In recent years, data centers, long treated mainly as real estate investments, are increasingly being classified as infrastructure assets by institutional investors. Leading private-equity infrastructure funds (such as Blackstone and Brookfield) have aggressively acquired data-center platforms, signaling a shift in how this sector is understood and capitalized. This article analyzes the factors driving that transition, examines why private-equity infrastructure funds are particularly motivated to buy, and critically assesses whether treating data centers as infrastructure might distort the conventional understanding of infrastructure given their continued exposure to market cycles.

Data centers occupy physical real estate: land, buildings, cooling systems, and power connections. But unlike traditional commercial property, they also deliver a mission-critical service layer, compute capacity, essential for the digital economy. This combination of real estate footprint and service delivery makes them what many call a “hybrid” asset class [1]. According to CBRE Investment Management, data centers provide both infrastructure-like cash flows and real estate-like lease characteristics. Their criticality to digital operations, high power demand, and long-term customer contracts make them more akin to infrastructure such as utilities than typical commercial property [2]. McKinsey similarly frames data centers within the evolving definition of infrastructure: they are becoming part of the “expanding foundations of modern society,” as digital and energy sectors converge [3].

One of the main reasons data centers are being viewed as infrastructure rather than simple real estate is the increasing electricity demand they impose. McKinsey’s analysis projects global data center capacity demand to grow from about 60 GW today to 171–219 GW by 2030, driven by both AI and non-AI workloads [4].

This surge in compute capacity requires a commensurate build-out of power infrastructure, transmission, and grid resources. McKinsey anticipates that by 2030, data centers will need around 1,400 TWh annually, equivalent to about 4 percent of global power demand [1]. In Europe, data center power consumption is projected to nearly triple by 2030, necessitating an estimated $250–$300 billion in infrastructure investment (excluding generation) to support that growth [5]. This power–compute coupling is not merely a technical matter: it underlines how data centers function as infrastructure, because their existence depends on and drives capital-intensive upgrades in the energy.

Data centers have become a focus for institutional capital, particularly private-equity infrastructure funds. According to CBRE’s 2024 Global Data Center Investor Intentions Survey, 97 percent of respondents plan to increase capital deployment in the data center sector, and 44 percent allocate more than $500 million annually [6]. The same survey showed a clear preference for turnkey hyperscale developments. More than 30 percent of investors believe such projects are the most attractive over the next two years, highlighting the appeal of large, scalable platforms with long-term clients [6].

Large private-equity firms are making significant bets on data centers:

• Blackstone acquired AirTrunk for ~$16 billion, significantly expanding its global digital infrastructure presence. The deal builds on Blackstone’s existing $55 billion portfolio in data centers and its ambitious development pipeline [7].

• Brookfield Asset Management pledged up to ~$9.9 billion for a new AI data-center campus in Sweden. The scale and duration (planned over 10–15 years) reflect long-horizon infrastructure thinking [8].

From the perspective of these investors, data centers deliver stable, contracted cash flows, often underpinned by large tech tenants (hyperscalers), and benefit from long-term power and lease commitments. These attributes align closely with infrastructure-style risk-return profiles.

McKinsey’s private-market research identifies value-creation levers for infrastructure funds investing in data centers. Instead of passive or purely financial ownership, these funds are increasingly engaging in active management: optimizing design and construction, strategically acquiring sites with favorable power access, and working with grid operators to secure future capacity [9]. For example, building large-scale campuses rather than dispersed small facilities allows economies of scale. McKinsey’s design insight suggests that scaling from tens of megawatts to gigawatt-scale campuses will be critical to meeting demand efficiently [1]. Funds that can mobilize capital, expertise, and long-term power commitments are well positioned to capture that value.

While data centers are being framed as infrastructure, there are important caveats that challenge this classification. Unlike traditional infrastructure (roads, water, power lines), data centers remain exposed to market cycles, technology risk, and lease risk.

