33. Is Infrastructure a Distinct Asset Class from Private Equity?

4 November 2026
By Gilles Mougenot — founder of Argos, Senior Advisor at Argos Fund, former Chairman of France Invest

Unlisted infrastructure has long been presented to institutional investors as a distinct allocation: essential assets, relatively predictable revenues, partial inflation protection and lower volatility than buyout. But infrastructure funds now invest in service companies, growth platforms, data centers and complex projects. Their value-creation strategies sometimes come close to those of traditional buyout. Does the distinction still hold economically?

The question is not merely semantic. It bears on the allocation policies of pension funds and insurers, on managers’ return targets, on how risk is measured and, ultimately, on the treatment investors grant to an asset class presented as specific.

A distinction that rests first on the economics of the assets

Buyout consists of acquiring companies, improving their performance and selling them after several years. Unlisted infrastructure uses largely similar vehicles and financial techniques: closed-end funds, LP commitments, management fees, carried interest, leverage and asset sales. It would therefore be artificial to claim that the two activities rest on entirely different management models.

A first difference nonetheless appears in the life of the vehicles. According to the fund-terms database of the law firm Goodwin (November 2024), 56% of Private Equity funds have a 10-year term, whereas half of infrastructure funds are designed to run 12 years or more. Cleary Gottlieb adds that closed-end core and core-plus funds often have a term of around 15 years, with possible extensions, and sometimes 15 to 20 years or more, while value-add and opportunistic funds resemble a traditional Private Equity fund in both term and holding period. Fund life therefore follows the same fault line as the hypothesis set out below: core is a long-dated product, value-add is a buyout product.

One caution is in order. The fund term is a contractual ceiling, not a realized duration: buyout funds routinely extend their term when exits are slow.

The historical difference lies mainly in the underlying assets. Electricity grids, water distribution, transport infrastructure or concessions often offer high barriers to entry, a long life, demand that is relatively insensitive to the economic cycle and, in some cases, contracted, indexed or regulated revenues. These characteristics can make their cash flows more predictable than those of an industrial or commercial company exposed to competition.

But they are neither universal nor automatic. A toll road exposed to traffic, a regulated network, a merchant power plant and a data center exposed to technological change do not carry the same risk. The right unit of analysis is therefore not the “infrastructure” label, but the combination of demand, contracts, regulatory framework, debt, reinvestment needs and terminal value.

Record fundraising hides a spectacular concentration

In 2025, according to McKinsey using Preqin data, infrastructure funds raised $200 billion, against $126 billion in 2024, an increase of 58%. At the same time, Bain describes a broadly stable private capital market and a decline in buyout fundraising. At first sight, infrastructure thus seems to enjoy an enthusiasm that buyout no longer commands.

The reality is less simple. PitchBook counts, within its broader real assets perimeter, ten funds of $5 billion or more that raised $127.9 billion in 2025, about 60% of the total for that category. Bain observes at the same time a 27% drop in the number of infrastructure funds closed, despite a 58% rise in the amounts raised. The record owes largely to the size and timing of mega-funds, not to a growing number of teams able to raise capital.

2026 marks a turn in the market. According to Infrastructure Investor, first-half fundraising came to $40.8 billion, against $126.3 billion a year earlier. Infralogic confirms the order of magnitude over nine months: $97.2 billion in 2026, against $159.5 billion in 2025. Concentration shows up in the third quarter: of $65.5 billion in final closes, nearly two-thirds (64.5%) came from just three funds, KKR Global Infrastructure Investors V, Partners Group Direct Infrastructure IV and Energy Capital Partners VI. Excluding those three, the quarter totals only $23.2 billion. Figures vary with databases and definitions, but the finding is robust: less capital, and increasingly concentrated.

The shift toward value-add changes the nature of the risk

The traditional segmentation of infrastructure distinguishes core, core plus and value-add. In core, most of the value is supposed to come from mature assets and recurring cash flows. Core plus adds more operational or commercial risk. Value-add involves transformations, development, bolt-on acquisitions or growth investment.

Yet capital is moving toward the riskiest strategies. According to Infrastructure Investor, in the first half of 2026 value-add accounted for 52% of capital raised, against 20% for core plus and 9% for core. These are shares of the fundraising of a given period, not shares of the global stock of assets or of all investments made. It would therefore be excessive to conclude that 91% of existing infrastructure is no longer core.

The change is nonetheless significant. McKinsey describes a widening of the investable universe to infrastructure-related services and to assets that do not always fit the classic definition of the sector: data center maintenance, pipeline inspection and even, in some classifications, billboards or pharmacies. Value creation then depends less on the yield of an existing asset alone than on the execution of an industrial plan.

This is precisely where the boundary with buyout blurs. Building a platform, making acquisitions, optimizing costs, growing revenues and selling at an attractive multiple are methods common to both worlds. The difference may remain in the resilience of demand or in the contracts, but it has to be demonstrated asset by asset.

Energy and data centers: essential assets, new risks

Investment needs are considerable. McKinsey estimates that $106 trillion of global infrastructure investment is needed by 2040, of which $23 trillion for energy, and points to nearly $7 trillion for data centers by 2030.

Data centers illustrate the ambiguity particularly well. Some benefit from long-term contracts with strong customers. But they can also carry risks of customer concentration, electricity costs, grid connection, technological obsolescence, overcapacity and refinancing. Physical infrastructure that is indispensable to the digital economy is not necessarily a low-risk investment.

Energy calls for the same caution. Regulated networks are not the same as power plants exposed to market prices, projects under construction or renewable development platforms. In each case, profitability depends on distinct parameters: tariffs, purchase contracts, interest rates, construction costs, permits, energy prices and exit conditions.

