ROCKSTRATEGE · ES GIBT EINE KI DAFUER

EGEKID DOSSIER · ENGLISH EDITION

When the Regulator Is Reading Over Your Shoulder: What BaFin’s AI Supervision Means for Your Portfolio

EGEKID Dossier · 30 July 2026 · Udo A. Schacht (“Rockstratege”) · Reading time approx. 7 min · Translated from the German original

1 – What happened

BaFin, Germany’s financial supervisory authority, has announced that it will begin formally reviewing how banks and insurers use artificial intelligence [1]. Until now, this was largely an internal matter: an institution would build an AI algorithm for credit assessment or fraud detection, test it in-house, done. Now a formal supervisory obligation is being added – comparable to what we already know from banking regulation, only for algorithms instead of capital ratios.

At first glance, that sounds like more paperwork. And it is. But for you as an investor, the real story lies elsewhere: a previously invisible grey area is becoming a visible, measurable criterion. And whatever becomes measurable eventually gets priced in – in the share price.

2 – The mechanics behind it

Why does an announcement like this move valuations at all? Quite simply: regulation shifts costs and risks between “invisible” and “visible”. Until now, the risk of a poorly controlled AI – say, a credit-scoring model that discriminates, or a trading algorithm that spins out of control under market stress – sat on the balance sheet like a hidden option. Nobody priced it, because nobody could see it clearly.

With BaFin’s supervision, this changes in two directions at once:

• The cost side: Documentation, model monitoring (the ongoing verification that an AI still behaves as intended) and compliance software cost money. In the short term, that weighs on margins.
• The risk side: An institution that documents its AI properly and makes it auditable reduces its legal and reputational risk. Less risk potentially means cheaper access to capital – lenders and investors demand a smaller risk premium.

This tension – more expensive in the short run, safer in the long run – is the essence of every new wave of regulation. The only question is: who benefits in the end, and who gets stuck with the bill?

3 – The historical evidence

We have seen this movie before – with Basel II and Basel III, the international frameworks that tightened capital requirements for banks after the financial crisis. Back then, too, it played out in two phases:

• Phase 1 – Uncertainty and volatility: Nobody knew exactly who would be best placed to meet the new rules. Share prices swung more sharply as the market had to reprice.
• Phase 2 – A valuation premium for the well-prepared: Institutions with solid capital positions and clear governance ultimately earned a trust premium. Weakly capitalised players traded at a permanent discount – or disappeared from the market altogether.

I experienced this first-hand several times during my 21 years at NORD/LB: new supervisory rules initially feel like sand in the gears for those affected. But in the end, it is often that very sand that weeds out the poorly oiled machines – while the well-oiled ones run more smoothly afterwards and are rewarded for it in the capital markets.

Applied to AI in finance, this means: whoever can already demonstrate today how their models work, what data goes in and how errors are detected, will hold an advantage tomorrow. Whoever runs their AI as a black box will be priced as riskier going forward – not because the technology is worse, but because the risk behind it remains opaque.

4 – What matters now

For you as a private investor, this does not mean selling or buying financial stocks tomorrow. It means: a new criterion has appeared on the radar – one you should track going forward when investing in banks, insurers or their technology providers.

• Watch for annual reports and investor relations communications that explicitly address “AI governance” or “model risk management”. This is no longer marketing filler – it is becoming a supervisory expectation.
• Institutions that communicate transparently today are sending a signal: “We have nothing to hide.” In a new regulatory environment, that signal is worth real money.
• Keep an eye on the supplier side as well: companies offering compliance software, audit tools or model monitoring for financial AI could benefit from the new supervisory regime – regardless of whether any individual bank rises or falls.

Important: this is not a buy signal – it is a filter. Craft over hype means exactly that: you look more closely before you judge.

5 – The investor’s lens

How do you apply this concretely to your own portfolio or your next investment decision? Here are three tools you can use right away.

In the end, the conclusion stands: regulation is neither a drama nor a reason to panic. It is a filter that – with some delay – reveals who has mastered their craft and who merely rode the hype. And it is precisely in that filtering process that the interesting valuation gaps of the coming years will emerge.

Sources

[1] Germany’s financial watchdog to monitor AI use at banks and insurers
Further sources from today’s German briefing are listed in the original edition.

← BACK TO ROCKSTRATEGE.COM