Payments giants pivot from lending to AI agents as margin logic shifts

Toast, Block, Shopify and Adyen are quietly sidelining their lending arms as agentic AI rewrites the merchant-services value chain.

By Disrupts Editorial Team Mon, 31 August 2026

Long rows of dark server racks with reflective doors displaying green and blue indicator lights line a brightly lit white data center aisle.

A structural pivot is under way across the global payments industry. The companies that spent a decade turning merchant transaction data into loans are now turning that same data into AI agents, and their earnings calls are the clearest signal yet that the strategic calculus has shifted decisively.

Toast, the US restaurant-payments platform, reported second-quarter revenue of $1.9 billion, up 23% year-on-year, with net income nearly doubling to $154 million and a record 9,500 net new locations added. A year ago, Toast Capital, the company's embedded-lending arm, was positioned as its next growth engine. On this quarter's call, it received two sentences. The spotlight fell instead on Toast IQ, a new family of AI agents beginning with Toast IQ Grow, which automates a restaurant's digital marketing by drawing on the platform's own transaction data, guest ordering patterns, visit frequency, and lapsed customer signals, and ties every campaign directly to the orders it generates.

"Over time, we plan to build on Toast IQ Grow and roll out a series of agentic products on top of our software platform using our data and context advantage. This opens up a market opportunity beyond software," said CEO Aman Narang. "Marketing, scheduling and payroll, and bookkeeping and tax are services restaurants often pay for today. In many cases, they're spending a multiple of what they spend on software."

The lending product that disappeared from earnings scripts

The pattern repeats across the sector. Block spent more than a decade developing Square Loans on top of card-sales data from its merchant base; on its most recent earnings call, the product earned a passing mention in a list of gross profit contributors. The featured product was Managerbot, an AI agent embedded in the Square dashboard that monitors inventory, staffing, and marketing in real time. At Shopify, Capital surfaced once, inside a line about transaction losses, while Sidekick, the company's AI assistant for merchant analytics, drew the substantive airtime.

The economic logic behind the shift is straightforward. Software margins are structurally cleaner: a loan book carries credit losses, funding costs, and capital set aside against defaults, whereas an AI agent's primary cost is inference compute. The addressable market is also wider. Merchant lending is bounded by borrower capacity, but replacing the external agencies, bookkeepers, and marketing firms that merchants already pay for is a substantially larger revenue pool. And the risk profile is more forgiving: bad underwriting decisions surface as charge-offs two years after origination; a poorly performing agent surfaces as churn within a quarter.

Convergence implications: fintech meets agentic infrastructure

The strategic shift carries consequences well beyond the payments sector. What Toast, Block, and Shopify are building are, in effect, vertical AI agents trained on proprietary transaction graphs, a model that sits at the intersection of fintech infrastructure and the agentic-AI layer that enterprise software vendors, CRM platforms, and ERP providers are all racing to occupy. The competitive threat is no longer primarily from other payments processors; it is from general-purpose AI platforms that could replicate merchant analytics without the incumbents' data moats.

For investors, the recalibration is already visible in how analysts are framing these stocks. Adyen, which reported first-half processed volume up 24% year-on-year to €803.8 billion and EBITDA at a 49% margin, broke a two-decade no-acquisition rule twice in 2026, buying loyalty platform Talon.One and usage-based billing software provider Orb. The move signals that even the most engineering-purist operator in global payments now regards the post-transaction software layer as the real battleground.

Meanwhile, Nubank's Q2 results, $5.9 billion in revenue, up 39%, and a record $1.1 billion in net income, demonstrate that lending-led growth is far from exhausted in emerging markets. With 139 million customers and a credit portfolio of $39.4 billion, Nubank is actively shifting deposits into loans rather than away from them. The divergence in strategy between mature-market incumbents (pivoting to agents) and emerging-market challengers (doubling into credit) may prove one of the more consequential fault lines in fintech capital allocation over the next investment cycle.