Compute infrastructure, most prominently large‑scale data centers, has rapidly shifted from a conventional capital expense in technology budgets to a core asset class within global capital markets. Driven by exponential growth in artificial intelligence (AI), cloud computing, and data‑driven services, compute capacity now underpins economic activity across sectors and borders. As a result, institutional investors, including sovereign wealth funds, infrastructure funds, and long‑duration capital providers, are increasingly allocating capital to compute assets that resemble traditional physical infrastructure rather than short‑lived technology projects.

This article analyzes how capital markets are repricing compute, the mechanisms through which institutional capital participates, the implications for cost of capital and risk, and the broader structural forces shaping the evolution of compute investment globally.

Approximately $6.7 trillion in capital expenditures will be required to build and upgrade data centers worldwide to meet anticipated compute demand, with AI‑related workloads accounting for the majority share of that need [1][2]. Within this total, it is estimated that $5.2 trillion is required specifically for AI‑optimized infrastructure capacity, while the remaining demand stems from non‑AI workloads [1][2]. Total data center capacity needs could almost triple by 2030, driven by a dramatic rise in AI and general cloud computing usage [1]. This trajectory suggests that compute investment requirements now exceed many traditional infrastructure categories, such as transportation or power transmission in terms of cumulative capital needs over the next decade.

Compute infrastructure’s position within global investment flows is well documented in the World Investment Report 2025 from the United Nations Conference on Trade and Development (UNCTAD). Data centers accounted for over one‑fifth of global greenfield foreign direct investment (FDI) project value in 2025, with announced investment exceeding $270 billion, a level that places compute investment among the most capital‑intensive greenfield sectors worldwide [2][3].

UNCTAD’s data further show that global FDI flows rose about 14% in 2025, to an estimated $1.6 trillion, and much of this increase stems from technology‑intensive and digital infrastructure segments such as data centers and semiconductor manufacturing [3]. This reflects a shift in cross‑border capital allocation towards assets that exhibit stable, predictable cash flows and long useful lives, characteristics that institutional investors often seek in infrastructure allocations. Importantly, UNCTAD notes that while compute investment is rising sharply in developed economies (e.g., United States, France, Republic of Korea), many developing countries remain underfunded in core infrastructure categories tied to digital economy development [3][4]. This unevenness reflects differences in policy frameworks, energy reliability, regulatory environments, and perceived investment risk.

Sovereign wealth funds (SWFs) and long‑duration institutional capital pools are increasingly active in compute infrastructure investment. While comprehensive public disclosure of SWF investments is limited, UNCTAD’s analysis of global FDI flows demonstrates that sovereign and state‑linked capital participates meaningfully in digital economy projects, including data centers, especially in middle‑income and strategic markets [4].

Institutional investors are drawn to data centers for several reasons:

• Long‑term contracted revenue streams — data centers often operate under long‑term service agreements with corporate or hyperscale tenants.

• Stable cash flows — predictable capacity usage and contractual pricing reduce volatility relative to other technology ventures.

• Low correlation with cyclical markets — compute demand often remains resilient across economic cycles due to its foundational role in digital services.

These characteristics align closely with the risk/return preferences of infrastructure funds, pension funds, and sovereign investors, which typically seek long‑duration investments with stable yield profiles.

The cost of capital for data center projects is evolving as institutional capital enters the market. Unlike traditional corporate finance, which relies on balance sheet debt or internal cash flows, large compute projects increasingly use project finance, blended capital stacks, and infrastructure‑like debt instruments. Institutional investors often participate via yield‑oriented vehicles, infrastructure debt funds, or joint venture equity arrangements that provide stable long‑term returns. Data center investment’s risk profile, anchored to contracted leases and diversified client bases, allows capital market participants to price risk differently than for typical venture or technology sector investments. The expectation of long‑duration cash flows and potential inflation protection, for example through contractual pricing escalators, has attracted institutional capital into compute assets that might previously have remained outside core infrastructure portfolios.

Yet this emerging capital stack is not without complexity: aligning return expectations across private equity, long‑term lenders, institutional equity, and project sponsors requires finely tuned structuring, increasing transaction costs and operational oversight compared with traditional corporate financing.

Despite rapid growth in compute investment overall, UNCTAD’s research highlights that the benefits remain unevenly distributed. While FDI into the digital economy has climbed from about 5.5% to 8.3% of total FDI over the past decade, developing countries accounted for only a fraction of core digital infrastructure investment, with most greenfield ICT infrastructure projects funded in a small number of economies [4]. In 2024, for example, just $9 billion was invested in ICT infrastructure in developing countries, far below the $62 billion annually required globally, indicating persistent infrastructure gaps that could limit long‑term equitable development [4]. This pattern implies that markets with strong governance frameworks, reliable energy grids, and robust connectivity have become preferred destinations for compute capital, reinforcing regional disparities.

