Q2 2026 M&A: record value, the thinnest deal count since 2010

By the CogniSuite team

Before any of the numbers below, one caveat that matters more than usual. The quarter ended on 30 June 2026, and every tracker cited here published a first cut between 1 and 14 July. Those first cuts revise upward as late-reported and undisclosed transactions are captured, and at least one publisher states outright that its figures already include estimates for late reporters. Nothing here is settled. Where providers disagree, we give both numbers and name both providers rather than picking the one that reads better.

There is no single headline number

Three credible totals exist for first-half 2026 global M&A value, and they are not reconcilable.

Reuters, reporting LSEG data on 1 July, put announced H1 2026 volume at US$2.8 trillion, up 48% year on year and the highest first half since LSEG records began in 1980. Two days later, Investment Executive, working from a slightly later LSEG cut, reported US$2.85 trillion and 50% growth. Same provider, same half, two different totals, separated by a cut-off date.

The wider gap is methodological. Mergermarket, in its 8 July league table release, tallied US$3.16 trillion for H1 2026, up 44% year on year. Mergermarket counts assumed debt and includes transactions LSEG excludes, so the roughly US$300 billion spread against LSEG is definitional rather than an error by either side.

The same divergence shows up at the level of a single transaction. The quarter's largest announcement, NextEra Energy's combination with Dominion Energy, carries a value of US$66.8 billion on LSEG's equity basis and US$118.8 billion on Mergermarket's. Megadeal counts differ too: LSEG counted 47 deals above US$10 billion in the half, Mergermarket counted 48.

The practical rule is to cite one provider per number and name it, which is what we do below.

Value rose. Deal count fell to a level not seen since 2010

On LSEG data via Investment Executive, Q2 2026 alone accounted for roughly US$1.6 trillion, up 31% on Q1 2026 and a record quarterly total on that basis. We were not able to verify a Q2-only year-on-year comparison from any provider; every growth figure available at this stage compares half to half, so treat quarter-on-quarter as the only clean sequential read.

Deal count moved the other way, and hard. The same Reuters cut of LSEG data counted roughly 24,000 announced transactions in H1 2026, down 9% year on year and a six-year low. A&O Shearman's Global M&A Insights for H1 2026 is blunter still: 10,309 transactions in Q2 2026, the lowest quarterly count since 2010, with global first-half volumes down 14% against the second half of 2025.

That divergence is the defining feature of the quarter: more capital moving through fewer processes.

Concentration, not breadth

Nearly all of the value growth sits in a small number of very large transactions. Per Reuters on LSEG data, 47 deals above US$10 billion totalled more than US$1.3 trillion, close to half of all global deal value and the highest concentration on record.

Bain & Company's midyear report, published 29 June, adds structure to that picture, though its data runs only through May and therefore excludes June. Bain counts megadeal volume up 52% and megadeal value up 53% year on year, with overall deal count up just 2%. Consideration mix shifted: 35% of megadeals used stock plus cash, which Bain describes as a historical high, while all-cash fell to 55%, a cyclical low. Median EV/EBITDA was flat at 11.6 times.

Concentration shows in adviser rankings too. On the legal side, Mergermarket put Kirkland & Ellis at US$600.5 billion across 392 deals, up 94.6% year on year for a 20.6% share, and Sullivan & Cromwell at US$433.1 billion across only 99 deals, up 126.7%. Latham & Watkins recorded US$431.8 billion across 400 deals, up 15.4%. On the banking side, Investment Executive's LSEG-based table shows Goldman Sachs at 38.4% share, Morgan Stanley at 24.2% and JP Morgan at 24.1%, with Lazard jumping to fifth from fourteenth a year earlier.

Sectors and regions, on a half-year basis

No provider we could verify published a Q2-only sector split, so the following are H1 2026 figures.

Technology led by value at US$649 billion, 24% of global deal value and up 90% year on year on LSEG data via Investment Executive. Industrials activity rose 57% and energy and power rose 41% on the same basis.

Regionally, the same LSEG cut via Investment Executive put the United States at US$1.5 trillion, 54% of global value against 45% a year earlier, up 80%. Europe more than doubled to US$676 billion. Asia Pacific fell 2% to US$360 billion. Cross-border activity had its strongest start since 2018 at US$893 billion, up 62%, with the United States the top target country at 25% of cross-border deals.

The record was corporate, not sponsor-led

Financial sponsors did not participate in the headline. Mergermarket put sponsor-related volume at US$745 billion for H1 2026, up only 2% year on year and down 34% from US$1.13 trillion in H2 2025, with strategic acquirers accounting for 78% of volume.

