functions like routing, risk assessment, and checkout personalization under a single intelligence layer, moving away from disparate machine learning models. Early implementations have already shown significant improvements, including an 8-10% rise in payment success rates and enhanced fraud detection capabilities for customers such as Blinkit and redBus.

Razorpay's introduction of Vulcan signifies a pivotal advancement in India's digital payments landscape, demonstrating the increasing sophistication of AI applications in fintech across Asia. By consolidating multiple payment intelligence functions into a single foundation model, Razorpay is setting a new standard for efficiency and security. This move not only addresses critical pain points like failed transactions and fraud but also positions Razorpay at the forefront of AI-driven innovation, potentially influencing how other payment providers in the region approach their technology stacks.

The collaboration with NVIDIA and AWS underscores the growing trend of leveraging global tech giants' infrastructure for developing powerful, localized AI solutions. The reported improvements in payment success rates and fraud detection highlight the tangible benefits of such advanced models, which will likely drive greater adoption of digital payments in India and inspire similar initiatives in other high-growth Asian markets. This development also signals a shift towards more autonomous and self-improving payment systems, where every transaction contributes to the system's overall intelligence, promising continuous enhancement of user experience and operational efficiency.