UNCTAD’s policy research underscores that regulatory conditions are a key determinant of where compute infrastructure investments flow. Stable regulatory regimes, transparent investment incentives, and supportive digital economy strategies strengthen investor confidence and reduce perceived risk, which can, in turn, lower the cost of capital for projects [5]. For example, clear permitting processes for grid connections, data protection frameworks that enable cross‑border operations, and investment guarantees can significantly enhance a jurisdiction’s appeal to institutional investors targeting compute infrastructure. Conversely, complex regulatory requirements and policy uncertainty raise perceived risk premiums, particularly in emerging economies, reducing capital inflows.

The magnitude of compute investment needs, measured in the trillions, suggests that traditional corporate financing alone cannot sustain the pace of build‑outs required. Public financial markets, internal cash flows, and commercial banking sources can only partially address these needs. As a result, institutional capital from sovereign funds, pension assets, and infrastructure debt markets plays an increasingly central role. By integrating compute infrastructure into mainstream capital markets, investors are repositioning it alongside assets such as long‑haul transportation, power transmission, and communication networks, long regarded as foundational to economic development. This repositioning reflects compute infrastructure’s dual identity: a provider of essential services for the digital economy and a provider of stable long‑term cash flows desirable to institutional capital.

Compute demand, particularly from AI workloads, will shape how capital markets price infrastructure risk and performance. AI workload capacity could grow by approximately 3.5 times between 2025 and 2030, and AI workloads may constitute around 70% of total data center capacity demand by 2030 [1]. This unprecedented growth trajectory creates opportunities, but also introduces risk. Project sponsors and capital providers must navigate uncertainties around demand forecasts, technology shifts (e.g., more efficient AI models), and operational cost structures (e.g., power and cooling costs). These uncertainties can widen risk premiums or make long‑term contracts more attractive to institutional investors seeking predictable returns.

Institutional investors, particularly those with long investment horizons, are attracted to compute infrastructure because it aligns broadly with infrastructure asset characteristics: durable physical capital, higher entry costs, and long useful lives. But compute also presents unique risk factors, including:

1. Technology obsolescence — rapid changes in compute hardware or software can shorten asset lifespans;

2. Energy dependencies — data centers require reliable, often heavy, energy infrastructure;

3. Regulatory uncertainties — data sovereignty, privacy laws, and cross‑border data flow regulations vary widely;

4. Demand uncertainties — while AI demand is strong, long‑term patterns and elasticity remain imprecise.

Institutional capital spreads these risks across larger pools, for example, through diversified portfolios of data center assets or phased financing structures, and often pairs capital with operational expertise from experienced partners in hyperscale computing.

Capital markets are actively repricing compute by treating data centers and related digital infrastructure as long‑duration, infrastructure‑like assets. This reclassification reflects not only the enormous cumulative capital required to support AI and cloud compute demand but also structural shifts in how institutional investors allocate capital. Sovereign wealth funds, infrastructure debt providers, and long‑duration equity pools are increasingly participating in compute financing, attracted by stable contractual cash flows, long asset lifespans, and essential service characteristics.

However, these investment flows are concentrated in jurisdictions with supportive regulatory frameworks and substantial infrastructure readiness, leaving significant global inequities. Bridging these gaps will require coordinated policy action, targeted risk mitigation mechanisms, and blended finance approaches that align public objectives with private capital incentives. As compute becomes central to global competitiveness and economic growth, understanding how capital markets price and finance compute infrastructure will be essential for investors, policymakers, and industry leaders alike.

1. The Cost of Compute: A $7 Trillion Race to Scale Data Centers | McKinsey & Company (2025)
   https://www.mckinsey.com/industries/technology-media-and-telecommunications/our-insights/the-cost-of-compute-a-7-trillion-dollar-race-to-scale-data-centers 

2. Data Center Demands and Capacity Forecasts | McKinsey & Company (2025)
   https://www.mckinsey.com/featured-insights/week-in-charts/data-center-demands 

3. World Investment Report 2025: International Investment in the Digital Economy | UNCTAD (2025)
   https://investmentpolicy.unctad.org/publications/1310/world-investment-report-2025-international-investment-in-the-digital-economy 

4. Global Investment in the Digital Economy Surges but Remains Uneven | UNCTAD (2025)
   https://unctad.org/news/global-investment-digital-economy-surges-remains-uneven 

5. The Investment Dimension of Digital Strategies | UNCTAD (2025)
   https://unctad.org/publication/investment-dimension-digital-strategies 

Compute infrastructure, most prominently large‑scale data centers, has rapidly shifted from a conventional capital expense in technology budgets to a core asset class within global capital markets. Driven by exponential growth in artificial intelligence (AI), cloud computing, and data‑driven services, compute capacity now underpins economic activity across sectors and borders. As a result, institutional investors, including sovereign wealth funds, infrastructure funds, and long‑duration capital providers, are increasingly allocating capital to compute assets that resemble traditional physical infrastructure rather than short‑lived technology projects.