In the United States, PitchBook's Q2 2026 US PE Breakdown, published 6 July, records US$177.3 billion of private equity deal value in Q2, down 37.5% quarter on quarter and 23.9% year on year, against 2,384 announced or estimated transactions that were roughly flat sequentially and up 11.5% year on year. First-half value came to US$461 billion, down 10.6% against H1 2025. Sponsors kept transacting. They transacted smaller.

That pattern holds globally. Aranca's Global Private Equity Factbook for Q2 2026 counts 1,978 global PE deals in the quarter against 2,057 in Q1, with average deal size falling to US$318 million from US$349 million and capital invested down 13% sequentially.

Exits, the binding constraint on the asset class, got worse. PitchBook puts US PE exit value at US$102.6 billion, down 46.3% quarter on quarter. Corporate asset sales fell 63% to US$38.5 billion across 107 deals, and sponsor-to-sponsor exits fell 57% to US$24.5 billion across 94 deals, the lowest quarterly count in at least a decade. Public listings were the exception: 12 listings at US$27.6 billion post-money, double the Q1 count, up 42.2% sequentially and roughly 31% of all exit value.

By sector, PitchBook records US software private equity value at US$10.7 billion in Q2, down 65.7% year on year and 90.3% below its Q3 2025 peak, which it attributes to a reassessment of software business models under AI pressure. Energy ran the other way, with year-to-date value up 80.5% against H1 2025 on datacenter power demand.

Financing explains the split

Corporates could pay and sponsors could not, and the reason is visible in the rate and credit data.

The Federal Reserve held the target range at 3.50% to 3.75% on 17 June 2026 by a 12-0 vote, noting that inflation "remains elevated relative to the Committee's 2 percent goal". PitchBook's read of the June projections has the median year-end 2026 fed funds path rising to 3.8%, with roughly half the committee penciling in at least one increase. The European Central Bank went further, raising all three key rates by 25 basis points on 11 June 2026, its first increase since 2023, taking the deposit facility to 2.25% effective 17 June against projected 2026 headline inflation of 3.0%.

For completeness, and after quarter-end, the Fed held again on 29 July 2026 by 9-3, with three members preferring a quarter-point increase. That vote shapes second-half conditions; it had no bearing on Q2 deal-making.

Investment-grade issuance stayed abundant. Global investment-grade corporate debt reached US$3.4 trillion in H1 2026, up 10% and the highest first half on LSEG's record. Leveraged conditions tightened at the same time. PitchBook LCD records the broadly syndicated loan market down 7% sequentially, private equity deal volume within it down 38%, and sponsored M&A loan volume down 54% quarter on quarter. Sponsor-backed borrowers made up just 45% of non-refinancing issuance against a five-year average near 70%, recent LBO spreads averaged S+509 against S+474 in Q1, and private credit spreads widened by roughly 25 to 50 basis points.

An investment-grade acquirer paying partly in its own stock had a market. A sponsor underwriting a leveraged buyout to a falling rate path did not.

The transactions that defined the quarter

NextEra Energy and Dominion Energy. Announced 18 May 2026 as an all-stock combination at a fixed exchange ratio of 0.8138 NextEra shares per Dominion share plus a one-time cash payment of US$360 million, leaving NextEra holders with 74.5% and Dominion holders with 25.5%. Closing is expected in 12 to 18 months, subject to HSR clearance, FERC Section 203 approval, NRC approval and reviews by the Virginia State Corporation Commission, the North Carolina Utilities Commission and the South Carolina Public Service Commission, alongside shareholder votes.

SpaceX and Anysphere. A US$60 billion all-stock acquisition announced 16 June 2026, described by Forbes as the largest acquisition of a venture-backed startup on record, with closing expected in Q3 2026.

Equity Residential and AvalonBay. An all-stock merger of equals announced 21 May 2026 at 2.793 Equity Residential shares per AvalonBay share, producing roughly US$52 billion of pro forma equity market capitalisation and about US$69 billion of enterprise value across more than 180,000 apartment homes, with AvalonBay holders at roughly 51.2% and US$175 million of gross synergies targeted.

KONE and TK Elevator. Announced 29 April 2026 at an enterprise value of EUR 29.4 billion, combining roughly EUR 20.5 billion of annual sales with about EUR 700 million of run-rate synergies. Notably it is a large sponsor exit, out of a consortium, in a quarter that produced very few. Closing is not expected before Q2 2027.

One clarification on attribution. Warner Bros. Discovery stockholders approved the Paramount Skydance transaction on 23 April 2026, but the deal was announced well before the quarter, so only the vote belongs to Q2. Its financing was large enough to move the whole market: PitchBook LCD notes the US$13 billion term loan makes Warner Bros. Discovery the largest single borrower in the US leveraged loan market.

Merger control moved during the quarter, and again just after it

The European Commission published draft revised Merger Guidelines on 30 April 2026, consolidating the 2004 horizontal and 2008 non-horizontal guidelines into a single instrument. The public consultation closed on 26 June 2026 and the review is expected to conclude in Q4 2026, which leaves European filers working against guidance that is known to be changing.

