If you just read the surface headlines, the merger and acquisition (M&A) market appears to be recovered. According to BCG, deal value in the first half of 2026 came in around $1.6 trillion, up roughly 28% year over year and the strongest first half since the 2021 boom. Bain, measuring a slightly different window, put the first five months at $2.4 trillion, up 41%, and PwC has full-year value on track for about $4 trillion.
But these surface headlines do not describe the market most founders and middle-market companies are experiencing. When you look deeper, PwC found nearly half of global deal value now comes from transactions above $5 billion. If you remove those megadeals, then aggregate value is down. So, we do not have one M&A market in recovery. We instead have two markets moving at very different speeds.
AI is driving the market, but not evenly
AI is the clearest driver of large strategic transactions. Of the 100 largest corporate deals of 2025, PwC found roughly a third cited AI in the strategic rationale. When you narrow that to just deals in the technology sector, nearly all the largest deals did.
But even when it comes to AI-driven deals, the money isn’t necessarily flowing evenly. BCG describes a split between the infrastructure layer and the application layer. Data centers, compute, and power command for extraordinary valuations because the demand is visible and under contract. Application-layer companies are seeing valuations correctly. Traditional software is under structural pressure, which is why BCG’s sentiment index has technology at 52, the lowest of any sector, in the same period technology led all sectors in deal value.
That gap between the money going in and the confidence behind it is the most interesting aspect of the market right now. Buyers are committing capital at pace while openly questioning the valuations they’re paying.
The second-order effect is that AI is pulling M&A into sectors that have nothing to do with software. Power availability has become the binding constraint on AI growth, so hyperscalers have moved from contracting for power to buying the generation behind it. Utilities and energy deal value hit a record, and nuclear, including small modular reactors, went from the periphery to the center of the conversation.
The middle market is still waiting
The other M&A market is far less exuberant. For companies valued between roughly $50 million and $500 million, the familiar obstacles remain: valuation gaps, more expensive financing, and cautious buyers. Large corporates and scaled sponsors have the balance sheets, private-credit access, and strategic mandates to act, but smaller buyers are being selective.
Capital allocation is a part of this story. Estimates suggest that building AI infrastructure could require $5 trillion to $8 trillion over five years. Not every dollar committed to a data center would otherwise fund an acquisition, of course, but the direction is clear: the AI buildout is competing for enormous pools of capital.
That does not mean demand for middle-market companies has disappeared. Private equity sponsors still hold substantial dry powder and face growing pressure to return capital to limited partners. Strategic buyers also need capabilities that they cannot develop quickly enough on their own. But neither group is likely to pay a premium simply because the headline numbers say M&A is back.
What founders and boards should do now
For companies in or adjacent to AI infrastructure, this is a seller’s market, but the window may be narrower than it appears. While premiums are available now, valuations can change quickly as soon as capacity catches up or investor’s priorities shift.
Application-layer software companies should assume buyers will test every AI claim. A credible seller needs more than a narrative. It must show, with data, how AI drives revenue, improves margins, deepens customer relationships, or creates a defensible advantage.
Companies outside the AI orbit should not wait for a broad market recovery to lift their valuation. The better course is to prepare their own fundamentals: clean financials, defensible growth, durable customer relationships, a clear competitive position, and a realistic view of value. A well-prepared company can transact in a selective market, but an unprepared one will not be rescued by favorable headlines.
The improving IPO market also matters, but mostly as a signal. Successful high-profile listings can validate private valuations, give sponsors another exit route, and improve risk appetite across the board. But for most founders, the practical path to a full exit still runs through a strategic buyer or a financial sponsor, not a public listing.
What to watch next
The question for the second half of 2026 is whether confidence spreads beyond megadeals and AI infrastructure. If prominent technology listings perform well and financing conditions continue to improve, buyers may become more willing to pursue transactions outside the market’s hottest themes. If not, the divide will continue.
Either way, the strategic case for M&A is getting stronger. AI is reshaping competitive position faster than most companies can build capability internally, and that’s true in industrials, healthcare, and financial services, as well as technology. Boards should be reviewing portfolios, identifying gaps, and deciding where buying, selling, or partnering makes the most sense.
The companies that do well moving forward will be the ones who are prepared. They understand how AI affects their business, know which assets or buyers matter, maintain discipline on valuation, and have a credible plan for integration. While that work has certainly always mattered, in a market split in two, it can determine on which side of the divide a company lands.
Sources: BCG, “Mid-2026 M&A Insights,” July 15th, 2026. PwC, “Global M&A industry trends: 2026 outlook” and 2026 mid-year outlook. Bain & Company M&A Midyear Report, 2026. Figures reflect the reporting windows used by each firm, which differ, so headline totals are not directly comparable across sources.
