Bengaluru-based payments technology company Juspay reported revenue growth of 30% to ₹664 crore in FY26, up from ₹514 crore in FY25, while swinging to a net loss of ₹90 crore from a profit of ₹62 crore a year earlier, the company announced on 9 October 2026. The loss stems from accelerated spending on artificial intelligence capabilities, GPU infrastructure and international expansion.
Revenue has now more than tripled from ₹213 crore in FY23, a compound annual growth rate of about 46%. Adjusted EBITDA stayed positive at ₹33 crore excluding non-cash ESOP costs, and the core India payments business remained profitable. Its platform processed 400 million transactions a day during the year, with annualised total payment volume crossing the $1 trillion milestone at 99.999% reliability.
The SoftBank-backed firm now serves more than 500 enterprises and banks globally — including Amazon, Google, Flipkart, Swiggy, IndiGo and HSBC — across Asia-Pacific, Europe, the US, Latin America and West Asia. Less than 10% of revenue currently comes from overseas, a share the company expects to reach double digits in FY27.
The bigger strategic pivot: Juspay is turning the AI tools it built for internal use into SaaS products for global enterprises. Flagship products include Xyne, an open-source enterprise AI platform, Breeze Buddy, an AI voice platform handling a few hundred thousand merchant calls a day, and Juspay Genius, which runs payments operations on AI. Rather than paying AI labs per token, the company rents its own GPU clusters — an eight-node cluster costs ₹25–35 lakh a month versus 5–10x that as token spend.
Why it matters for founders
Juspay is the clearest Indian example of a profitable payments SaaS engine being redeployed into enterprise AI products. The lesson: internal tooling that saves you real money (its GPU economics alone cut AI costs 5–10x) can become the product line — but only if you invest before the market forces you to.
Find these funders and apply in one click — browse investor profiles at /investors and prepare your application with the /apply-kit.