Add-Ons at 76% — When the Modal Acquired Entity Is Also the Exit
When add-on transactions reach 76% of US PE buyouts, the integration choice stops being a sequencing afterthought and starts being an underwriting question asked of every acquisition.
PitchBook’s Q1 2026 US PE Breakdown records a print that crosses a structural threshold the asset class has been approaching for a decade. Add-ons reached 76.3% of all US PE buyouts year-to-date through Q1 2026 — a record, up from 72.7% in 2025 — across 1,422 add-on transactions versus 442 non-add-on buyouts in the partial quarter. The 4:1 ratio is the more telling figure. For most of the prior decade, an add-on was the supporting role inside a platform-led narrative; at four to one, it is no longer the supporting role at all.
The trajectory that produced the 76% print was already visible. Capital pressure of the kind traced in Dry Powder and the Pressure to Move has favoured add-ons for years, because they deploy faster, face less competition than platform deals, and absorb capital without requiring a new operating thesis. What’s new is not the trend; it is the share threshold. When add-ons account for three out of every four buyouts the asset class makes, the question of what an add-on actually is — to the platform, to the eventual buyer, to the operating model that has to integrate it — stops being a portfolio-construction detail and becomes an architectural one.
The signal at large-cap, the mechanism at mid-market
Two of 2025’s largest sponsor-to-sponsor exits were add-ons being sold separately rather than platforms — Hometown Food’s $601M Chef Boyardee acquisition and NRG’s $560M Texas gas-gen pickup — both documented in PitchBook’s 2025 Annual US PE Middle Market Report. Both transactions were large-cap and should not be read as evidence of MM dynamics. The pattern of add-ons being carved out and sold individually is showing up first in large-cap, where buyer pools are deeper and the standalone underwriting case for a single asset is easier to construct.
The mid-market mechanism is different. MM platforms typically exit whole, with their add-ons going to the buyer as part of the platform sale rather than being separated out. What the 76% print changes for the mid-market is not the exit pattern itself; it is the cumulative effect of an architecture where most of the platform’s growth is now happening through add-ons rather than organically. The MM operator question that follows is not about whether add-ons will be sold separately at exit — most won’t — but about what the platform actually becomes when 76% of its acquired components are add-ons whose individual cases were never required to hold up beyond the integration moment.
What “add-on as the modal acquisition” actually changes
When add-ons were one in three buyouts, a platform could be defined by its initial acquisition and its operating thesis, with add-ons absorbed into that defining identity. When add-ons are three in four, the platform is increasingly defined by the cumulative shape of its add-on architecture — which acquisitions were chosen, how they were sequenced, and what each one contributed beyond capital deployment. The platform’s identity becomes a function of its add-on history rather than the other way around, and most operating models built for the prior ratio do not produce a coherent answer to the question of what the platform now is.
The platform-and-add-on selection criteria the buy-and-build literature has long distinguished are doing more work in this environment than they did in the prior one. When add-on selection drives most of the platform’s eventual shape, applying platform-grade selection criteria to add-on choices is no longer optional — it is the difference between a platform whose add-ons cohere into something valuable and one whose add-ons stack into something that doesn’t.
The compounding question
The 76% print also raises a question that runs underneath the velocity itself: whether the add-on-heavy architecture is producing compounding or merely stacking. Compounding through buy-and-build requires that each acquisition reduce the integration load on the next one — that operating capacity, customer overlap, and capability adjacencies grow through addition. Stacking, by contrast, produces growth in revenue and EBITDA without producing growth in coherent capability, and the difference is usually not visible at the velocity level. A platform doing eight add-ons a year and a platform doing three may both look healthy in deal-flow terms; only one of them is producing a more valuable platform with each acquisition.
The platforms I evaluated as acquisition candidates from a strategic vantage fell into two categories — those whose add-ons still had identifiable cases at the deal table, and those that had absorbed everything into a single identity that was harder to value at exit. The distinction wasn’t velocity-related. The platforms with high add-on counts that read as compounding had usually preserved enough definition in each acquisition that a downstream buyer could see what each piece contributed. The ones that read as stacking had often run integration as the elimination of distinction — converting every add-on into the platform’s existing identity as quickly as possible — and the resulting asset was harder to value precisely because the components had been blended past the point of being individually visible.
The underwriting question, on every acquisition
When the modal acquisition is an add-on and a meaningful share of platforms will eventually be evaluated by buyers who underwrite at the asset level rather than the platform level, every add-on now carries an implicit underwriting requirement that the prior environment’s add-ons did not. Each acquisition has to make sense not only as a contribution to the platform’s current operating thesis but as a piece whose case can still be reconstructed if a downstream buyer asks why it was acquired and what it now contributes. The buy-and-build platforms I’ve watched run integration as absorption — adding companies to a unified identity — produced different results from the platforms that ran integration as portfolio architecture, where each acquired piece kept enough standalone definition to be valued independently.
The buyer-fit question that becomes operationally specific in When the Buyer Profile Flips lands here in a related form. The platforms whose add-ons retain identifiable cases preserve more options at exit — both because some add-ons may be sold separately and because corporate strategics underwrite synergy at the asset level even when they buy whole platforms. The platforms whose add-ons have been absorbed into indistinguishability have narrowed their buyer pool to sponsors willing to underwrite the platform as a single, blended whole, and that pool is meaningfully smaller than it used to be.
What the operating model now has to do
The integration playbook for an environment where 76% of acquisitions are add-ons is not the integration playbook the prior environment produced. Integration capacity becomes binding much faster when the cadence of acquisition is higher and each one carries its own underwriting requirement. The failure modes that produce indistinguishable absorption — speed-prioritized integration that erases standalone definition, technology consolidation that deletes operating-model variation, leadership consolidation that removes the people who carried the add-on’s case — are no longer just integration risks. They are exit-value destruction mechanisms, and they compound at the velocity the asset class is now operating at.
Most operating models built in the prior add-on environment treat integration as the work of making each acquisition disappear into the platform. The work the new environment requires is closer to the opposite — making each acquisition contribute to the platform while preserving enough of its individual case that it remains visible to the people who eventually have to value it.
What the underwriting question actually settles
When 76% of buyouts are add-ons and a meaningful share of those add-ons may eventually be sold separately, the integration choice is no longer a sequencing question — it is an underwriting question being asked of every acquisition the platform makes, and most operating models still answer it as if the add-on were going to disappear into the platform forever.