Skip to content
56North

PaperDial · AI Act evidence

What a stable rule is worth in an unstable world

Pascal Mennesson
By Pascal Mennesson

Founder, 56North · Published · 9 min read

Europe has lost the race for speed. It can win the race for trust.

In my previous paper, I argued a simple idea: an AI you can prove is worth more than an AI you can only describe. Several executives answered, in substance: "Fine for proof. But Europe overdoes it. It regulates while others build."

The objection is serious and I will not dodge it. I want to turn it around, though. I believe that, in the world as it has become, European regulation is not first a brake. It is an asset. It does not only protect citizens: it protects those who invest, lend, buy and partner in the region. And the more the world comes apart, the more that asset is worth.

What an investor really pays for

An investor does not flee rules. An investor flees unpredictability.

A demanding but stable rule can be priced. You put it in a business plan, you provision for it, you pass it on in a price. An environment that shifts with elections, decrees or moods cannot be priced. You pay for it another way: through a risk premium, the extra return demanded to compensate for what cannot be anticipated.

That is the whole question. Regulation has a visible cost, paid up front. Instability has an invisible cost, paid in the storm. The first shows up in a compliance budget. The second shows up in a discount rate, a valuation haircut, a loan refused. European rules are often compared with no rules. They should be compared with uncertainty.

Europe was born from a rule against chaos

We forget that European integration did not start with culture, nor with money, but with market regulation.

In the 1930s, a group of economists and lawyers in Freiburg, around Walter Eucken and Franz Böhm, drew a lesson from the collapse of the Weimar Republic that would shape post-war Europe: a market left without a framework does not stay free. It drifts toward cartels, then toward political instability. Their school, ordoliberalism, advocates a market economy held by strong and stable rules. It inspired the German social market economy, then European competition law.

On 9 May 1950, Robert Schuman proposed pooling the coal and steel of France and Germany, the raw materials of war, under a common authority. The stated aim was to make war between the two countries "not merely unthinkable, but materially impossible". The founding act of the European Community, signed in Paris in 1951, was therefore a market rule in the service of stability.

The same logic shaped the currency. Germany had lived through the hyperinflation of 1923 and the ruin of its savings. It required the future European Central Bank to be independent from political power, and the Maastricht Treaty set it in stone in 1992. This was not legal vanity: it was insurance against the temptation to print money whenever convenient.

What Europe really exports

People say Europe no longer exports technology champions. That is partly true. But it exports something else, which is measured less well: trust, given a legal form.

In 1985, Europe created a label for investment funds sold to the general public, known as UCITS. Diversification rules, an independent depositary, mandatory disclosure to savers: the framework is demanding. Forty years later, according to the European fund and asset management association, these funds are distributed in more than fifty countries outside the EU, and their assets held outside the EU reach EUR 5.7 trillion, up 133% in ten years. Savers in Asia, Latin America and the Middle East buy a European product not for its speed, but for its rule.

The data protection regulation followed the same path. Adopted in 2016, applied since 2018, it was first mocked as a form factory. California adopted its own privacy law the same year, and many countries took inspiration from the European model. The legal scholar Anu Bradford, a professor at Columbia, gave the phenomenon a name: the "Brussels effect". Because its market is large and its rules demanding, Europe sets standards that companies worldwide end up applying everywhere, rather than run two regimes.

Three storms, three lessons

You do not measure the value of a sea wall in fair weather. Three recent episodes show what the rule protects when the weather turns.

March 2023: the bank whose rules had been relaxed. In 2018, the US Congress raised from 50 to 250 billion dollars of assets the threshold above which a bank is automatically subject to enhanced prudential standards. Silicon Valley Bank fell below that threshold, with lighter requirements. On 9 March 2023, its customers withdrew 42 billion dollars in a single day, and were set to withdraw nearly 100 billion the next. The bank was closed on 10 March. The Federal Reserve report published in April puts the failure of the bank's management first. But it also finds that the 2018 relaxation impeded effective supervision. The European Union, for its part, applies the Basel prudential rules to all its banks, whatever their size, with proportionate relief for the smallest. Relaxation is an immediate and visible gain. The risk it creates stays invisible until the day it materialises.