Traditional infrastructure assets typically generate stable, regulated cash flows with limited sensitivity to market cycles. In contrast, data centers remain subject to fluctuations in demand, especially from tech companies and hyperscalers. AI workload demand, while currently strong, may evolve based on macroeconomic conditions, changes in enterprise adoption, or innovations in compute efficiency. Investors face a potential risk of overbuilding: if projected demand does not materialize, data center capacity could be underutilized, leading to margin compression or stranded assets. McKinsey itself cautions that uncertainty around AI adoption, model efficiency, and chip mix pose risks to capacity forecasts [4]. Because of such risks, data centers may not offer the same downside protection as more traditional infrastructure, making them more akin to real estate or private equity investments than regulated utilities.

Compute infrastructure is particularly subject to rapid technological change: new chip architectures, cooling technologies, and network designs can render existing facilities less competitive. A data center built purely for current generation GPUs may become suboptimal if future AI workloads require different hardware or more efficient cooling. Private-equity investors must therefore continuously invest not only in physical capacity but also in technology refresh cycles. This differs from many infrastructure assets, which often have decades-long lifespans with minimal technological turnover.

While the coupling with energy infrastructure is a strength, it also introduces regulatory and operational risk. Data center expansion depends on transmission upgrades, generation capacity, and favorable energy contracts. Delays in grid expansion, permitting, or power generation can slow development or raise costs. Moreover, given the concentration of data centers in power-constrained markets, there is a risk of grid congestion. Some academic work suggests that data centers’ spatial flexibility (the ability to shift compute loads) could mitigate congestion, but such flexibility may not always align with commercial leases or customer demand [10].

High energy consumption and carbon intensity are also a concern. Research shows that data centers already contribute significantly to emissions: in the U.S., they account for over 4 percent of electricity use, with over half derived from fossil fuels [11]. As infrastructure investors increasingly weigh ESG criteria, the carbon footprint of data centers may present reputational or regulatory risk. In addition, capital-intensive upgrades for decarbonization (e.g., on-site renewables, cooling innovations) are necessary but may compress returns or require new operational models.

To understand how data centers straddle real estate and infrastructure, it is useful to conceptualize them along three overlapping dimensions:

1. Physical real estate: land, buildings, cooling, and connectivity infrastructure.

2. Compute service: provision of processing capacity, often under contract to hyperscalers or enterprises.

3. Energy system integration: reliance on and impact on power grid infrastructure.

Institutional infrastructure investors are increasingly prioritizing the third dimension. As McKinsey notes, data centers are not just power consumers, they actively drive grid investment, generation planning, and transmission expansion [12]. Infrastructure funds that can align compute real estate with energy strategies (e.g., renewable PPAs, local generation, or storage) create a vertically integrated value chain. CBRE Investment Management articulates a similar view: data centers are “mission-critical digital infrastructure” with risk-return characteristics more aligned with utilities than speculative real estate [2]. This framing helps justify their inclusion in infrastructure portfolios.

The hybrid identity of data centers has several strategic implications for infrastructure funds:

• Long-Term Capital: Because data centers require large upfront capex and ongoing operating costs, they favor investors with long-term horizons and patient capital. Private equity infrastructure funds fit this profile.

• Active Management: Value is created not only through ownership but also through operational optimization, design innovation, and grid partnerships. Funds must engage actively across domains.

• Risk Mitigation: Investors must hedge technology and demand risks, perhaps through flexible leases, technology refresh planning, and co-investment in grids.

• Sustainability: ESG strategies, from renewable energy sourcing to carbon metrics, are increasingly essential. Infrastructure funds must integrate environmental objectives into their data center investments.

Labeling data centers as infrastructure raises a risk of conceptual distortion, especially when compared to classic infrastructure assets:

1. Market Cycle Sensitivity: Data centers are exposed to fluctuations in tech demand, unlike regulated utilities with predictable cash flows. If investors too loosely classify them as infrastructure, they may overestimate stability or under appreciate volatility.

2. Technological Obsolescence: The rapid pace of innovation could render large data centers redundant or inefficient. Traditional infrastructure, by contrast, often has longer operational lifetimes and slower technological turnover.

3. Regulatory Risk: Power grid constraints, permitting delays, and environmental regulation could derail data center development. Infrastructure investors may underestimate these risks if they conflate data centers with simpler, less regulated assets.

These distortions suggest that while data centers should be considered infrastructure in many respects, they also require a distinct investment framework, one that acknowledges both their infrastructure and private equity characteristics.

