Skip to content
Banking & Compliance

FCA Market Integrity Warning: Rathi Says The Threats Are The Most Serious He Has Seen, And Names AI Agent Coordination As A Test Case

At Mansion House on 22 October 2026, FCA chief executive Nikhil Rathi said threats to market integrity are as serious as he has seen them, naming 24/7 trading, prediction market insider dealing and the point where agentic coordination becomes manipulation. Two days earlier his deputy set out how the FCA uses AI itself. What the two speeches tell trading firms.

AlchmAI Editorial7 min read

60,000+

Professional services businesses being brought into the FCA's remit (FCA, Rathi speech, 22 October 2026)

8%

Share of banks' workforces dedicated to fighting financial crime (FCA, 22 October 2026)

+89%

Rise in applications to FCA innovation services in the strategy's first year (FCA, Pritchard speech, 20 October 2026)

+40%

Rise in UK investment fraud losses in 2025, a UK Finance figure cited by the FCA (20 October 2026)

FCA market integrity is under more pressure than at any point in Nikhil Rathi's tenure, by his own account. Speaking at the Corporation of the City of London's annual City Dinner at Mansion House on 22 October 2026, the FCA chief executive described the opportunities in digital markets as vast, then added: "But the threats to market integrity are as serious as I have seen them." He named 24/7 trading, insider dealing linked to prediction markets, cross-border crime and the question of "when agentic coordination becomes market manipulation".

Two days earlier, deputy chief executive Sarah Pritchard told the Investment Association's annual dinner how the FCA is using AI inside its own supervision. Read together, the speeches set out what the regulator worries about, and how it plans to find it. The FCA notes that both texts are drafted speeches that may differ slightly from what was delivered.

What Did Rathi Say At Mansion House On 22 October?

The speech, published as Market integrity in a digital age, made speed its theme. "We must move faster and focus on outcomes to protect market integrity without holding back innovation," Rathi said. His warning on the alternative was blunt: "If we maintain an analogue approach in a digital age, new products enter the market faster than we adapt."

He listed the threats as "Unprecedented volatility, concentration risks, new forms of market abuse and rampant cross-border financial crime". He placed the UK as the world's largest centre for cross-border banking, FX and derivatives trading, which makes it exposed to every one of those risks. He also pointed to scale on the crime side: banks dedicate 8% of their workforces to combating crime, and more than 60,000 professional services businesses are being brought into the FCA's remit.

He was also candid about resilience. Citing the AI minister, he said UK systems will likely be breached at some point, which is why reporting near misses and having insurance in place matter. He expects alignment with national and global AI standards to be a focus of the UK's G20 presidency, and he confirmed a joint FCA and Bank of England roadmap on digital assets is on the way. For trading firms, that is three separate pieces of work signalled in one evening: incident reporting, AI standards and tokenisation.

Which New Threats To FCA Market Integrity Did The Speech Name?

Four stand out for trading firms. The first is the change in wholesale market structure: decentralised finance, parallel market structures and 24/7 trading, where prices form over weekends while traditional venues are shut. The second is insider trading linked to offshore gambling sites and prediction markets, which happens largely outside the venues most surveillance tools watch.

The third is AI. Rathi said personal AI agents are changing how consumers use financial services and are expected to reach UK consumers soon, and that frontier models make it easier for agents to act with growing autonomy. His example of a hard call for experts was "when agentic coordination becomes market manipulation". The fourth is harm that starts outside the FCA's remit, such as mass digital harm on social media. He cited Meta Muse, launched six weeks before the speech and one of the fastest-downloaded apps in US history, as a case the Digital Regulators Forum is working on together.

The timing was pointed. The day before, Interactive Brokers had launched autonomous AI trading that lets connected AI platforms place orders without per-trade approval.

How Is The FCA Itself Using AI Agents?

Rathi said the FCA uses AI agents to assess cases, summarise regulatory returns and scan large datasets for missing or concerning information. Pritchard made the same point to asset managers on 20 October: "AI already helps us cut through large data sets to identify risks sooner and act more quickly." Her framing of the wider shift matched her chief executive's: "Technology and AI mean markets are changing faster, and more fundamentally, than ever before."

Pritchard's numbers show a regulator trying to be faster at both ends. Applications to its innovation services rose 89% in the strategy's first year, a Baillie Gifford tokenised fund tested in the FCA sandbox was authorised in under a year, and streamlining the asset manager rulebook could save firms £128m a year. On the enforcement side, investment fraud losses rose 40% in 2025, the FCA issued more than 2,300 warnings about scams and unauthorised firms last year, and its Firm Checker tool has been used nearly 2 million times since its January launch.

Neither speech announced new rules for AI trading. The FCA's position remains that existing regimes, from market abuse rules to the Consumer Duty and senior managers accountability, apply to AI as they do to anything else.

Our View: An Outcomes Regulator With AI Will Ask For Evidence, Not Policies

The combination matters more than either speech. A regulator that focuses on outcomes, and that now reads returns and case files with AI agents, will find gaps faster than one reading them by hand. That changes what good preparation looks like. A written AI policy is no longer enough; the question will be whether a firm can show, from its own data, which orders came from algorithms or agents, what limits applied and what its surveillance flagged.

The prediction market and 24/7 points deserve attention from firms that think neither applies to them. Weekend price moves in crypto and tokenised assets already reach retail clients through broker apps, and event contracts are spreading into mainstream broker apps. Surveillance built for a 9 to 5 equity market will miss both.

“When the regulator reads your returns with AI agents, the gap between a policy and the evidence behind it gets found quickly. Market integrity is now a data question.”

What It Means For Trading And Fintech Firms

  • Brokers and CFD providers: check that surveillance runs on the hours you actually trade, including extended sessions and any weekend pricing, and that it can tell agent-originated orders from manual ones.
  • Prop trading firms: coordinated behaviour across funded accounts, whether from copy trading, shared bots or agents, is the pattern Rathi described. Your rule engine should look for it across accounts, not only within one.
  • Compliance leads: prepare an evidence pack, not a policy: a list of AI and algorithmic tools in use, who owns each, what limits apply and how alerts are reviewed. That is what an outcomes regulator will ask for.
  • Research and quant teams: if your models trade or inform trades, log inputs and outputs so you can show why a signal fired. Similar models reacting to the same data is the core of the agentic coordination concern.
  • Fintech and trading startups: the FCA's innovation services are busier and faster than they were. Firms building agent or tokenised products have a clearer route to talk to the regulator early.

The Bottom Line

Rathi's Mansion House speech put FCA market integrity at the top of the agenda: threats "as serious as I have seen them", 24/7 trading, prediction market insider dealing and agentic coordination named in one breath. Pritchard showed the other half: a regulator using AI agents on its own data, with innovation applications up 89% and fraud losses up 40%. AlchmAI, a London trading technology and AI studio, builds surveillance dashboards and AI controls for trading firms, and the message from both speeches is that firms will be judged on what their data can show, not on what their policies say.

References & Further Reading

Banking & ComplianceFCA market integritymarket abuseAgentic AI24/7 tradingNikhil RathiAI Agency UK
Share Email
AI

AlchmAI Editorial

Research and analysis, London

The AlchmAI team writes about the markets, technology and regulation we work with every day. We build trading platforms, real-time charts and AI analysis tools for brokers, prop firms and fintech teams from our office in Mayfair, London. Every article lists its sources. Nothing we publish is investment advice.

This article is general information and commentary. It is not investment advice or a recommendation to buy or sell any investment. Important information

Keep reading

All articles