Halfway through 2026, the story of AI mergers and acquisitions has split cleanly in two. Overall corporate dealmaking has become more selective, with buyers running fewer processes and digging deeper before committing capital. Yet within that more cautious market, AI-driven acquisitions have kept moving at a pace and at valuations that outstrip nearly everything else on the board. The result is a bifurcated landscape where discipline reigns almost everywhere except where artificial intelligence is involved, and where the definition of what actually counts as a compelling AI target has narrowed considerably since the freewheeling years of 2022 through 2025.
THE HEADLINE NUMBERS ARE STRONG, BUT THE CONFIDENCE UNDERNEATH IS THIN
Global M&A value reached roughly $1.6 trillion in the first half of 2026, up about 28% from the same period a year earlier, according to Boston Consulting Group’s Jens Kengelbach, Daniel Friedman, and Dominik Degen. That marked the strongest first-half total since the 2021–2022 boom, and megadeal activity, transactions of $10 billion or more, hit 31 in the first six months of the year, nearly double the 17 recorded over the same stretch in 2025. Yet BCG’s own M&A Sentiment Index, which blends fundamentals like valuation levels and business confidence with AI-driven analysis of corporate communications, sits at 84, still below its long-term average of 100. Volume outside the largest transactions remains subdued, and activity is heavily concentrated in a handful of sectors and regions, mostly North America.
Technology, media, and telecommunications generated more first-half deal value than any other sector tracked by BCG, and simultaneously posted the weakest sentiment reading of any sector, at 52. That gap is not a contradiction so much as a description of where the market actually stands: buyers are pouring capital into AI-adjacent assets while remaining deeply skeptical of the broader software business models that AI is disrupting. The firm describes a sharp bifurcation within the AI value chain itself, in which infrastructure-layer assets, data centers, computing capacity, and power, command extraordinary valuations tied to visible, contracted demand, while application-layer AI companies are seeing their valuations correct downward.
WHERE THE MONEY IS ACTUALLY GOING
That infrastructure hunger shows up clearly in the deal data. NextEra Energy’s agreement to acquire Dominion Energyfor approximately $67 billion, the largest transaction tracked globally in May 2026 according to Intellizence, was framed explicitly around power. Dominion’s exposure to Virginia and other data-center-heavy regions was cited as a key piece of strategic value, since AI infrastructure buildouts continue to strain electricity supply. BCG’s sector analysis echoes the same theme industry-wide: hyperscalers are increasingly moving from simply contracting for power to owning the generating capacity behind it, and nuclear assets, including small modular reactors, have moved from a niche interest to a central subtheme in energy M&A.
Direct AI-company acquisitions are also active, if smaller in absolute dollar terms than the infrastructure megadeals. OpenAI acquired Ona, a company focused on cloud-based infrastructure for AI agents, in a deal valued at roughly $2.5 billion, according to data compiled by NewsCatcher’s CatchAll tracker. TDK agreed to acquire Fabric8Labs, a maker of AI data center cooling components, for around $400 million, another sign that thermal management and physical infrastructure supporting AI compute have become acquisition targets in their own right. Separately, Apollo Global Management and Blackstone finalized a $35 billion financing package to expand Anthropic’s AI infrastructure, illustrating how capital is flowing into AI buildouts through structures well beyond traditional acquisitions, including large-scale credit and equity financing arrangements.
FINANCIAL SERVICES, HEALTH CARE, AND ENERGY ARE THE CONVICTION SECTORS
Outside of technology itself, BCG identifies financial services as the sector with the clearest, least conflicted case for dealmaking, posting the highest sentiment score of any industry at 108. Incumbent banks and insurers are acquiring the scale, technology, and AI capability they cannot build quickly enough internally. In North America, fintech M&A has rebounded following an earlier valuation reset, and asset and wealth managers are racing to build out private-market capabilities. European banking consolidation looks increasingly likely on economic grounds, even though lengthy and politically sensitive regulatory approval processes remain the main obstacle to closing deals rather than any lack of strategic logic.
