AI Due Diligence

State of play.

• AI infrastructure is driving record M&A deal value. Q1 2026 opened as the strongest start for large deals on record, with AI-related transactions a primary engine alongside geopolitical repositioning .
• AI capabilities are now a discrete valuation input in consumer-sector deals. Foley & Lardner's analysis of fashion and beauty M&A identifies AI-powered virtual try-on and demand forecasting as explicit buyer priorities alongside Gen Z relevance and supply chain resilience — meaning AI infrastructure is being priced at the deal level, not treated as a generic operational asset .
• Environmental self-disclosure frameworks are load-bearing in M&A diligence. EPA, DOJ, and state audit policies — anchored by EPA's December 1998 CERCLA guidance — shape how environmental liabilities are valued and allocated in transactions, with Phase I assessments and compliance audits as the standard toolkit .
• For counsel advising acquirers in AI-adjacent or AI-enabled targets, the practical baseline is that AI diligence now spans at least three distinct tracks: infrastructure and capability valuation, environmental and operational compliance tied to data center buildout, and sector-specific digital-asset diligence (IP, data, model ownership) that standard deal frameworks have not yet standardized.

Where things stand.

• Record deal volume is compressing diligence timelines. Q1 2026's record deal value, driven by AI boom and geopolitical repositioning, creates pressure on diligence depth at exactly the moment AI-specific diligence requirements are expanding .
• AI infrastructure is a discrete line item in consumer-sector valuations. The Foley & Lardner fashion and beauty analysis documents buyer premiums for digital infrastructure including AI-powered tools — a signal that AI capability diligence is migrating from tech-sector deals into general consumer M&A .
• Environmental diligence frameworks are well-established but allocation mechanics are unsettled. EPA's CERCLA audit protocol and Phase I assessments are standard; how buyers and sellers are currently structuring disclosure obligations and liability splits in transactions involving AI-driven or data-center-heavy assets remains incompletely documented .
• Gen Z spending power — estimated at $360 billion per the Foley analysis — is shaping premium valuations for culturally resonant, digitally native brands, with AI-enabled demand forecasting and virtual try-on treated as value-additive infrastructure rather than commodity features .

Latest developments.

• Q1 2026 opened as the strongest start for large deals on record, with AI-related transactions identified as a primary driver alongside geopolitical repositioning .
• Foley & Lardner published a 2025-2026 fashion and beauty M&A analysis documenting AI-powered virtual try-on and demand forecasting as explicit buyer priorities, alongside three major closed transactions: Prada/Versace (€1.25B), e.l.f. Beauty/Rhode (up to $1B), and L'Oreal/Creed (€4B) .
• Environmental self-disclosure as a transactional risk management tool is documented as standard M&A practice, with EPA CERCLA audit protocols and Phase I assessments as the baseline framework; specific liability allocation mechanics in current transactions remain incompletely documented .

Active questions and open splits.

• What does AI diligence actually require? No standardized framework exists for evaluating AI capabilities, model ownership, training data provenance, or AI infrastructure in M&A targets — buyers are improvising across deals, creating inconsistent risk allocation.
• How should AI infrastructure be valued in non-tech-sector targets? The fashion and beauty sector is now pricing AI tools as premium assets; whether acquirers are applying consistent valuation methodologies or simply accepting seller representations remains an open question .
• Environmental liability allocation for AI-driven data center assets. Data centers carry significant environmental footprint — power consumption, water use, site contamination risk — and the intersection of EPA CERCLA frameworks with AI infrastructure acquisitions has not been tested in published guidance .
• IP and data ownership in AI-enabled targets. Deals involving AI-powered tools require diligence on training data rights, model licensing, and third-party API dependencies — none of which map cleanly onto standard IP representations and warranties.
• Geopolitical risk as a diligence variable. Record Q1 deal volume is occurring against a backdrop of geopolitical turmoil; how acquirers are diligencing cross-border AI supply chain exposure and export control risk is unsettled .

What to watch.

• Whether deal volume sustains through Q2-Q3 2026 or geopolitical headwinds compress the pipeline — the answer reshapes how much leverage buyers have to demand robust AI diligence.
• Publication of any standardized AI diligence frameworks by bar associations, deal counsel, or institutional buyers that could become market-standard.
• Whether EPA or DOJ issue updated guidance on environmental compliance audits that address AI data center footprints specifically.
• Whether fashion and beauty deal premiums for AI-enabled brands hold through 2026 or compress as AI tooling becomes commoditized .
• Litigation or rep-and-warranty claims arising from AI capability misrepresentations in closed deals — the first such published dispute will set the template for AI diligence standards.