The sale of Aligned Data Centers, valued at $40 billion in the data cited by Bain, reveals how porous the categories are: this major transaction, which comes from the infrastructure world, appears in the buyout statistics. The same asset can thus feed two different market narratives depending on the nature of the transaction and the classification used.

Is performance really different?

This is the decisive question, and also the one the public data answer least clearly.

McKinsey’s Global Private Markets Report 2025 puts the aggregate net IRR of infrastructure at 6.1% in 2024, after 8.5% in 2023, and cites a median of 9.3% for certain 2012-2021 vintages. These figures cannot be set directly against buyout returns without aligning vintages, geography, currency, fees, maturity stage and calculation methods.

Two comparisons nevertheless exist in the same source, on an identical scope and method. McKinsey calls infrastructure the best-performing private asset class of 2024, with that 6.1% net IRR, which is also its lowest level since 2011. And, across 2012-2021 vintages, the spread between performance percentiles is 8 points in infrastructure against 15 points in Private Equity. In other words: a modest median return, but dispersion half as wide. This is the strongest argument in favor of a distinct allocation.

McKinsey’s European study also finds that listed European infrastructure companies beat the market over 2014-2019 (total shareholder return, dividends included, of 83% over the period against 36% for “the market,” McKinsey not specifying the index), then clearly underperformed the Euronext 100 index over 2019-2025 (35% against 82%). The reversal is interesting for gauging the sector’s sensitivity to interest rates and the cycle.

To settle the matter seriously, one would need to compare, for identical vintages, net IRRs, capital multiples (TVPI), actual distributions (DPI), losses, dispersion among managers and performance after accounting for leverage. Above all, one would need to distinguish core, core plus and value-add rather than treating infrastructure as a homogeneous block. The sources gathered here do not yet allow that full comparison.

The question of risk is just as delicate, and valuations can be read two ways. On one hand, an EDHEC survey of 79 institutional investors, conducted in 2023-2024 and published in August 2025, indicates that 76% of them sell their infrastructure assets above the last net asset value, with typical premiums of 6% to 20%: net asset values would therefore be rather conservative. On the other, more than 60% of respondents rely mainly on management forecasts, only 28% revalue quarterly and 32% adjust their values in periods of stress. This means that the low volatility the asset class displays partly reflects slow, little-challenged valuation practices.

Liquidity: the promise of income does not replace distributions

One of the traditional arguments for infrastructure is its ability to generate regular income. But cash flows produced by the assets do not automatically turn into distributions to LPs. They may go to debt service, maintenance, growth investment or acquisitions.

The data cited by McKinsey show a weakening of five-year DPI for some recent vintages: the median cited for the 2020 vintage is 13%, against levels close to 40% for vintages of the mid-2010s. The comparison must be read with caution, because market conditions, fund composition and strategies may differ, and because longer-dated funds mechanically distribute later.

Is a distinct allocation still justified?

For an institutional investor, a stand-alone asset class is justified if it offers an identifiable combination of return, risk, liquidity and diversification that cannot be obtained as efficiently elsewhere.

Infrastructure can meet that test. A portfolio of regulated networks, concessions or assets backed by long contracts can show economic exposures different from those of a diversified buyout portfolio. The duration of the assets, the structure of revenues and certain contractual protections are genuine specificities.

But it is much less obvious that a value-add fund specializing in the acquisition and transformation of infrastructure-related service companies provides the same diversification. If its performance depends mainly on EBITDA growth, acquisition discipline, leverage and exit multiple, it shares several risk drivers with buyout. The sector label is no longer enough to justify an independent allocation bucket.

LPs should therefore put more precise questions to managers than “what share of the portfolio is invested in infrastructure?”: what proportion of revenues is contracted or regulated? What exposure to volumes and market prices? What share of the expected return comes from distributions, operational growth, deleveraging or the exit multiple? What are the construction, technology and refinancing risks? What economic correlation with the rest of the portfolio?

France offers a complementary perspective. France Invest reports €6.6 billion of infrastructure fundraising in the first half of 2026, against €3.5 billion a year earlier, while investments fell from €5.3 billion to €3.8 billion. This divergence is a reminder that strong fundraising says nothing, on its own, about the quality or pace of future investment. It reinforces the need to look at assets and strategies rather than at fundraising statistics alone.

Conclusion: one asset class, or several strategies under one label?

Infrastructure has not become buyout. The characteristics of many assets — natural monopolies, concessions, regulated or contracted revenues, essential needs — continue to justify a distinct analysis. But infrastructure is not an economically homogeneous category either.

The widening of its perimeter and the rise of value-add strategies bring part of the funds closer to the methods and risks of buyout. The concentration of fundraising on a few platforms and the growing sophistication of vehicles accentuate this trend without, on their own, proving it.

The hypothesis this analysis allows us to formulate is the following: the word “infrastructure” covers two economies. Core is a yield asset, regulated or contracted, whose appeal lies in its lower dispersion. Core plus and value-add belong to sector-specific buyout: platforms, bolt-on acquisitions, execution, exit at a multiple. Yet it is toward this second group that capital is flowing (52% and 20% of first-half 2026 fundraising).

A separate allocation is justified for the first economy. For the second, it risks mainly masking an additional exposure to the same risk factors as buyout, under a sector label that reassures. The question to put to managers is therefore simple: which of the two are you selling, and why charge for it as a separate asset class?

Sources

Gilles Mougenot — founder of Argos, Senior Advisor at Argos Fund, former Chairman of France Invest, author of Tout savoir sur le Capital Investissement.

The weekly Private Equity press review

Get PE Research’s weekly edition by e-mail — the essentials, without the noise.

Read next