November 2022: the platform without a framework. FTX, then one of the largest crypto-asset platforms in the world, collapsed within days, with its customers' money. In 2023 Europe adopted the MiCA regulation, the first comprehensive framework for the sector in a major economic area, fully applicable since 30 December 2024. The result: players who intend to last come to Europe to seek a licence. The rule turned a market of distrust into a market of access.

January 2021: the aircraft that was not recertified by delegation. After two crashes, the Boeing 737 MAX was grounded worldwide in March 2019. In 2017, the European Union Aviation Safety Agency had validated the aircraft by relying on the US certification. When the time came for its return, it learned the lesson and refused to rely on the US authority alone, even though that authority had given its green light in November 2020. It ran its own review, its own test flights, and set its own conditions before authorising the return on 27 January 2021. Its executive director, Patrick Ky, stressed one point: the assessment was conducted fully independently of Boeing and of the US authority. For a passenger, as for an airline buying these aircraft, that refusal to delegate is a guarantee.

Why the argument is stronger today than yesterday

In a stable world, regulation looks like a burden. In an unstable world, it becomes insurance.

We have entered an unstable world. Tariffs announced overnight, regulatory reversals from one administration to the next, pressure on the independence of institutions: every shock elsewhere raises the value of a space where the rule is written, voted, dated, and applied by independent judges.

Markets expressed it in their own way. In spring 2025, after the US tariff announcements, capital moved toward euro-denominated assets, and the euro appreciated markedly. On 7 October 2025, in Paris, Christine Lagarde drew the lesson of that episode: Europe's democratic checks and balances and the independence of its central bank are powerful anchors of stability, but Europe cannot remain a merely passive safe haven. I share both halves of that diagnosis. The stability of European rules attracts. It is not enough to build.

Artificial intelligence offers a direct illustration. The European AI regulation sets a calendar: obligations for high-risk uses apply from 2 December 2027, and from 2 August 2028 for systems embedded in regulated products. These dates result from a postponement voted this summer, and I do not hide it: the rule moved. But it moved once, through a published text, with new and firm dates. One can debate its content, and I have. It is harder to dispute its clarity. A company knows what is expected of it and by when. In the United States, by contrast, the 2023 federal executive order on AI was revoked in January 2025, and the federal government now challenges the laws that states pass. For an investor, the second situation is not freer. It is more uncertain.

What the critics say, and what to grant them

A thesis that does not look at its objections does not deserve a board's attention. Here are the most serious ones.

The Draghi report. In September 2024, Mario Draghi delivered to the Commission a severe diagnosis of European competitiveness. It counts around a hundred texts affecting the digital economy and more than 270 digital regulators across the member states. More than 60% of European companies see regulation as an obstacle to investment. Of the 147 unicorns founded in Europe between 2008 and 2021, 40 moved their headquarters abroad. The finding is fair. But it targets layering and fragmentation first, twenty-seven markets and twenty-seven interpretations, far more than the principle of the rule. The answer it calls for is a single, simple rule applied the same way everywhere. Not the absence of rules.

The cost for young companies. A study published in 2021 in Marketing Science measured, in the year after the data protection regulation came into force, a 26% drop in the monthly number of venture deals in young European technology companies, relative to US ones. It must be acknowledged plainly: a rule costs a young company proportionally more than a large group. That is a reason to simplify it for small firms, and to give them tools that lower the cost. It is not a reason to give it up.

Supervisory failures. Europe had Wirecard in 2020: EUR 1.9 billion of cash that did not exist, despite public warnings the German regulator had not followed up. And it was researchers, then the Californian and US regulators, who exposed the diesel emissions fraud in 2015, not Europe. Both cases teach the same thing: a text protects only if it is enforced. A declared rule is not enough. It takes a measured rule.

Europe is stepping back itself. The simplification texts adopted in 2025 and 2026, and the postponement of some obligations of the AI regulation, show that Brussels acknowledges excesses. I do not see that as a disavowal. Adjusting a setting without changing course is the mark of a stable system. Instability is not adjusting a rule: it is never knowing whether it will still exist next year.

What an executive can do with this now

If the European rule is an asset, it still has to be used as one. Three decisions are within reach of any executive committee.

