Industry Outlook: Banking & Financial Services — Week of June 15, 2026

Market Outlook

• Tokenised deposits move from pilots to platforms. Visa’s plan to build a technology layer for tokenised deposits signals that programmable, 24/7 bank money is shifting from proofs-of-concept to commercial infrastructure. Combined with Digital Asset’s $355m raise for the Canton Network, institutional capital markets and retail payments are both converging on on-chain, but permissioned, rails. Banks that delay a tokenised-deposit roadmap risk ceding control of future money formats to card schemes and specialist fintechs.
• Real-time payments expand with Zelle’s India push. Zelle’s move to enable US users to send money to recipients in India, alongside work on a stablecoin for global expansion, underscores intensifying competition in cross-border retail payments. This will pressure banks’ legacy remittance products on price, UX and speed, while raising complex FX, sanctions and travel-rule compliance requirements. The direction of travel is clear: instant, API-driven and increasingly wallet- or token-based flows across borders.
• Fintechs climb regulatory ladder toward full banking. Canada’s Koho hitting unicorn status and explicitly targeting a federal banking licence, alongside Barclays’ acquisition of youth-focused GoHenry, illustrate a maturing fintech cohort moving into regulated balance-sheet models. The line between digital bank, payments fintech and broker/platform is blurring, particularly for younger and underbanked segments. Incumbents must assume that today’s partners can become tomorrow’s full-stack competitors in deposits, lending and payments.

Discussion: This week’s market signals point to programmable deposits, cross-border instant payments and licence-seeking fintechs reshaping the competitive baseline. CTOs should reassess their core payments and ledger strategies with an eye to tokenisation, 24/7 operations and partner-versus-competitor dynamics.

Headwinds

• AI hallucinations erode trust in external reporting. KPMG’s withdrawal of an AI adoption report after hallucinated claims about UBS and others is a warning shot for banks relying on AI-generated analytics, marketing content or investor communications. Misstatements about technology capabilities or risk controls can rapidly become reputational and regulatory issues, especially in prudential and conduct-sensitive domains. The incident will harden supervisors’ expectations around model governance, documentation and human oversight for any AI impacting disclosures.
• Regulatory uncertainty and politicisation of oversight. Fed Governor Michael Barr’s warning that deregulation threatens financial stability, combined with vacant minority-party seats at key US regulators and leadership changes at the CFPB, point to a volatile supervisory environment. Enforcement priorities may whipsaw across consumer protection, capital, AI use and third‑party risk, complicating multi‑year tech planning. Banks that treat compliance as a static checklist instead of an adaptable capability will struggle to keep pace with shifting expectations.
• Rising legal and IP risk in data-driven partnerships. Pagaya’s lawsuit accusing Klarna of trade secret misappropriation over underwriting technology highlights the fragility of data- and model-sharing arrangements between banks and fintechs. As more credit, fraud and marketing decisions are powered by third‑party AI, questions around IP ownership, retraining rights and reverse engineering are moving from legal fine print to material business risk. Poorly structured contracts or weak data segregation can turn strategic partnerships into protracted litigation.

Discussion: Governance, not just innovation speed, is emerging as the binding constraint on AI and partnership-driven strategies. CTOs should tighten AI model risk frameworks, strengthen IP and data controls in vendor/fintech contracts, and ensure compliance tooling can adapt quickly to regulatory swings.

Tailwinds

• AI-native fraud and identity tools gain momentum. Visa’s Scam Disruption unit flagging $2.6bn in fraud attempts and Fraudio raising fresh capital for AI-powered, real-time fraud prevention both underscore the payoff from advanced analytics in risk management. Parallel research on KYC/KYB/KYA shows identity orchestration is now a lever for operating-cost reduction and revenue expansion, not just compliance. Banks that industrialise machine learning across fraud, AML and onboarding can reduce losses while lifting approval rates and customer satisfaction.
• Youth and retail investor segments present growth runway. Barclays’ acquisition of GoHenry and TrueLayer’s data showing a 27% jump in Pay by Bank funding of trading accounts around the SpaceX IPO highlight two high-growth segments: youth banking and retail investing. Both are digital-first, event-driven and highly sensitive to UX frictions in payments and account opening. Modern APIs and embedded finance partnerships can let incumbents capture these flows without building every experience themselves.
• Quantum preparedness guidance from G7 central banks. The G7 Central Bank Quantum Technologies Working Group’s report on preparing for quantum technologies provides rare, concrete guidance from top regulators on crypto-agility and operational resilience. While practical quantum attacks on today’s cryptography are not imminent, supervisory expectations for migration planning and risk assessment are now on the record. Early movers can spread the cost and complexity of crypto-modernisation over several years instead of facing a compressed, regulator-driven scramble.

Discussion: Risk and compliance modernisation is becoming a strategic growth enabler, not a sunk cost. CTOs should prioritise AI-enabled fraud and identity capabilities, design products for digitally native youth and retail investors, and begin structured planning for quantum-safe cryptography.

Tech Implications

• Architecture for tokenised deposits and on-chain assets. Visa’s tokenised deposit layer and Digital Asset’s Canton Network funding round both assume banks will expose core balances as programmable, token-like instruments. This requires a clear architectural stance on how ledgers, smart contract platforms and existing payment rails interoperate, including settlement finality, reconciliation and regulatory reporting. Banks that treat tokenisation as an isolated pilot rather than a core architectural concern will face costly rework when volumes scale.
• Industrialising AI fraud, KYC and real-time monitoring. The growth of Fraudio and Visa’s scam unit, plus emerging KYC/KYB/KYA trends, point toward an always-on risk fabric embedded in transaction flows, not batch-based post-processing. Technically, this demands low-latency feature stores, streaming data pipelines, model deployment platforms and clear feedback loops from case management back into model training. Legacy, siloed fraud and AML systems will struggle to support the granularity and speed regulators and customers are beginning to expect.
• Governed AI and data pipelines to avoid hallucinations. KPMG’s AI hallucination issue is a reminder that large language models need strong guardrails, especially when used for analytics, reporting or customer-facing advice. Banks must invest in retrieval-augmented generation, provenance tracking, model evaluation and human-in-the-loop workflows before scaling LLMs into regulated domains. Robust data lineage and access controls are now preconditions for any serious AI deployment in risk, finance or compliance functions.

Discussion: Core architecture decisions around ledgers, data streaming and model platforms can no longer be deferred. CTOs should align tokenisation experiments with core banking modernisation, converge fraud/AML/KYC onto a shared real-time data and ML stack, and embed AI governance into every new data product.

CTO Action Items

This week, prioritise an enterprise view on tokenisation: assess how Visa’s tokenised deposit model and networks like Canton would integrate with your core ledger, and define a reference architecture rather than isolated pilots. Direct your risk and data teams to design a unified, streaming-based platform for fraud, AML and KYC/KYB/KYA, with clear ownership of model governance and feedback loops. Review all current and planned LLM use cases for exposure to hallucination risk in reporting, marketing and advisory contexts, and implement retrieval, provenance and human review where needed. Finally, initiate a quantum-readiness assessment of your cryptographic estate and ensure new systems are designed to be crypto-agile, so regulatory expectations can be met without a disruptive future retrofit.