Data centers are unequivocally moving beyond mere real estate: their evolving role as power-intensive compute platforms, coupled with deep integration into the energy system, positions them as next-generation infrastructure. This hybrid identity makes them highly attractive to private-equity infrastructure funds such as Blackstone and Brookfield, which bring the capital, operational expertise, and long-term horizon required to capitalize on this shift. However, the classification of data centers as infrastructure should not obscure their exposure to market cycles, rapid technological change, and energy risk. Their dual nature demands a nuanced investment approach, one that neither treats them as traditional utilities nor as speculative real estate.

Ultimately, data centers represent an emergent asset class: infrastructure by necessity, real estate by form, and compute by function. As investors continue to commit capital, the challenge will be balancing the infrastructure-like attributes that drive stability with the market, technology, and energy risks that could undermine it.

1. Scaling bigger, faster, cheaper data centers with smarter designs | McKinsey & Company
   https://www.mckinsey.com/industries/private-capital/our-insights/scaling-bigger-faster-cheaper-data-centers-with-smarter-designs
   

2. Perspectives: Infra-AI – Infrastructure ready to run | CBRE (2024)
   https://mediaassets.cbre.com/-/media/project/cbre/bussectors/cbreim/insights/articles/2024-media-folder/infra-ai/infra-ai-2024.pdf?rev=4f061cd763ec4e3a8d22ed7f574fd4e3
   

3. The infrastructure moment: Investing in the expanding foundations of modern society | McKinsey & Company
   https://www.mckinsey.com/~/media/mckinsey/industries/infrastructure/our%20insights/the%20infrastructure%20moment/the-infrastructure-moment-investing-in-the-expanding-foundations-of-modern-society.pdf
   

4. AI data centre growth: Meeting the demand | McKinsey & Company https://www.mckinsey.com/industries/technology-media-and-telecommunications/our-insights/ai-power-expanding-data-center-capacity-to-meet-growing-demand
   

5. Europe’s data centre power demand expected to triple by 2030, McKinsey report says | Reuters https://www.reuters.com/technology/europes-data-centre-power-demand-expected-triple-by-2030-mckinsey-report-says-2024-10-23/

In recent years, data centers, long treated mainly as real estate investments, are increasingly being classified as infrastructure assets by institutional investors. Leading private-equity infrastructure funds (such as Blackstone and Brookfield) have aggressively acquired data-center platforms, signaling a shift in how this sector is understood and capitalized. This article analyzes the factors driving that transition, examines why private-equity infrastructure funds are particularly motivated to buy, and critically assesses whether treating data centers as infrastructure might distort the conventional understanding of infrastructure given their continued exposure to market cycles.

Data centers occupy physical real estate: land, buildings, cooling systems, and power connections. But unlike traditional commercial property, they also deliver a mission-critical service layer, compute capacity, essential for the digital economy. This combination of real estate footprint and service delivery makes them what many call a “hybrid” asset class [1]. According to CBRE Investment Management, data centers provide both infrastructure-like cash flows and real estate-like lease characteristics. Their criticality to digital operations, high power demand, and long-term customer contracts make them more akin to infrastructure such as utilities than typical commercial property [2]. McKinsey similarly frames data centers within the evolving definition of infrastructure: they are becoming part of the “expanding foundations of modern society,” as digital and energy sectors converge [3].

One of the main reasons data centers are being viewed as infrastructure rather than simple real estate is the increasing electricity demand they impose. McKinsey’s analysis projects global data center capacity demand to grow from about 60 GW today to 171–219 GW by 2030, driven by both AI and non-AI workloads [4].

This surge in compute capacity requires a commensurate build-out of power infrastructure, transmission, and grid resources. McKinsey anticipates that by 2030, data centers will need around 1,400 TWh annually, equivalent to about 4 percent of global power demand [1]. In Europe, data center power consumption is projected to nearly triple by 2030, necessitating an estimated $250–$300 billion in infrastructure investment (excluding generation) to support that growth [5]. This power–compute coupling is not merely a technical matter: it underlines how data centers function as infrastructure, because their existence depends on and drives capital-intensive upgrades in the energy.

