EU AI Act · Regulation (EU) 2024/1689 · M&A

The AI Act value gap in M&A transactions

How (non-)compliance with Regulation (EU) 2024/1689 became a recurring repricing factor for AI assets — and what the leading firms already say about it.

David Roux · Founder, MB AI Value Intelligence, SKEMA Business School · 10 June 2026

Executive summary

AI regulatory risk attaches to the asset, not to the transaction: the acquirer inherits it from closing. In 2026 this risk became a standalone due-diligence workstream that translates into a discount, an escrow, or a walked-away deal. The figures and positions below — drawn from first-rank sources — document it. The decisive variable is not the risk itself but who prices it, and when: the acquirer in the deal room, or the asset holder, upstream.

1. The market reality

AI risk has moved out of technical meetings to become a diligence workstream in its own right, with its own price impact.

Bain & Company — Global M&A Report 2026

« One in five strategic dealmakers tell us that they have walked away from a deal because of the anticipated impact of AI on the target's business. » AI adoption in M&A processes also doubled, to 45% of practitioners.

In other words: one in five strategic acquirers walked away from a deal in 2026 because of AI's anticipated impact on the target. Risk no longer merely worries buyers — it kills deals.

2. The mechanism: risk travels with the asset

Seasoned acquirers know this principle because they have already paid for it in the personal-data arena — the direct antechamber of the AI Act.

Precedent — Verizon / Yahoo (2017)

After massive data breaches surfaced, Verizon cut its acquisition price by $350M (from $4.83B to $4.48B). The risk wasn't in the contract: it was in the asset.

Precedent — Marriott / Starwood

By acquiring Starwood, Marriott inherited a breach undetected in diligence — and a £18.4M GDPR fine. The UK regulator explicitly faulted inadequate verification at purchase. Regulatory liability survives closing.

The EU AI Act sets up this very context for AI assets. The Article 9–15 obligations (risk management, technical documentation, human oversight, robustness) attach to the system; Article 99 penalties reach €35M or 7% of global annual turnover.

3. What the leading firms say

Skadden, Arps — "M&A in the AI Era" (Jan 2026)

« RWI insurers are taking a closer look at AI-specific issues, which could lead to policy exclusions for data provenance, model performance or other AI-related representations. »

a leading audit firm — "AI & M&A due diligence" (Dec 2025)

Adherence to standards such as GDPR and the EU AI Act prevents legal exposure, bias-related issues and reputational risks « that could affect the valuation ».

a leading audit firm — "Impact of the EU AI Act in private equity" (Sept 2024)

"The EU AI Act could potentially alter deal valuations. Companies that have proactively addressed compliance may command a premium, while those lagging behind could face valuation discounts or require additional investments… The Act may also shape deal structures, with indemnities and warranties becoming more nuanced." — Melchior Ballreich, a leading audit firm Strategy & Transactions

a leading audit firm (same) — premium at exit

"A portfolio company that showcases its adherence to the EU AI Act is effectively future-proofing its business operations… This forward-thinking approach can lead to a premium at exit, since buyers are often willing to pay a good price for a business that has strategically mitigated regulatory risks." — Dr. Dierk Buß, a leading audit firm

a leading audit firm — 2026 Global M&A Outlook (March 2026)

"Dealmakers are increasingly incorporating AI exposure assessment into standard diligence practice, recognizing that pricing confidence now depends not only on financial performance, but on structural defensibility." In early 2026, nearly $1 trillion in software-services market value was repriced in a matter of weeks — not from collapsing revenues, but from multiple compression.

a leading audit firm — How AI is reshaping software valuations in M&A (Feb 2026)

"Trust and auditability are a defensible moat in regulated environments." Assets with embedded compliance logic, in regulated sectors, prove the most resilient: compliance isn't a cost, it's a defensible barrier.

Fieldfisher — "AI in M&A: Assessing transaction risks" (Dec 2024)

"The DD review of governance, internal risk management compliance and record keeping and the impact on valuation for AI systems has now become an important part of any transaction." Buyers now require AI-specific representations & warranties — "as 'fundamental' warranties… with longer survival periods and higher caps" — backed by price holdbacks. AI Act fine cited: €35M or 7% of worldwide turnover.