  • Make compliance a valuation argument, not a cost line. In front of an acquirer, a lender or a partner outside Europe, present what the rule makes you document as a guarantee, just like audited accounts.
  • Choose Europe as a trusted jurisdiction. For a contract, a fund, data hosting or an AI system, European jurisdiction is not a constraint to endure: it is an argument to sell to your customers and investors.
  • Prove rather than declare. Europe's failures all share one cause: a rule written but poorly enforced. Companies that measure for themselves, continuously, what the rule asks of them have nothing to fear from an inspection. They have everything to gain.

The danger, the advice

The dangerThe advice
Compliance is filed as a cost line: it never shows up in a valuation, a fundraising or a negotiation.Present what the rule makes you document as a guarantee, just like audited accounts.
The rule is declared but not enforced: evidence is rebuilt on the day of an audit or an incident.Measure continuously what the rule asks of you and keep dated evidence as you go.
Your contracts, data and AI systems depend on jurisdictions whose rules change from one year to the next.Make European jurisdiction an explicit choice and an argument with customers and investors.

My conviction

Europe will not win the race for speed. It has lost it, and nostalgia will not change that. But another race is starting, the race for trust, and it is won with assets Europe took seventy-five years to build: written rules, independent institutions, a market large enough to impose its standards.

In an unstable world, stability becomes the scarce resource. Whoever holds it has no need to apologise for it. They have to learn to sell it.

Pascal Mennesson, founder of Maltem Consulting Group, founder of 56North. If you lead a company and would like to discuss this, request a first conversation.

Questions and answers

Does European regulation hold back innovation?

It has a cost, especially for young companies, and the Draghi report points to layered texts. But a stable rule reduces uncertainty, which investors charge for as a risk premium. The right answer is a simpler, better enforced rule, not no rule.

What is the Brussels effect?

It is the name the legal scholar Anu Bradford gave to the way Europe, through the size of its market and the demands of its rules, sets standards that companies worldwide end up applying everywhere.

When do AI Act obligations for high-risk systems apply?

Since the 2026 postponement (regulation 2026/1744), on 2 December 2027 for high-risk uses, and on 2 August 2028 for systems embedded in regulated products.

Sources

Every source was opened and dated before publication.

  1. Commission européenne, « La déclaration Schuman, mai 1950 »
  2. Walter Eucken Institut, « Walter Eucken and the Freiburg School of Ordoliberalism »
  3. EFAMA, « Latest research shows success of UCITS outside Europe » (23 septembre 2025)
  4. Anu Bradford, The Brussels Effect: How the European Union Rules the World, Oxford University Press (2020)
  5. CNBC, « SVB customers tried to withdraw nearly all the bank's deposits over two days, Fed's Barr testifies » (28 mars 2023)
  6. Réserve fédérale, communiqué du 28 avril 2023 sur le rapport de Michael Barr
  7. Gibson Dunn, « Agencies Release Reports Detailing Supervision and Regulation of Silicon Valley Bank and Signature Bank » (mai 2023)
  8. Règlement (UE) n° 575/2013 (CRR)
  9. TRM Labs, « What the First Quarter of MiCA Tells Us About CASP Authorizations » (2025)
  10. EASA, « EASA declares Boeing 737 MAX safe to return to service in Europe » (27 janvier 2021)
  11. BCE, Christine Lagarde, discours du 7 octobre 2025
  12. Règlement (UE) 2024/1689 sur l'intelligence artificielle
  13. Règlement (UE) 2026/1744 (omnibus numérique)
  14. Paul Hastings, « President Trump Signs Executive Order Challenging State AI Laws » (décembre 2025)
  15. Tech Policy Press, « Draghi's European Competitiveness Report: Key Findings » (septembre 2024)
  16. INFORMS, « GDPR Has Dampened Investment in European Tech Ventures, Study Finds » (Jia, Jin et Wagman, Marketing Science, 2021)
  17. CNBC, « German payments company Wirecard files for insolvency » (25 juin 2020)
  18. EPA, « Learn About Volkswagen Violations » (avis du 18 septembre 2015)

Do you know which AI systems run in your company, and who answers for them?

Request an assessment

A free 30-minute first conversation, no commitment.