Hard assets took 47 percent of the buyout dollar. Computers and electronics fell to 10.

Where did the buyout dollar go in the first half?

To hard industries, by a margin that did not exist eighteen months ago. Carlyle estimates that utilities, energy, chemicals, transportation and machinery, with real estate, construction, oil and gas, metals and aerospace inside its definition, made up 47 percent of global leveraged buyout volume in the first half of 2026, on Dealogic data to July. The same group was 27 percent of volume for all of 2025 and approximately 13 percent a year across 2021 to 2024.

The other side of the ledger is computers and electronics: 10 percent of first-half deal volume, down from a 30 percent share for full-year 2025, and the sector's slowest quarter since the second quarter of 2020. Utility and energy deal value reached 63.4 billion dollars in the first half alone, within 10 percent of 2025's full-year total, which was itself the largest annual figure for the sector since 2007. Machinery was 10 percent of global buyout activity in the second quarter at 12.5 billion dollars, against 2 percent in 2025.

Carlyle reads the rotation as structural rather than cyclical, driven by the demand for power, materials and physical assets that the build-out of computing infrastructure is creating, and says in the same paragraph that its durability remains to be seen. We would add the arithmetic. A group of sectors moving from 13 percent to 47 percent of the buyout dollar, in a half in which US buyout volume fell 45 percent from the first quarter to the second on the same report, is not only new money arriving. It is the same buyers leaving somewhere else. The somewhere else is computers and electronics and, in the credit market that finances buyouts, software.

How much of the record is a handful of transactions?

Most of the headline. Global M&A reached 1.8 trillion dollars in the second quarter, but the top ten transactions accounted for over 30 percent of global deal volume, up from an already concentrated first quarter. One listing, at 86 billion dollars, was 74 percent of the record 117 billion dollars of quarterly IPO proceeds. In buyout fundraising, the top three funds captured 35 percent of all commitments in the quarter.

Carlyle's advice to its readers is to compare deal counts against dollar volumes to see whether activity is broadening or narrowing, to use median rather than average multiples, and to look at manager-level activity outside any single mega-deal. We would apply that advice to Carlyle's own 47 percent. It is a share of volume, and the report's own figures put 63.4 billion dollars of the first half in utility and energy, a sector whose transactions are large by nature. Carlyle does not publish a deal count by sector, so the question a founder in a hard industry needs answered, whether the number of sponsors bidding in the sector broadened or a few very large cheques moved the share, cannot be answered from this report. It can be answered from a buyer list built for the sub-sector, and that is the first thing to ask for.

What happened on the other side of the rotation?

It lost its financing before it lost its buyers. The weighted average bid on software loans fell to approximately 86 cents on the dollar by June 2026 from 92 cents in January, on PitchBook LCD's second-quarter wrap as cited by Carlyle, widening the price gap against non-software loans to roughly 11 points from about 5 at the start of the year. Software's share of new-issue leveraged buyout loan volume fell to 14 percent year to date from 35 percent for full-year 2025.

Carlyle's reading is that this is a valuation story rather than an operational one, that the wider market is open and functioning, and that the stress is sector-specific rather than systemic. Its evidence is the rest of the market: US high yield issuance rose 27 percent from the first quarter, European leveraged loan volume held up, and private credit fundraising reached an estimated record above 100 billion dollars in a single quarter, led by direct lending. For a software founder the distinction between valuation and operations is thin comfort. A buyer that cannot borrow against the company at the old price will not pay the old price, whatever the operating numbers say, and a lender that marks the sector's existing loans at 86 cents is not writing new ones at par. What that reset looks like from the lender's side, and where the exposure is concentrated, is covered in our companion pieces on software loans.

Is the buyer under pressure of its own?

Yes, from behind. After two stronger quarters, buyout LP cash flows turned negative again in the first quarter of 2026, the latest period for which the data exists, with limited partners receiving approximately 1 billion dollars less in distributions than they paid in capital calls. Carlyle puts the cumulative private equity cash-flow deficit since 2022 at nearly 260 billion dollars on MSCI Burgiss data, and at roughly 550 billion dollars across all private market strategies. The sponsor bidding for a company is also a seller whose investors are waiting to be paid.

Two consequences follow for anyone selling to that sponsor. The first is that a buyer which must return capital prefers assets it can see its way out of, which is one more reason the buyout dollar has moved toward sectors where exits are clearing; the exit multiples Carlyle reports improving, covered in our companion piece, are for the deals that did exit, a shrinking count. The second is that the buyer's own investors have started selling: Carlyle records a record first half in the secondary market and a secondary dry powder figure of about 194 billion dollars, less than one year of activity, so the supply of stakes is now arriving faster than the capital raised to absorb it. A founder selling into a hard industry is selling into demand. A founder whose company a sponsor already owns should expect the company to be either the asset the sponsor keeps or the asset it moves into a continuation vehicle, and should ask which.

Which side of the rotation is your company on?

Three questions settle it. First, is the business inside Carlyle's list: utility and energy, real estate, transportation, chemicals, construction and building, machinery, oil and gas, metal and steel, aerospace? If so, the share of the buyout dollar pointed at the sector roughly tripled between 2024 and the first half of 2026, but that is a share of volume, and the useful number is the count of sponsor platforms that closed in the sub-sector in the last four quarters. Second, if the business is in computers, electronics or software, how is any offer being financed? The lender moved before the buyer did, so the structure of a bid, the equity share, any seller note, any earnout, tells more than the headline, and it is the question to ask before price. Third, in either case, what does the buyer's own fund need to do this year? A sponsor with a deficit to its investors and a fund to raise behaves differently from one that has just closed, and both are bidding in the same process.

The rotation's durability remains to be seen, in Carlyle's words twice over. A founder in a hard industry who is thinking of selling inside the next two years has the uncommon condition of a buyer universe that has moved toward the company rather than away from it. Conditions like that are rented rather than owned, and the way to collect the rent is to run the process while the tenant is still there.