AI product management · Industry brief

Global M&A through July 2026 reached US$3.19 trillion, up 36% year-over-year, but deal count fell 10%, concentrating value in mega-deals. According to LSEG and PwC data cited in the article, deals above US$10 billion represented roughly 40% of activity and deals above US$5 billion now account for 48% of global deal value, up from 39% in 2025. The AI deal market has split into two distinct segments: infrastructure-layer assets like data centers and power are commanding extraordinary valuations based on contracted demand, while application-layer company valuations are correcting, per BCG analysis. For product management teams, this bifurcation means infrastructure deals and application deals now require different investment theses and underwriting approaches. Cross-border M&A reached US$1.05 trillion through July, the highest January-to-July total since 2007 per LSEG, introducing additional regulatory and jurisdictional complexity for corporate development teams managing international transactions.

Grant Thornton's $5 billion acquisition of CBIZ, the accounting sector's largest deal in 25 years, is driven by AI capability expansion and competitive pressures to modernize technology infrastructure. The merger is expected to close in Q4 and would create the nation's fifth-largest accounting firm, with Grant Thornton having committed $1 billion over three years to AI strategy, including specialized advisory practices and AI-integrated audit and tax platforms. CEO Jim Peko highlighted that demand for AI governance advisory services is growing as companies rely more on automated tools, and predicted regulators providing guidance will create additional revenue streams, while noting that larger competitors like BDO USA and RSM US LLP are making similar AI investments alongside Big Four firms pouring billions into their AI infrastructures.

The website content contains three articles relevant to your search intents: The first article discusses ransomware attacks targeting AI infrastructure, specifically the Jadepuffer attack that exploited vulnerabilities in LangFlow, a popular AI development framework. It details how attackers destroyed trained AI models and provides practical security recommendations for protecting AI development tools, including inventory management, third-party risk assessment, and strengthening security controls for AI development processes. The second article covers the EU AI Omnibus regulation (Regulation EU 2026/1744), which amends the EU AI Act with extended compliance deadlines for high-risk AI systems, new prohibitions on nonconsensual intimate imagery, expanded bias testing provisions, and reduced administrative burdens for small and medium-sized enterprises, effective July 27, 2026. The third article addresses compliance representations and warranties in M&A and third-party contracts, noting that compliance language increasingly must address AI-related risks, including requirements that companies comply with laws like the EU AI Act and that training data was collected lawfully.