Special Half-Day Event Series

AI in Financial Markets 2026: From Hype to Impact

20 October 2026

Half Day Event: 13.00 to 19.00 London Time

CMS London, Cannon Place, 78 Cannon Street, London EC4N 6AF

Get Tickets Now ↗

See Photos & Takeaways ↗

Loading details....

About the Event

Artificial intelligence is reshaping how financial markets operate, from trading and risk management to compliance, research, and client engagement. While debate continues about the sustainability of AI’s current boom, there is no doubt that the technology is transforming the way financial institutions think, act, and compete. AI in Financial Markets 2026 will explore how financial institutions can capture real value from AI while navigating the challenges of cost, scalability, regulation, and trust. The event brings together leaders from investment banks, asset managers, market infrastructures, technology firms, and regulators to discuss what’s working today, what’s overhyped, and where the real breakthroughs are emerging. The event aims to go beyond the buzzwords and focus on where AI is truly delivering measurable impact, and where the next opportunities lie for investors, institutions, and innovators.

Order of the Day

13.30 to 14.00: Registration Opens with coffee and biscuits

14.00 to 14.15: Opening Address: Future of Finance and Sponsor/Host

14.15 to 14.30: Keynote Address: A visionary keynote on the transformative impact of AI on capital markets and the emerging challenges around transparency and accountability

14.30 to 15.15: Panel 1 - What is different about the AI Bubble?

15.15 to 16.00: Panel 2 - Does AI work?

16.00 to 16.45: Coffee Break and Networking

16.45 to 17.30: Panel 3 - What are the use cases for AI in capital markets?

17.30 to 19.00: Networking drinks and canapés

Audience

Financial Institutions and Market Participants: Investment Banks, Broker-Dealers, Asset and Wealth Managers, Hedge Funds and Quant Funds, Exchanges and CSDs, Custodians and Post-Trade Infrastructure Platforms. Regulators, Policymakers and Central Banks. Technology and Infrastructure Providers. Consultants, Law Firms and Advisors. Corporate and Institutional Investors

All tickets are subject to approval. If your purchase is not approved, you will receive a full refund with no deductions. Refund requests may be made up to seven days before the event date. By purchasing a ticket, you agree to these terms.

Chatham House Policy

All panels are held under the Chatham House Rule to encourage open and candid discussion. All participants are free to use the information shared, but neither the identity nor the affiliation of speakers or attendees may be revealed.

To maintain this environment, sessions are not officially recorded or filmed, and no official event recordings are provided. Following each event, we will share key takeaways on our website and social media.

Panels & Key Discussion Topics

Keynote Address

What is different about the AI Bubble?

The scale of the investment in AI is far out of proportion to the revenues generated by it, with even paying customers not covering the costs of serving them. Is there any development in prospect that is likely to yield a viable, commercial business model for the current version of AI?
We have had AI bubbles before, notably in the 1960s (“GOFAI”), and again in the 1980s (rules-based “expert systems”), where different approaches still ended in disappointment. Why should this Bubble have a different outcome?
The Dot Com Bubble deflated 25 years ago with an 80 per cent fall in the Nasdaq Composite Index, and more than half of Dot Com companies failed. Are Nvidia, Open Ai and Anthropic the Amazon, eBay and Facebook of the current Bubble or its Boo.com, Pets.com and Global Crossing?
The technology underlying the Dot Com boom was not in question. Can the same be said about AI based on Machine Learning?
It used to be said that the current AI Bubble is different because it is funded by cash not borrowing, but the big AI companies are increasingly funding their investments in AI with debt. Is the difference narrowing?
Start-ups are claiming to be using AI because it increases their chances of getting funded. What does that signify?
Established firms (buyers and sellers) are pretending to use AI, claiming to be AI-powered or AI-enabled or AI-driven. Does that have real consequences (such as ignoring solutions which are not AI but might be better and cheaper) or is it just marketing?
The sums being invested in data centres, GPUs and electricity generation to support AI are already running into hundreds of billions of dollars, but the amount of capital needed to scale AI is estimated at trillions. Is the gap between investment need and investment available bridgeable?
Is the scale of the investment in AI constraining investment in other sectors?
To what extent does the massive investment in AI erect barriers to entry to smaller companies that might achieve more with less - and should regulators do anything about it?
Is the current approach to AI leading to concentration risk among AI service providers (akin to Cloud services)?
The flows of funds between AI companies in Silicon Valley are often described as incestuous and circular. Should this worry investors and lenders active in the AI industry?
There is a view that AI can overturn the economics of the Internet, which are currently based on harvesting the attention of consumers. Does that explain the willingness of the leading Internet technology companies to invest so heavily in AI?