Data centers have become a focus for institutional capital, particularly private-equity infrastructure funds. According to CBRE’s 2024 Global Data Center Investor Intentions Survey, 97 percent of respondents plan to increase capital deployment in the data center sector, and 44 percent allocate more than $500 million annually [6]. The same survey showed a clear preference for turnkey hyperscale developments. More than 30 percent of investors believe such projects are the most attractive over the next two years, highlighting the appeal of large, scalable platforms with long-term clients [6].

Large private-equity firms are making significant bets on data centers:

• Blackstone acquired AirTrunk for ~$16 billion, significantly expanding its global digital infrastructure presence. The deal builds on Blackstone’s existing $55 billion portfolio in data centers and its ambitious development pipeline [7].

• Brookfield Asset Management pledged up to ~$9.9 billion for a new AI data-center campus in Sweden. The scale and duration (planned over 10–15 years) reflect long-horizon infrastructure thinking [8].

From the perspective of these investors, data centers deliver stable, contracted cash flows, often underpinned by large tech tenants (hyperscalers), and benefit from long-term power and lease commitments. These attributes align closely with infrastructure-style risk-return profiles.

McKinsey’s private-market research identifies value-creation levers for infrastructure funds investing in data centers. Instead of passive or purely financial ownership, these funds are increasingly engaging in active management: optimizing design and construction, strategically acquiring sites with favorable power access, and working with grid operators to secure future capacity [9]. For example, building large-scale campuses rather than dispersed small facilities allows economies of scale. McKinsey’s design insight suggests that scaling from tens of megawatts to gigawatt-scale campuses will be critical to meeting demand efficiently [1]. Funds that can mobilize capital, expertise, and long-term power commitments are well positioned to capture that value.

While data centers are being framed as infrastructure, there are important caveats that challenge this classification. Unlike traditional infrastructure (roads, water, power lines), data centers remain exposed to market cycles, technology risk, and lease risk.

Traditional infrastructure assets typically generate stable, regulated cash flows with limited sensitivity to market cycles. In contrast, data centers remain subject to fluctuations in demand, especially from tech companies and hyperscalers. AI workload demand, while currently strong, may evolve based on macroeconomic conditions, changes in enterprise adoption, or innovations in compute efficiency. Investors face a potential risk of overbuilding: if projected demand does not materialize, data center capacity could be underutilized, leading to margin compression or stranded assets. McKinsey itself cautions that uncertainty around AI adoption, model efficiency, and chip mix pose risks to capacity forecasts [4]. Because of such risks, data centers may not offer the same downside protection as more traditional infrastructure, making them more akin to real estate or private equity investments than regulated utilities.

Compute infrastructure is particularly subject to rapid technological change: new chip architectures, cooling technologies, and network designs can render existing facilities less competitive. A data center built purely for current generation GPUs may become suboptimal if future AI workloads require different hardware or more efficient cooling. Private-equity investors must therefore continuously invest not only in physical capacity but also in technology refresh cycles. This differs from many infrastructure assets, which often have decades-long lifespans with minimal technological turnover.

While the coupling with energy infrastructure is a strength, it also introduces regulatory and operational risk. Data center expansion depends on transmission upgrades, generation capacity, and favorable energy contracts. Delays in grid expansion, permitting, or power generation can slow development or raise costs. Moreover, given the concentration of data centers in power-constrained markets, there is a risk of grid congestion. Some academic work suggests that data centers’ spatial flexibility (the ability to shift compute loads) could mitigate congestion, but such flexibility may not always align with commercial leases or customer demand [10].

High energy consumption and carbon intensity are also a concern. Research shows that data centers already contribute significantly to emissions: in the U.S., they account for over 4 percent of electricity use, with over half derived from fossil fuels [11]. As infrastructure investors increasingly weigh ESG criteria, the carbon footprint of data centers may present reputational or regulatory risk. In addition, capital-intensive upgrades for decarbonization (e.g., on-site renewables, cooling innovations) are necessary but may compress returns or require new operational models.

