rewrite this title XDC AI and the Rise of Agentic Finance: When AI Agents Learn to Pay – Decrypt

Today’s AI assistants can plan your trip, draft your emails, and find you a coffee shop down the street. What most still can’t do is buy the coffee.

The gap between recommending and doing is the missing piece of the “agentic economy,” where AI agents act on behalf of users to book services, pay for data, and settle transactions. For that, agents need a way to move money that is instant, programmable, and trusted.

AI companies have raised tens of billions of dollars, but much of what they’ve built sits in what the industry calls a “knowledge layer”—systems that generate recommendations rather than take actions.

“They’re not actually doing actions for you, and that’s for a reason,” Atul Khekade, co-founder of XDC Network, said in a recent FintechTV interview, citing the absence of “real-time settlement… compliance, dispute resolution, KYC, AML, and that exact trust layer which AI needs.” It’s a gap that companies from card networks to crypto startups are now racing to close.

The x402 standard

One key building block has emerged from that race: x402, an open payment standard introduced by Coinbase in 2025 that revives the HTTP “402 Payment Required” status code—reserved since the 1990s but never used.

In an x402 exchange, a server can charge for an API call or piece of data, and a client—including an autonomous AI agent—can settle it in stablecoins over the same request, with no account, API key, or human sign-off.

XDC Network is among the platforms building on the standard. XDC AI pairs x402 with “gasless” USDC settlement, so users don’t pay network fees directly—letting AI agents pay per request.

That combination, said XDC Network co-founder Ritesh Kakkad in a LinkedIn post, is “building the actual rails” for agentic commerce, enabling AI agents to “autonomously pay per API call, book services, and settle trades in real-time at sub-cent fees.”

Inside XDC AI

XDC AI is designed as shared infrastructure for agent payments. Rather than every business building its own rails, the company says, developers can plug into a common marketplace. In Khekade’s telling, that could let “coffee shops or DoorDash or travel companies” build “agentic applications” on the network.

XDC Network is no newcomer to institutional plumbing. Launched in 2019 by XinFin and co-founded by Khekade and Kakkad, the EVM-compatible Layer-1 is aligned with ISO 20022—the messaging standard used across traditional banking—and has focused on trade finance, payments, and real-world asset tokenization.

XDC Tech, the U.S. based institutional arm of XDC Network, has integrated Bridge, a Stripe company, bringing stablecoin settlement into its agentic AI roadmap. XDC calls the partnership a foundational piece of its plan to become the settlement layer for the agentic economy, a future where AI agents transact with other agents, businesses and humans at machine speed. XDC is plugging into infrastructure Bridge already runs, a faster path toward that vision than building it alone. “Every layer of finance is being rebuilt for a world where software, not just people, initiates the payment,” Khekade said. “This partnership gives our ecosystem stablecoin infrastructure that already meets that bar.”

The company recently showcased its XDC AI marketplace at its New York office, in a demonstration that drew more than 100 industry representatives from leading banks, tech partners, VC, family offices, AI tech firms and ecosystem builders, who are “very excited to see what we are building right now,” Khekade said.

The event demonstrated how agentic finance has become the “key translation point for AI to really transact and operate,” he said, adding that with XDC AI, “You can order a coffee. You can book your tickets.”

Connected through XDC AI’s MCP connector API in an assistant’s settings, a chatbot that could once only point you to a coffee shop “will actually do that transaction for you,” Khekade said. That, he added, is “quite transformative in a way, considering that they lack that kind of monetization layer right now.”

Khekade cast compliance and risk controls as central, arguing that autonomous software needs “a risk compliance and spending limits layer” to operate safely. XDC has “already built it and perfected it,” he said, with the network acting as a “cushion” for mainstream systems from AI compute and its ability to transact.

The road ahead

Agent-driven payments still have a way to go before they become as routine as tapping a card, with nascent standards and the industry still working through security, privacy, and control. But TradFi and crypto builders alike are already building for a future in which agents transact.

Kakkad, for one, is betting that the “convergence of AI agents, stablecoins, and enterprise blockchain,” particularly in trade finance and RWAs, will “unlock massive value.”