Across three continents, the financial services industry is quietly dismantling decades-old infrastructure and replacing it with something fundamentally different: systems designed not to assist human decision-making, but to operate autonomously within it. The shift is no longer theoretical. It is showing up in earnings reports, regulatory filings, and strategic partnerships from New York to Helsinki.
Mastercard's fourth-quarter 2025 results offer one of the clearest windows into how far this transformation has progressed. The payments giant reported 15% net revenue growth year-over-year on a currency-neutral basis — a figure that masks where the real momentum lies. Its Value-Added Services segment, which encompasses AI-powered analytics, fraud detection, and identity verification, grew at 22%, with 19% of that organic. More than 40% of all Mastercard transactions are now tokenized. Over 70% flow through its switched network, up 10 percentage points since 2020. These are not incremental improvements. They are indicators of an infrastructure that has been systematically rebuilt around machine-readable, programmable transaction flows — a foundation compatible with the autonomous systems now being layered on top.
The most consequential of those systems is AgentPay, Mastercard's emerging framework for enabling AI agents to initiate and settle payments within defined policy guardrails. The concept addresses a problem that financial institutions on every continent are grappling with: how to grant an AI system spending authority without sacrificing compliance, auditability, or regulatory standing. AgentPay's answer is to embed controls directly into the payment rail itself, rather than relying on downstream human approval workflows. The implications extend well beyond convenience. In markets where financial compliance infrastructure is underdeveloped — across much of Sub-Saharan Africa, South and Southeast Asia, and parts of Latin America — programmable, policy-embedded payment rails could leapfrog traditional oversight mechanisms entirely, for better or worse.
JPMorgan Chase is approaching the same challenge from the capital allocation side, backing a portfolio of AI startups focused on financial infrastructure. The bank's strategic thesis holds that the next generation of financial tooling will be built on foundation models capable of reasoning simultaneously across regulatory frameworks, risk models, and live market data — a capability that static, rule-based legacy systems cannot replicate. That bet has global implications: JPMorgan operates in over 100 countries, and the AI infrastructure it backs today will shape the compliance and risk architecture of cross-border finance for years to come.
In retail investment, the divergence between AI-native platforms and traditional brokerages is becoming measurable. eToro, the Israeli-founded platform with a significant European and emerging-market user base, has outperformed benchmarks by integrating AI-driven portfolio construction with social signal aggregation — a model that blends quantitative analysis with crowd sentiment in ways that conventional fund managers have struggled to replicate at scale. In the United States, Robinhood and Coinbase are integrating AI into customer-facing advisory features and fraud prevention. Coinbase, notably, has demonstrated revenue resilience through sustained cryptocurrency market volatility — an outcome its leadership attributes in part to an AI-enhanced risk management stack capable of real-time exposure adjustment.
The governance dimension is drawing increasing attention from regulators in Brussels, London, and Washington alike. Firms such as FairPlay AI and Cleareye.ai are embedding explainability and bias-detection directly into lending and underwriting pipelines, responding to regulatory pressure — particularly from the European Union's AI Act and evolving guidance from the US Consumer Financial Protection Bureau — for auditable, non-discriminatory AI in credit decisions. This is not a peripheral concern. Access to credit remains one of the sharpest fault lines in global economic inequality, and the algorithms now determining creditworthiness in Lagos, Jakarta, and São Paulo carry consequences that extend far beyond financial returns.
Contact center automation is following a parallel trajectory. Cresta is deploying large language models inside financial services call centers, automating complex customer interactions while generating audit trails for compliance review. In markets where multilingual support and regulatory documentation requirements have historically made customer service operationally expensive — much of continental Europe, the Gulf states, and East Asia — LLM-based automation offers both cost reduction and, potentially, improved consistency.
At the research frontier, the most forward-looking experiment may be unfolding in Finland. OP Pohjola, one of the Nordic region's largest cooperative banks, has partnered with quantum computing firm Qutwo to explore quantum-AI hybrid models for financial risk assessment. The Nordic financial sector has long served as a global laboratory for banking innovation — it pioneered real-time payment infrastructure and led early open banking adoption — and this latest experiment suggests the next frontier lies at the intersection of quantum computation and machine learning. If viable at scale, quantum-AI risk models could render current stress-testing and portfolio optimization methods obsolete, with consequences for every institution managing systemic exposure.
The aggregate picture is of an industry in structural, not cyclical, transition. AI is not being added to financial systems. It is becoming the substrate on which those systems run. The institutions that recognized this earliest — and rebuilt accordingly — are already reporting the returns. The institutions, regulators, and economies that have not yet confronted this shift face a widening gap that, unlike previous technological cycles, may prove difficult to close incrementally.
Sources:
1 Yahoo Finance, "Earnings live: eToro surges after Q4 profit beat, Medtronic stumbles, DTE Energy pops" (February 17, 2026)
2 Yahoo Finance, "Earnings live: Palo Alto Networks stock sinks after company cuts full-year-forecast" (February 17, 2026)
3 Yahoo Finance, "Earnings live: Supermicro, Eli Lilly stocks pop on upbeat forecasts, AMD and Uber slide" (February 04, 2026)
4 Yahoo Finance, "H&R Block Reports Fiscal 2026 Second Quarter Results" (February 03, 2026)
5 Nasdaq, "Mastercard (MA) Q4 2025 Earnings Call Transcript" (January 29, 2026)