To understand how data centers straddle real estate and infrastructure, it is useful to conceptualize them along three overlapping dimensions:

1. Physical real estate: land, buildings, cooling, and connectivity infrastructure.

2. Compute service: provision of processing capacity, often under contract to hyperscalers or enterprises.

3. Energy system integration: reliance on and impact on power grid infrastructure.

Institutional infrastructure investors are increasingly prioritizing the third dimension. As McKinsey notes, data centers are not just power consumers, they actively drive grid investment, generation planning, and transmission expansion [12]. Infrastructure funds that can align compute real estate with energy strategies (e.g., renewable PPAs, local generation, or storage) create a vertically integrated value chain. CBRE Investment Management articulates a similar view: data centers are “mission-critical digital infrastructure” with risk-return characteristics more aligned with utilities than speculative real estate [2]. This framing helps justify their inclusion in infrastructure portfolios.

The hybrid identity of data centers has several strategic implications for infrastructure funds:

• Long-Term Capital: Because data centers require large upfront capex and ongoing operating costs, they favor investors with long-term horizons and patient capital. Private equity infrastructure funds fit this profile.

• Active Management: Value is created not only through ownership but also through operational optimization, design innovation, and grid partnerships. Funds must engage actively across domains.

• Risk Mitigation: Investors must hedge technology and demand risks, perhaps through flexible leases, technology refresh planning, and co-investment in grids.

• Sustainability: ESG strategies, from renewable energy sourcing to carbon metrics, are increasingly essential. Infrastructure funds must integrate environmental objectives into their data center investments.

Labeling data centers as infrastructure raises a risk of conceptual distortion, especially when compared to classic infrastructure assets:

1. Market Cycle Sensitivity: Data centers are exposed to fluctuations in tech demand, unlike regulated utilities with predictable cash flows. If investors too loosely classify them as infrastructure, they may overestimate stability or under appreciate volatility.

2. Technological Obsolescence: The rapid pace of innovation could render large data centers redundant or inefficient. Traditional infrastructure, by contrast, often has longer operational lifetimes and slower technological turnover.

3. Regulatory Risk: Power grid constraints, permitting delays, and environmental regulation could derail data center development. Infrastructure investors may underestimate these risks if they conflate data centers with simpler, less regulated assets.

These distortions suggest that while data centers should be considered infrastructure in many respects, they also require a distinct investment framework, one that acknowledges both their infrastructure and private equity characteristics.

Data centers are unequivocally moving beyond mere real estate: their evolving role as power-intensive compute platforms, coupled with deep integration into the energy system, positions them as next-generation infrastructure. This hybrid identity makes them highly attractive to private-equity infrastructure funds such as Blackstone and Brookfield, which bring the capital, operational expertise, and long-term horizon required to capitalize on this shift. However, the classification of data centers as infrastructure should not obscure their exposure to market cycles, rapid technological change, and energy risk. Their dual nature demands a nuanced investment approach, one that neither treats them as traditional utilities nor as speculative real estate.

Ultimately, data centers represent an emergent asset class: infrastructure by necessity, real estate by form, and compute by function. As investors continue to commit capital, the challenge will be balancing the infrastructure-like attributes that drive stability with the market, technology, and energy risks that could undermine it.

1. Scaling bigger, faster, cheaper data centers with smarter designs | McKinsey & Company
   https://www.mckinsey.com/industries/private-capital/our-insights/scaling-bigger-faster-cheaper-data-centers-with-smarter-designs
   

2. Perspectives: Infra-AI – Infrastructure ready to run | CBRE (2024)
   https://mediaassets.cbre.com/-/media/project/cbre/bussectors/cbreim/insights/articles/2024-media-folder/infra-ai/infra-ai-2024.pdf?rev=4f061cd763ec4e3a8d22ed7f574fd4e3
   

3. The infrastructure moment: Investing in the expanding foundations of modern society | McKinsey & Company
   https://www.mckinsey.com/~/media/mckinsey/industries/infrastructure/our%20insights/the%20infrastructure%20moment/the-infrastructure-moment-investing-in-the-expanding-foundations-of-modern-society.pdf
   

4. AI data centre growth: Meeting the demand | McKinsey & Company https://www.mckinsey.com/industries/technology-media-and-telecommunications/our-insights/ai-power-expanding-data-center-capacity-to-meet-growing-demand
   

5. Europe’s data centre power demand expected to triple by 2030, McKinsey report says | Reuters https://www.reuters.com/technology/europes-data-centre-power-demand-expected-triple-by-2030-mckinsey-report-says-2024-10-23/