Digital rupee at the intersection of financial innovation, instant settlement and AI-driven risk.
MUMBAI: Union Finance Minister Nirmala Sitharaman used her keynote at the Global Fintech Fest 2026 to push the Reserve Bank of India toward a more ambitious digital rupee, while issuing a pointed warning about artificial intelligence outpacing the safeguards meant to contain it. “I urge the Reserve Bank of India to further advance and also develop what it has initiated through both the wholesale and retail CBDC pilots,” she said, directly linking this call to the newly launched Demat 2.0 corporate bond tokenisation pilot unveiled by RBI Governor Sanjay Malhotra and SEBI Chairman Tuhin Kanta Pandey.
Why the Digital Rupee Suddenly Matters More
For years, India’s digital payments story centred on UPI, and retail CBDC often seemed redundant next to it, fast, free and already ubiquitous. Union Finance Minister Sitharaman’s remarks reframed the question entirely. UPI moves commercial bank money through clearing intermediaries; CBDC is direct central bank money with programmatic finality, exactly what tokenised securities and high-velocity capital markets need for their settlement leg. As she put it, “every tokenisation architecture being tested across the world eventually returns to the foundational question: what is the money leg made of.” Demat 2.0 supplied a live answer during the festival itself: REC Limited, Larsen & Toubro and IIFL Finance jointly raised ₹1,025 crore through bonds settled instantly using wholesale e-rupee via the RBI’s Unified Market Interface, with the bond token and the payment moving simultaneously so neither side is exposed to the other failing.
A Structural Upgrade
This atomic settlement model replaces the older system, where ownership records and payments moved on separate rails and took one to two days to reconcile, exposing all parties to counterparty risk in the interim. Under Demat 2.0, if either leg of the transaction fails, the whole trade aborts automatically and smart contracts now handle coupon payments and redemptions directly into bondholders’ digital wallets without manual processing. Existing investor protections and depository oversight remain fully intact; only the underlying technology has changed. Phase 1, covering institutional primary issuance, is live. Phases 2 and 3, covering secondary trading and eventual retail access, remain pending, with no published timeline yet from either regulator.
The Warning Nobody Wanted to Hear
Alongside this progress, Union Finance Minister Sitharaman delivered a considerably more sobering message on artificial intelligence. “AI acts as a double-edged sword,” she said, “it helps us spot fraud faster, but it also allows attackers to automate larger, more sophisticated cyberattacks.” Most strikingly, she referenced, without naming, a leading AI lab that now judges autonomous cyber capability to be moving faster than its own controlling mechanisms, a unsettling admission coming from within the industry itself rather than an outside critic. She went further, warning that certain AI models are capable of influencing public opinion and elections without people even realising they are being shaped.
Speed Cuts Both Ways in Finance Too
This concern connects directly back to the settlement infrastructure being built. “Speed works both ways,” Sitharaman noted, “when decisions and transactions become instantaneous, software errors, fraud and market shocks can spread at the same rapid pace.” Traditional settlement delays, however inconvenient, acted as informal circuit breakers, giving risk managers time to catch bad trades before they cascaded. Atomic settlement removes that buffer entirely, meaning any error introduced by an autonomous trading agent could propagate through the financial system in an instant.
SBI’s Answer: Know Your Agent
State Bank of India Chairman C.S. Setty offered a complementary framework at the same festival, arguing that as AI shifts from advisory tool to independent executor, banks need “Know Your Agent” protocols alongside existing Know Your Customer rules, covering verified digital identities for AI agents, hard transaction limits, immutable audit trails and instant kill-switches. His standard was blunt: “There is nothing like 99% accuracy. You need to be 100% accurate and every time, all the time.”
Regulation Without Strangling Innovation
Union Finance Minister Sitharaman’s proposed answer is “soft-touch regulation,” developed closely with the RBI, designed to preserve room for experimentation while closing jurisdictional gaps. Her clearest line on this was direct: “where the regulator’s writ ends, your own standards must begin.” She welcomed the RBI’s recognition of the United FinTech Forum as a self-regulatory organisation on September 10 and proposed a federated industry platform to give India’s technology sector a stronger collective voice in global standard-setting.
What This Actually Means Going Forward
Three genuine tensions now sit at the centre of India’s financial technology policy: instant settlement speed against the loss of traditional risk buffers, a light regulatory touch against the risk of activity slipping through gaps between agencies and the growing role of sovereign digital money against the health of commercial banking balance sheets. None of these are simple problems, and Union Finance Minister Sitharaman’s own remarks reflect that complexity rather than offering easy resolution.
What to Watch
The coming months should show whether SEBI issues clear guidelines for Demat 2.0’s secondary trading phase, whether the RBI publishes formal guidance on Know Your Agent requirements for AI systems, and whether meaningful cross-agency coordination actually materialises among the RBI, SEBI and other regulators. For now, India has built genuinely functional settlement infrastructure. Whether its governance architecture can keep pace with the AI systems now operating inside that infrastructure remains the open question Union Finance Minister Sitharaman chose to leave squarely on the table.
About the Author