Brown Rudnick — "EU AI Act: Impact on Corporate Finance, VC, Exits" (Jan 2024)

« Compliance with the act may become the golden ticket to a successful exit… Noncompliance could act as a market access barrier, confining companies to domestic markets or niche buyers. »

Reed Smith — "The AI M&A playbook" (2026)

AI risk is already structured into contracts: dedicated escrows with 18- to 24-month terms, indemnities targeting training-data rights.

CMS — "The impact of the AI Act on Tech M&A due diligence" (Jun 2024)

A target's AI Act classification — unacceptable, high, limited or minimal risk — becomes a due-diligence item in its own right in technology deals.

Aird & Berlis — "Artificial Intelligence, Real Risks" (Mar 2025)

Representations & warranties now cover training-data licences and consents; targeted indemnities address open-source and third-party data.

This is no longer a fringe topic: it is covered across the advisory and corporate-law market. a leading audit firm and a leading audit firm on valuations and private equity, a leading audit firm in its 2026 Global M&A Outlook, Capgemini on compliance, CMS, Aird & Berlis, Fieldfisher and Hunton Andrews Kurth on diligence and clauses — all now document AI Act exposure as a transaction factor. The consensus is in place; what was missing is the instrument to price it per asset.

4. The magnitude, quantified

Measure	Magnitude	Source
AI deals walked away from (anticipated AI risk)	1 in 5	Bain & Company, 2026
Multiple reduction (regulatory + privacy + technical risk)	−15 to 30%	FE International, 2026
Observed discount on a consumer-AI target (despite strong growth)	−25%	FE International, 2026
Repricing for asset-borne data risk (precedent)	−$350M	Verizon / Yahoo, 2017
Maximum AI Act fine (prohibited practice)	€35M / 7%	Regulation 2024/1689, Art. 99

5. Implications — the same gap, two readings

For the acquirer

Unresolved risk becomes a haircut, an escrow or a walk-away. Quantifying exposure before the LOI means negotiating on facts — and drafting R&W, escrow and the regulatory MAC on a numbers basis.

For the asset holder

Handled during the deal, remediation costs 2 to 3 times more — and scares bidders off. Established upstream, it protects price. Arriving in the deal room with an attestation means refusing to let the buyer set the discount for you.

« AI Act compliance is not a cost line. It is an asset-price factor. The only question is who sets it: you, upstream — or the buyer, in the deal room. »

6. The MB AI response

The D7™ score characterises an AI asset's exposure across seven dimensions, with a per-claim evidence level [PROVEN / INFERRED / ABSENT]. It is established (on public data), proven (an opposable attestation, defensible at signing and in W&I underwriting) and defended over time (MB Shield™) — because the framework evolves, and so does the asset. A fraction of the cost of a leading audit firm (€150–500K), to protect 15–30% of the asset's value.

Methodological note. The first-rank sources above document that AI risk — of which the AI Act is an explicit and growing component — reprices transactions. The discount figures (15-30%) cover the whole of AI risk (regulatory, data, technical); they are not attributed to the AI Act alone. The rigorous wording used in this note is therefore "recurring repricing factor", not a quantified discount imputable to Regulation 2024/1689 alone.

Quantify the gap on a real asset

D7™ score · established then defended by MB Shield™ · scope: EU AI Act + GDPR

AI Act M&A due diligence →

Sources: Bain & Company, Global M&A Report 2026 · Skadden, M&A in the AI Era · a leading audit firm, AI & M&A due diligence · Brown Rudnick, EU AI Act & exits · Reed Smith, The AI M&A playbook · FE International, AI Business Valuation Model 2026 · a leading audit firm, Impact of the EU AI Act in private equity (2024) · a leading audit firm, AI software valuations in M&A · a leading audit firm, 2026 Global M&A Outlook · Capgemini Invent, EU AI Act Compliance · CMS, AI Act & Tech M&A due diligence (2024) · Aird & Berlis, Artificial Intelligence, Real Risks (2025) · Fieldfisher, AI in M&A: Assessing transaction risks · Hunton Andrews Kurth, GDPR, AI & Cybersecurity in M&A · CNN/TechCrunch, Verizon–Yahoo (2017) · Cybersecurity Dive, Marriott/GDPR · Regulation (EU) 2024/1689, Art. 99.