Health care sentiment has shifted from cautious to constructive, landing at 92, among the fastest-improving readings in BCG’s index. Large pharmaceutical companies face a coming wave of patent expirations and are turning to bolt-on acquisitions of biotech firms with validated platforms rather than pursuing transformational mergers. Medical technology and AI-enabled diagnostics form a second pillar of activity as care increasingly moves outside traditional hospital settings. In May, Intellizence tracked several transactions reflecting this same push toward technology-enabled health platforms as part of the month’s broader mega-deal activity.
Energy and utilities, at a sentiment score of 90, has undergone a thematic shift of its own. The renewables and transition narrative that dominated 2023 through 2025 has given way to a more urgent one centered on power availability as the binding constraint on AI-driven growth. Oil and gas majors retain strong balance sheets and are continuing to make acquisitions even as they shed other assets, and private capital has become an increasingly active buyer across the power value chain.
WHAT STRATEGIC BUYERS ARE ACTUALLY LOOKING FOR IN AI TARGETS
Beneath the macro numbers, the criteria strategic acquirers apply to AI companies have tightened considerably, according to Telegraph Hill Advisors. Buyers are no longer paying premiums simply for models, which the firm describes as increasingly commoditized, but for the proprietary data, pipelines, and workflows that make an AI product defensible over time. Vertical depth has also become more valuable than horizontal breadth, since large strategics already have generic infrastructure but lack the trust, regulatory familiarity, and workflow embeddedness that comes from years of focused execution in a single industry. Aptean’s acquisition of supply-chain orchestration company OpsVeda was cited as a clean example of a buyer paying for narrow, proven depth rather than broad ambition.
Agentic capability has become a similarly sharp filter. The market has largely stopped rewarding polished demonstrations of autonomous AI behavior and now demands evidence that agents are deployed in production, generating measurable revenue and outcomes for paying customers. Net revenue retention above 120% has emerged as one of the clearest underwriting signals in AI-specific deals, since it indicates both that customers are expanding usage over time and that the AI capability is compounding value without a proportional increase in sales effort.
Telegraph Hill Advisors is equally direct about what is not moving valuations in 2026. Adding “AI-powered” language to a product description without underlying depth is treated as essentially worthless once technical diligence teams look under the hood. Impressive prototypes without production deployments or paying customers are valued as research projects rather than acquisition targets. Fast growth paired with weak retention is treated as a warning sign rather than a premium, and heavy founder dependency on the underlying AI system is priced as a key-man risk that buyers discount heavily.
PRIVATE EQUITY IS PLAYING A DIFFERENT GAME
Private equity firms are pursuing a distinct strategy from strategic acquirers, according to Telegraph Hill Advisors, building platforms by combining multiple vertical AI companies under a single umbrella rather than making standalone acquisitions. The thesis rests on shared infrastructure, cross-selling across overlapping customer bases, and scale that no individual company could reach alone. BCG’s data supports the broader trend: private equity and venture capital firms are sitting on nearly $2 trillion in undeployed capital, and mounting pressure from limited partners for distributions is pushing sponsors to deploy that capital more actively across technology-enabled services, health care, and financial services, often through add-on acquisitions, take-private deals, and continuation vehicles.
WHAT COMES NEXT
BCG points to a coming wave of large technology IPOs, following SpaceX’s recent listing and with OpenAI and Anthropic expected to follow, as a potential catalyst that could reset risk appetite across the broader deal market. A successful public listing at scale would validate private valuations, reopen a major exit channel for financial sponsors, and signal to corporate boards that the window for ambitious strategic action remains open. Until that test plays out, the pattern likely to define the second half of 2026 is the one already visible in the first: infrastructure and power assets commanding premium prices on the strength of visible AI demand, application-layer AI companies facing a harder valuation reset, and the buyers who move with a clear thesis and disciplined diligence taking the advantage over those chasing narrative alone.
References and Further Reading
