VERSE PRESS

Crypto News, Global First.

Visa Bets on Stablecoins and AI Agents as the Next Layer of Global Commerce

Visa is building infrastructure for a world where AI systems, not people, initiate and complete purchases. Its stablecoin settlement volume was running at approximately $7 billion annualised as of April 2026, and the company is now explicitly tying that infrastructure to the emerging market for autonomous AI commerce.

|

At the Visa Payments Forum in June 2026, the company unveiled a suite of tools under its "Intelligent Commerce" platform designed to let AI agents browse, select, and pay for goods without human input at each step. The announcements included Agent Score (a tool that rates merchant websites for AI navigability), an Agentic Directory (a verified registry of legitimate agents and merchants), a Large Transaction Model for AI-driven fraud detection, and Tokenized Deposits, which enables banks to convert traditional deposits into programmable digital money without moving funds off balance sheet. Visa also announced a partnership with OpenAI to provide global network, credentialing, and security infrastructure within OpenAI's own agentic commerce platform.

Around the same period, Visa was named a design partner on Circle's purpose-built Layer 1 blockchain, Circle Arc, and committed to operating a validator node on the network. That relationship deepens Visa's multi-chain settlement strategy, a theme running through much of its recent infrastructure work.

The strategic logic Visa is advancing frames card rails and stablecoins as complementary rather than competing. Card infrastructure handles the consumer-facing interface; stablecoins (digital currencies pegged to fiat, in this case primarily USDC on multiple blockchains) handle settlement underneath. "AI is transforming the front end of commerce," said Jack Forestell, Visa's Chief Product and Strategy Officer. "Stablecoins are reshaping the back end." Forestell described Visa's role as enabling this to "work securely, reliably and at global scale, for every participant in the ecosystem."

Why Traditional Card Rails Fall Short for Machines

The architectural mismatch between legacy payment infrastructure and AI agents is practical, not theoretical. An AI agent executing a purchase does not browse a checkout page or enter a PIN. It needs machine-readable pricing, programmable permission structures, and the ability to settle transactions instantly, including those worth fractions of a cent. Standard card networks require human authentication, static payment pages, and one-to-three-day settlement windows. Stablecoins, which settle in seconds on public blockchains and operate continuously without banking hours, address those gaps directly.

Early data from the x402 protocol, an open-source micropayment standard built on Coinbase's Base blockchain, illustrates the scale this can reach quickly. The protocol processed over 100 million cumulative transactions by the first quarter of 2026, with approximately 69,000 active agents as of late April 2026 and around $50 million in total volume. These figures originate from Coinbase and Base network analytics and have not been independently verified against on-chain block explorer data.

The majority of those transactions fall between $0.01 and $0.10. Visa's own stablecoin settlement program now spans nine blockchains, including Ethereum, Solana, Avalanche, Stellar, Base, and Polygon, after expanding in April 2026 from four chains. According to Visa, more than 160 stablecoin-linked card programs are live or in development globally.

Emerging Markets: Opportunity With a Structural Caveat

The regional implications of this shift are sharpest in Sub-Saharan Africa, South Asia, and Southeast Asia, where payment infrastructure is both rapidly growing and structurally fragmented.

Africa already has significant on-chain activity. Sub-Saharan Africa received approximately $205 billion in on-chain value between July 2024 and June 2025, a 52% year-on-year increase, with Nigeria accounting for $92.1 billion of that, according to the Chainalysis Geography of Cryptocurrency 2025 report.

The average cost of sending remittances to the region remains around 8.46%, according to the World Bank, compared with a global average of approximately 6.4% and a bank-channel average of approximately 15%. Proponents of stablecoin-based settlement argue that programmable, near-instant transactions on public blockchains could compress those costs considerably, though independent projections on the scale of that compression remain limited.

The risk is exclusion rather than inclusion. Visa's agentic infrastructure is currently designed around card rails and USDC settlement. It is not yet built to interface with M-Pesa, MTN Mobile Money, or Airtel Money, the dominant last-mile financial systems across East and West Africa. Oluwatobi Ajayi, CEO of African payments startup Ivorypay (a company with a direct commercial interest in this space), put it plainly: "Agentic payments will remain useful to crypto-native developers in the United States and Europe, while African businesses are left with the same old problem: global demand, but no efficient way to collect."

Southeast Asia presents a contrasting regulatory picture. In April 2026, Singapore issued a Know-Your-Agent (KYA) framework, becoming the first major regulatory jurisdiction to address AI agent identity in financial transactions directly. Regional startups are already building to that standard. Stables, for example, is developing AI-native payment middleware targeting Asia's B2B e-commerce market, which is projected to reach $28.9 trillion in 2026.

India is moving differently. In February 2026, at the India AI Impact Summit, Razorpay and the National Payments Corporation of India piloted agentic payments over UPI (India's state-built instant payment rail) within Anthropic's Claude. Users could order food from Zomato, Swiggy, and Zepto inside a Claude conversation, with payments authorised in advance rather than confirmed per transaction, via a feature called UPI Reserve Pay. NPCI is now developing a Unified Agent Protocol to standardise how AI agents authenticate and transact within that system.

UPI processed 22.71 billion transactions worth roughly 28.92 trillion rupees in June 2026 alone (approximately $340 billion USD at prevailing exchange rates; the dollar figure is an approximation not sourced directly from the underlying data). India's approach keeps AI-agent commerce on sovereign rails domestically while leaving room for stablecoin settlement in cross-border flows, an arrangement that fits Visa's hybrid model naturally.

What Comes Next

The total stablecoin market sits at approximately $303 billion as of mid-July 2026, with USDC at $73.4 billion (24% share) and USDT at $184.2 billion (59% share). This figure is drawn from market data as of July 12, 2026 and should be checked against current sources at the moment of publication, given the pace of change in stablecoin supply. Citi projects the total supply could reach $1.9 trillion by 2030 under a base-case scenario.

The regulatory environment is taking shape, though critical provisions are still arriving. The US GENIUS Act, signed in July 2025, provided legal clarity for domestic stablecoin issuance. Singapore's Know-Your-Agent framework, issued in April 2026, was the first major jurisdiction to address AI agent identity in financial transactions directly. The EU AI Act's high-risk provisions become enforceable on August 2, 2026, with agents that execute financial transactions expected to fall under regulated categories. That development will directly shape how agentic commerce platforms are designed and deployed across European markets.

The infrastructure layer is largely being built, and multiple organisations are competing to define its standards. The x402 protocol from Coinbase and Google's Agent Payments Protocol both represent attempts to establish the technical foundation for machine-to-machine commerce. The unsolved problem is agent identity and reputation. No portable, cross-chain standard currently exists for verifying who or what an agent is before it spends money.

That gap represents the most significant open engineering and regulatory challenge in the space, and whoever solves it will sit at the centre of a market that is moving faster than most of the frameworks meant to govern it.