From the rise of agentic AI and evolving regulatory frameworks in Singapore and the UK to accelerating Wall Street adoption, rapid fintech innovation and Scotland’s push to become a green-AI hub, AI continues to reshape global financial markets. Entering 2026, the momentum shows no sign of slowing. Helen Jewell, BlackRock’s CIO of Fundamental Equities for EMEA, expects AI-driven investment flows to remain strong – accompanied by heightened market volatility and a growing shift in algorithmic trading from raw speed to sophisticated AI models.

High-frequency trading still matters, but the true competitive edge now lies in data science. Firms are scaling machine-learning and deep-learning models to read sentiment and model risk, moving trading firmly toward “intelligence over speed” and requiring secure environments that protect proprietary models – and increasingly, autonomous agents.

This next phase of automation goes beyond today’s “copilot” tools that respond to prompts. Agentic AI can autonomously pursue goals, such as resolving trade discrepancies. Adoption is early but accelerating, alongside the wider democratisation of AI trading tools and stock screeners, here’s what stood out this week:

1. MAS Widens the Regulatory Net

MAS’ proposed Guidelines on AI Risk Management build on the FEAT principles, Veritas and Project MindForge, reflecting growing concern over risks from complex AI, GenAI and AI agents. Open for consultation until 31 January 2026, the Guidelines significantly broaden the definition of “AI” to include any model, system or use case that learns or infers. This could pull existing trading, optimisation and decision-support tools into scope, with proportionate obligations.

FIs must demonstrate:

• consistent AI identification
• a continuously updated AI inventory
• a risk-materiality framework based on impact, complexity and reliance

Boards must oversee AI risk appetite and governance, with enhanced scrutiny for material AI exposure. Firms must implement comprehensive lifecycle controls – data quality, explainability, fairness, testing/validation, secure deployment, drift monitoring, kill switches and change management – plus strengthened third-party oversight and adequate internal capabilities.
Read more: mas.gov.sg

2. From Singapore to the Thames: The UK’s Escalating Response

The Bank of England warned this week of systemic vulnerabilities linked to AI – not only inflated sector valuations but also leverage, funding pressures and herding behaviour. Its December Financial Stability Report signals that AI will sit at the centre of future regulatory reviews:
bankofengland.co.uk

The FCA also confirmed a shift in approach. CEO Nikhil Rathi noted that the pace of AI will require “a different relationship between regulator and regulated”:
ft.com

The FCA is pursuing a principles-based, flexible framework – intervening only in cases of significant, unmanaged risk – while encouraging innovation to support UK competitiveness. It has launched the first AI Live Testing initiative with Advai, complementing the Supercharged Sandbox. Applications for the second cohort open January 2026, with testing beginning April 2026.
Read more: https://lnkd.in/euyrS_KV

3. AI on Wall Street Keeps Growing

In line with MAS’s broadened definition of AI, Citadel has rolled out an AI-powered equities research assistant across most investor teams. CTO Umesh Subramanian says the $71bn hedge fund has used the tool for a year to scan transcripts and filings, summarise brokerage research and flag risks far faster than human analysts.

Subramanian emphasised caution: PMs must not outsource investment judgment to AI. He and founder Ken Griffin both noted that while AI improves efficiency, it will not, on its own, generate market-beating returns – performance will depend on how well investors use the technology.

Citadel’s move mirrors broader adoption across hedge funds including Man Group, Bridgewater, AQR and Viking Global, all deploying AI for alpha generation, strategy testing and signal discovery.
Read more: https://www.reuters.com/business/citadel-debuts-new-ai-tool-equities-investors-cto-subramanian-says-2025-12-03/

4. Integration with Traditional Finance & the Next Wave of Fintech

Announcements this week highlight how rapidly AI is permeating secondary markets – and how differently this plays out for incumbents versus challengers. J.P. Morgan notes that collapsing model-running costs are accelerating adoption but also driving early commoditisation, with foundational AI models likely to face margin pressure similar to telecoms and ISPs:
https://www.businessinsider.com/ai-stocks-market-outlook-dot-com-bubble-prediction-jpmorgan-2025

JPM argues value is shifting to the application layer: workflow integration, proprietary data and domain-specific tools – areas where incumbents hold the advantage.

Recent deals underline this, including LSEG’s expanded AI-analytics work and the FT’s partnership with RavenPack, where FT Ventures has taken a stake and licensed real-time and archival journalism into RavenPack’s Bigdata.com analytics platform:
https://www.ravenpack.com/blog/ravenpack-financial-times-investment-content-integration

These integrations give institutional clients richer context for backtesting and event-driven strategies, enable AI agents to merge structured data with narrative and sentiment, and support advanced quant, systematic and news-driven approaches. For incumbents, they monetise proprietary content, strengthen governance and embed AI into critical market infrastructure – advantages fintech challengers cannot easily replicate. Yet falling model costs still lower barriers for fintech experimentation, even as defensibility increasingly requires unique data or distribution. Showing just how fast AI is diffusing culturally, Perplexity’s new partnership with Cristiano Ronaldo signals AI’s shift into mainstream infrastructure – mirroring markets’ relentless search for any edge:
https://lnkd.in/g6uuwPSP

5. Scotland the Brave?

Scotland is on the brink of a major opportunity, with 11 proposed hyperscale data-centre projects representing 2–3 GW of demand – up to three-quarters of current winter peak load. This positions Scotland as a potential green-AI powerhouse, powered by abundant offshore wind, a cool climate ideal for advanced cooling systems, and new fibre routes linking the UK, Europe and North America.

These sites could form a unified digital-energy ecosystem converting renewable surplus into AI compute, accelerating grid reinforcement, anchoring transmission lines and spurring innovation across universities and industry.

The challenges, however, are significant: multi-year grid-upgrade delays, rural infrastructure gaps, EPC and cooling-OEM bottlenecks, GPU shortages and planning constraints. If Scotland clears these hurdles, it could build a resilient, energy-integrated AI corridor; if not, workloads may shift to Norway or the wider Nordics. Read more here – https://www.datacenterdynamics.com/en/news/planned-hyperscale-data-centers-in-scotland-could-demand-2-3gw-of-power-report/.

As AI becomes more deeply embedded in trading, asset management and the underlying infrastructure of financial markets, the industry will need increasingly advanced fintech tools and substantial data-centre capacity to support this transformation. But as AI moves further into the financial mainstream, regulatory scrutiny and expectations will rise in parallel. One to watch closely as we head into 2026.

Thank you for reading. I’d love to hear your thoughts – what you agree with, where you differ, and what you’d like to see more of next time.

Best wishes,
Rebecca