
In a first-of-its-kind publication, the FCA’s Emerging Tech & Research team has published Emerging Technology Horizon Scan 2026, setting out how emerging technologies could combine to reshape financial services, market infrastructure and the control environment that supports trust in UK markets.
The report is designed to foster debate and encourage collaboration across the financial services ecosystem. It is framed around three broad areas: Personalised Intelligence, Synthetic Security and Programmable Finance. Each points to a different dimension of technology-led innovation, but the stronger message is one of convergence. AI agents, synthetic media, distributed ledger technology, tokenisation, stablecoins, smart contracts, digital identity and programmable settlement are developing in parallel and may interact in ways that create new opportunities and new risks.For wholesale capital markets and treasury, the scan is most relevant where it moves beyond consumer-facing innovation and into market integrity, operational resilience, evidential standards and infrastructure modernisation.
Synthetic (in)Security
The FCA describes Synthetic (in)security as a future in which simulated data and fabricated ‘truth’ become difficult to distinguish from real data and actual truth. The report’s treatment of synthetic evidence is the most direct read-across for compliance, surveillance, financial crime and investigations teams.
Traditional controls often treat consistency, professional formatting, plausible explanations and complete documentation as positive signals. The FCA’s analysis suggests those same features may become easier to manufacture at scale.
In wholesale markets, the risk is that synthetic identities, fabricated transaction histories, simulated audit trails, false negotiation records and domain-specific narratives could make misconduct harder to identify. The control challenge becomes one of evidential integrity: how firms establish authenticity when false evidence can be produced with the structure, tone and supporting detail of legitimate records.The FCA’s reference to “suspicious perfection” offers a useful practical lens. A file, payment instruction, ownership explanation or trade rationale that appears unusually coherent may require deeper scrutiny, particularly where it supports higher-risk activity or masks opacity in beneficial ownership, source of funds or transaction purpose.
Agentic Conduct
The report also raises questions for market abuse surveillance. It considers scenarios in which autonomous agents could engage in collusion, spoofing, pump-and-dump strategies or sentiment manipulation, either through deliberate orchestration or emergent behaviour in multi-agent systems.
That presents a difficult conduct problem. Existing surveillance models are built around identifiable actors, trade patterns, communications and intent. Agentic systems complicate each of those elements. A model may act within a mandate, optimise towards a target or interact with other systems in ways that produce market impact without a trader manually directing each step.
For capital markets firms, this points to the need for stronger governance around agent permissions, trading constraints, escalation rules, kill switches, audit trails and explainability. Surveillance teams may need to monitor the behaviour of automated systems as a distinct conduct population, rather than treating them solely as extensions of human users.
The same issue applies to accountability. If an agent takes an action that appears manipulative, firms will need to evidence how the instruction was set, what limits applied, how the system behaved, and whether control functions could detect and interrupt harmful outcomes.
Programmable Finance
The FCA describes a shift towards financial infrastructure that embeds protocol capabilities into traditional finance, including distributed ledger technology, tokenisation, programmable money, stablecoins, Central Bank Digital Currencies and smart contracts.
For wholesale markets, the report points towards a future in which assets, money and transactions execute according to pre-defined rules, with fewer manual steps and reduced reconciliation between intermediaries. For treasury teams, the implications include intraday liquidity, collateral mobility, cross-border settlement, programmable cash, atomic settlement and the interaction between existing payment rails and emerging forms of money.
The FCA also links programmable finance to the UK’s wider growth and competitiveness agenda. Digital identity, smart data, payment infrastructure, settlement modernisation, programmable assets and cross-border interoperability are presented as interlocking layers of a future financial stack. For firms operating across jurisdictions, this raises strategic questions around connectivity, legal finality, liquidity fragmentation, sanctions controls and the operational risks of interoperating across divergent national and regional infrastructures.
The report’s distinction between a unified ledger model and modular sovereign-ledger approaches is particularly important for global treasury. Future market infrastructure may not converge on a single global layer. Firms may need to operate across a more heterogeneous environment, where value moves between different forms of money, different settlement systems and different legal frameworks.
Personalised Intelligence
The FCA explores a future in which AI agents and proxies become the main interface between consumers and financial services firms, taking on comparison, negotiation, switching and decision-making tasks.
For wholesale markets, the parallel is machine-to-machine intermediation. Similar delegation models could emerge in treasury optimisation, liquidity routing, client service, compliance monitoring and investment workflows. The governance questions are consistent: who does the agent serve, what authority has it been given, how is intent captured, and how can outcomes be explained, challenged or reversed?
This is especially relevant where firms deploy AI agents in control or execution environments. A treasury agent that optimises cash, a compliance agent that triages alerts or a client service agent that negotiates terms may improve efficiency, but each creates a new layer of delegated judgement.
Collaboration and Control
FCA Chief Executive Nikhil Rathi’s recent speech on Rethinking regulation for the age of AI pointed to agentic systems and tokenisation as two scaling opportunities, including agentic systems supporting liquidity management, trading workflows and other wholesale market functions. He described this as “A profound step change to the structure and operation of markets.”
For wholesale capital markets and treasury firms, the practical questions are how automated authority is granted, how intent is captured, how synthetic evidence is challenged, how market abuse controls adapt to agentic behaviour, and how programmable settlement interacts with liquidity, collateral and cross-border infrastructure. Rathi’s message reinforces the collaborative message of the Horizon Scan: the FCA is sharing its thinking to support debate and practical engagement across regulators, firms and technology providers.
As he put it: “We need to build it with you.”
Subscribe to our newsletter