In the United States, and after quarter-end, the DOJ Antitrust Division revived an expedited Second Request track. Per Davis Polk's client update, the process announced on 23 July 2026 commits the Division to a Front Office meeting within 21 days of a merging party's priority production and a decision within 14 days of that meeting. It is a second-half development, but it rewards parties who can assemble a complete, well-organised production quickly.

What the quarter meant for diligence practice

Four structural shifts follow from the data above, and they change how a diligence process should be run rather than merely how it feels.

Fewer processes, each far longer. Bain puts announcement-to-close for deals above US$10 billion at roughly seven months, with 24 to 36 months to realise the majority of cost synergies and more than 36 months for full integration. Regulated transactions run longer still: NextEra and Dominion are guiding to 12 to 18 months across six separate regulatory tracks plus shareholder votes, and KONE and TK Elevator are not expecting to close before Q2 2027. A data room opened in this quarter may be live through several changes of reviewer population, several waves of regulatory production, and staff turnover on every side. Access design that was correct at signing is rarely still correct nine months later, which argues for permissions that resolve per folder and inherit predictably, and for an access record that can be produced later without reconstruction.

Paper consideration makes diligence reciprocal. Bain's finding that stock-plus-cash megadeals hit a historical high while all-cash fell to a cyclical low is not a financing footnote. When target holders are paid in shares, they and their advisers must diligence the acquirer. Three of the quarter's largest announcements were all-stock, and one was an explicit merger of equals. That means two rooms, two disclosure postures, and a genuine need for tooling that understands which side of the table a participant sits on rather than assuming a single seller disclosing to many bidders.

Cross-border and regulatory breadth expand the reviewer population. Cross-border value up 62% and a European guidelines rewrite in flight both push in the same direction: more counsel, more jurisdictions, more people who need a narrow slice of a room and nothing else. That is a granularity problem before it is a security problem. It also raises the cost of the blunt instrument, which is to grant broad read access and hope. Watermarking helps, but only if it holds across the file types diligence actually uses, spreadsheets and presentations included, and only if a file that cannot be marked is refused rather than served clean.

AI exposure became a diligence question in its own right. In PitchBook's Q2 2026 US PE survey, taken as of 8 June 2026, respondents ranked interest rates at 59%, geopolitical risk at 59% and AI disruption or opportunity at 58% as the top macro factors, with tariffs last at 33%. Software was the sector most cited as one respondents had pulled back from. Disagreement about a target's AI exposure is now a distinct source of the bid-ask spread, and it lands in diligence as a demand for evidence: contract terms, customer concentration, data rights, model dependencies, and what the company's own product roadmap actually assumes.

On how diligence teams are using AI, the honest position is that we have no Q2 2026 measurement. The two useful studies both predate or straddle the quarter. Research conducted by FT Longitude across 1,000 senior dealmakers in 27 countries, with fieldwork in March 2026, found 50% reporting regular AI use during due diligence, the stage with the highest reported return, with 66% saying it helps de-risk transactions, 71% ranking accuracy and 70% ranking security as the top requirements, and 58% relying on human review to validate output. Separately, SRS Acquiom's diligence study with Mergermarket, published in Q4 2025 from a sample of 150 senior US investment bank executives, found 73% expecting diligence to become more complex over the following 12 to 24 months, one in five saying timelines had already lengthened, and 57% of that group reporting one to three additional months. Both are expectations surveys, not measurements of the quarter that just closed. They should be read as direction, not as evidence.

One last point, and it is the same problem this commentary ran into. This article had to publish three different totals for a single half-year because three reputable providers measured the same market differently. A data room presents the identical problem in miniature: a management presentation, an audited statement and a lender model routinely disagree, and the disagreement is usually definitional rather than dishonest. Our own AI is instructed to surface that disagreement and to attach a source and a date to every figure it reports, and drafted answers are checked server-side against the quoted source text before a person sees them. What we do not have is a reconciliation engine. Nothing extracts figures across an entire room, normalises units and flags contradictions on its own, so the surfacing depends on both conflicting documents landing in the same retrieval set. Reconciliation across a full room is still human work, and it is worth asking any vendor that claims otherwise to demonstrate it on your own documents.

Two related limits are worth stating plainly because they bear on how much weight a diligence record can carry. Audit events do capture a source IP, but the trail is not cryptographically tamper-evident, and it is readable by the advising firm's team rather than by counterparties, which makes it an oversight tool rather than a mutual evidentiary record. Single sign-on is available as OpenID Connect with authorization code and PKCE and a per-firm email-domain allowlist. It is opt-in per workspace and it sits alongside the emailed one-time code rather than replacing it. Our current posture on all of this is documented at /security.