PANEL 1

14.30 to 15.15

Dominic Hobson

AS MODERATOR

Does AI work?

Machine learning has delivered in narrow areas, such as document summaries, grammar and shopping recommendations, translation, transcription, and face and voice recognition. Is that narrowness a function of technological immaturity or a fundamental weakness of machine learning (i.e., that every task requires a unique model, assumptions and data set so AI on this model cannot ever develop into a general-purpose technology comparable to electricity, the internal combustion engine or digital computing)?
Machines "learn" by capturing the statistical regularities in work previously done by humans in the translation and transcription of texts and categorisation of objects. Does this reliance on the prior work of humans place an uncrossable boundary on their ability to excel at a task?
AI machines need to scan vast quantities of data to fill in the blanks in a line of code. Does this data inefficiency place limits on what is achievable or affordable by this method?
AI clearly improves incrementally, as shortcomings are fixed. Are the shortcomings getting harder to fix?
It is often said that AI aims only to aid or enhance the performance of workers, not to replace them. Is this a practical (or ethical) judgment or a confession of technical limitations?
Proofs of Concept and Pilot Tests regularly prove that AI works but, unlike software bugs, reliance on data rather than code means AI machines can still deliver predictions that are wrong but appear plausible. Is there a bias to success in AI experiments?
Is there survivorship bias in the literature about AI (i.e., are the lessons of failed AI projects being ignored)?
Is there a risk that humans will take AI too seriously, ignore its shortcomings and regard it as infallible, and therefore rely on it in situations where they should not - and does that require regulatory intervention?
AI jargon has expanded – Machine Learning, Generative AI, Large Language Models (LLMs), Natural Language Processing (NLP), Neural Networks, Agentic AI – but all models ultimately rely on the same thing: scanning large data sets. Is the jargon just marketing?
Is there a constituency inside large corporations that boasts of success in AI experiments because continued funding and employment depend on it - and are they hoodwinking senior management and outside investors?
Regulation (such as the EU AI Act) has so far focused on obliging providers to address AI issues such as safety, risk, privacy and bias, with fines for non-compliance. Are regulators over-estimating, under-estimating, understanding or misunderstanding, the current capabilities of AI?
Machine learning does not appear to offer a clear path to artificial general intelligence (AGI). Is there an alternative path that might?

PANEL 2

15.15 to 16.00

Dominic Hobson

AS MODERATOR

What are the use cases for AI in capital markets?

There are plenty of general, anonymous studies that claim to prove AI has achieved measurable improvements in processes, while individual cases that worked at scale inside a named organisation are much harder to find. What does that tell us?
The narrow but general use cases for AI (such as document summaries, transcriptions, translation, coding) are as useful in capital markets as in any industry. Can you think of a use-case specific to capital markets?
The common uses-cases in the front office (chatbots, algorithmic trading, investment research, robo-advice) and the middle and back office (transaction processing, financial crime compliance, market surveillance) do not feel transformational. Can you think of a transformational use case?
Can you think of a problem in capital markets that AI has solved definitively (as opposed to merely having a positive impact) ?
Can you think of an initially promising use-case that ultimately failed?
Might AI have deleterious effects in capital markets (e.g., increasing market speed and volatility, reducing transparency into market participants and widening the scope for fraud by manipulation of data sets)?
Some believe there are synergies between AI and Blockchain (e.g., portable data and identities, data verification, micro-payments, smart contracts). Do you agree?
Do you think AI and Blockchain are converging?
If the AI Bubble pops can the damage to the wider financial system be contained (e.g., are private credit markets lending to AI companies a contagion transmission mechanism)?
AI advertisements make lavish claims (such as "What if you could multiply yourself?" and "With the right tools, work doesn't have to feel like work"). How seriously should employers take these claims?
How would you describe the user experience of current versions of AI?
The Chinese Communist Party (CCP) has made AI the centre of its next five-year plan. They think AI is going to transform not just finance but science and technology, industrial production, defence, governance, global cooperation and the quality of life. Are you surprised to find the CCP and Silicon Valley saying the same thing?