On 24 June 2026, FCA chief executive Nikhil Rathi delivered a speech at techUK’s Agents of Change conference setting out how the regulator intends to “re-think what it means to be an effective regulator in the age of AI.” While framed around artificial intelligence, the speech carries direct signals for anyone building or licensing crypto and tokenised-asset businesses in the UK.

What happened

Rathi argued that financial services must sit “at the heart” of the UK’s ambition to be a world-leading AI economy, providing the capital, infrastructure and trust to scale AI. He identified two scaling opportunities of particular relevance to digital-asset firms: agentic systems — AI that does not just summarise or detect, but “coordinate and transact” across retail and wholesale markets — and tokenisation, which he described as the “more automated and programmable infrastructure that agentic finance demands.”

On the tokenisation front, the FCA confirmed it approved Baillie Gifford, alongside Bank of New York Mellon, to launch the UK’s first natively tokenised authorised fund, with “the entire transaction journey, end-to-end, on chain.” The FCA and the Bank of England have also set out a direction for tokenised wholesale markets, and Rathi urged firms to respond to the related Call for Input, which closes shortly after the speech.

What it means in practice

The clearest message is that the FCA is moving from pure rule-writing toward what Rathi called stewardship as well as supervision. He was explicit that “legislation will never keep up” and that in some areas “traditional rule-making simply won’t work anymore.” Crucially, the FCA signalled it is willing to act before legislation catches up — citing Buy Now Pay Later, where it intervened well before the activity formally entered the regulatory perimeter.

For applicants, this means engagement and intent matter as much as box-ticking. Three themes stand out:

  • Accountability stays human. Even with agentic systems coordinating decisions, Rathi stressed that “accountability for regulated activities and outcomes must remain clear,” designed “with the right human oversight.” Firms cannot outsource responsibility to a model.
  • Competition is being actively championed. The FCA noted AI “lowers barriers to entry” for challengers and that its role is “not to protect incumbents.” It expects to use system-wide powers — under the Enterprise Act and the Digital Markets, Competition and Consumers Act — “more frequently” rather than as exceptional interventions.
  • Resilience and third-party dependency are now central. With 98% of operational incidents reported to the FCA relating to technology and cyber issues, and almost £1.3 billion lost to payment fraud last year, the regulator wants dependencies on cloud, model and data providers “properly mapped and governed,” and is leaning harder on the Critical Third Parties regime.

Implications for licensees and applicants

If you are pursuing a crypto / VASP license in the United Kingdom, this speech reshapes how you should present your business case. Tokenised funds, tokenised deposits and on-chain settlement are no longer hypothetical — the FCA has now authorised live examples, which raises the baseline expectation for governance maturity among new entrants in the same space.

Equally, firms deploying AI in onboarding, monitoring, trading workflows or fraud prevention should expect supervisory interest in model governance, oversight controls and supply-chain mapping. Note too that the FCA is itself exploring “agentic AI as our ‘first responder'” to monitor wholesale markets faster — meaning weaker market-abuse and conduct controls are more likely to be detected.

Concrete next steps

  • Respond to the Call for Input on tokenised wholesale markets if your model touches that area — early engagement shapes the regime you will operate under.
  • Document AI governance now: clear human accountability, oversight points, and mapped dependencies on model, cloud and data providers.
  • Track the upcoming publications Rathi flagged — the Mills Review on AI in retail financial services, and a later paper on good and poor AI practice — and align your controls accordingly.
  • Consider the FCA’s safe-testing routes, including the Supercharged Sandbox, AI Lab and the new Agentic Academy, to validate models against real-world data before scaling.

The FCA’s framing is that it is rethinking regulation “not to stand in the way of innovation, but to support the UK to seize the opportunity.” For crypto and tokenisation businesses, the door is open — but it favours firms that treat accountability, resilience and competition as design choices from day one.

Source: Financial Conduct Authority — Rethinking regulation for the age of AI