This article analyzes how capital markets are repricing compute, the mechanisms through which institutional capital participates, the implications for cost of capital and risk, and the broader structural forces shaping the evolution of compute investment globally.

Approximately $6.7 trillion in capital expenditures will be required to build and upgrade data centers worldwide to meet anticipated compute demand, with AI‑related workloads accounting for the majority share of that need [1][2]. Within this total, it is estimated that $5.2 trillion is required specifically for AI‑optimized infrastructure capacity, while the remaining demand stems from non‑AI workloads [1][2]. Total data center capacity needs could almost triple by 2030, driven by a dramatic rise in AI and general cloud computing usage [1]. This trajectory suggests that compute investment requirements now exceed many traditional infrastructure categories, such as transportation or power transmission in terms of cumulative capital needs over the next decade.

Compute infrastructure’s position within global investment flows is well documented in the World Investment Report 2025 from the United Nations Conference on Trade and Development (UNCTAD). Data centers accounted for over one‑fifth of global greenfield foreign direct investment (FDI) project value in 2025, with announced investment exceeding $270 billion, a level that places compute investment among the most capital‑intensive greenfield sectors worldwide [2][3].

UNCTAD’s data further show that global FDI flows rose about 14% in 2025, to an estimated $1.6 trillion, and much of this increase stems from technology‑intensive and digital infrastructure segments such as data centers and semiconductor manufacturing [3]. This reflects a shift in cross‑border capital allocation towards assets that exhibit stable, predictable cash flows and long useful lives, characteristics that institutional investors often seek in infrastructure allocations. Importantly, UNCTAD notes that while compute investment is rising sharply in developed economies (e.g., United States, France, Republic of Korea), many developing countries remain underfunded in core infrastructure categories tied to digital economy development [3][4]. This unevenness reflects differences in policy frameworks, energy reliability, regulatory environments, and perceived investment risk.

Sovereign wealth funds (SWFs) and long‑duration institutional capital pools are increasingly active in compute infrastructure investment. While comprehensive public disclosure of SWF investments is limited, UNCTAD’s analysis of global FDI flows demonstrates that sovereign and state‑linked capital participates meaningfully in digital economy projects, including data centers, especially in middle‑income and strategic markets [4].

Institutional investors are drawn to data centers for several reasons:

• Long‑term contracted revenue streams — data centers often operate under long‑term service agreements with corporate or hyperscale tenants.

• Stable cash flows — predictable capacity usage and contractual pricing reduce volatility relative to other technology ventures.

• Low correlation with cyclical markets — compute demand often remains resilient across economic cycles due to its foundational role in digital services.

These characteristics align closely with the risk/return preferences of infrastructure funds, pension funds, and sovereign investors, which typically seek long‑duration investments with stable yield profiles.

The cost of capital for data center projects is evolving as institutional capital enters the market. Unlike traditional corporate finance, which relies on balance sheet debt or internal cash flows, large compute projects increasingly use project finance, blended capital stacks, and infrastructure‑like debt instruments. Institutional investors often participate via yield‑oriented vehicles, infrastructure debt funds, or joint venture equity arrangements that provide stable long‑term returns. Data center investment’s risk profile, anchored to contracted leases and diversified client bases, allows capital market participants to price risk differently than for typical venture or technology sector investments. The expectation of long‑duration cash flows and potential inflation protection, for example through contractual pricing escalators, has attracted institutional capital into compute assets that might previously have remained outside core infrastructure portfolios.

Yet this emerging capital stack is not without complexity: aligning return expectations across private equity, long‑term lenders, institutional equity, and project sponsors requires finely tuned structuring, increasing transaction costs and operational oversight compared with traditional corporate financing.

Despite rapid growth in compute investment overall, UNCTAD’s research highlights that the benefits remain unevenly distributed. While FDI into the digital economy has climbed from about 5.5% to 8.3% of total FDI over the past decade, developing countries accounted for only a fraction of core digital infrastructure investment, with most greenfield ICT infrastructure projects funded in a small number of economies [4]. In 2024, for example, just $9 billion was invested in ICT infrastructure in developing countries, far below the $62 billion annually required globally, indicating persistent infrastructure gaps that could limit long‑term equitable development [4]. This pattern implies that markets with strong governance frameworks, reliable energy grids, and robust connectivity have become preferred destinations for compute capital, reinforcing regional disparities.

UNCTAD’s policy research underscores that regulatory conditions are a key determinant of where compute infrastructure investments flow. Stable regulatory regimes, transparent investment incentives, and supportive digital economy strategies strengthen investor confidence and reduce perceived risk, which can, in turn, lower the cost of capital for projects [5]. For example, clear permitting processes for grid connections, data protection frameworks that enable cross‑border operations, and investment guarantees can significantly enhance a jurisdiction’s appeal to institutional investors targeting compute infrastructure. Conversely, complex regulatory requirements and policy uncertainty raise perceived risk premiums, particularly in emerging economies, reducing capital inflows.

The magnitude of compute investment needs, measured in the trillions, suggests that traditional corporate financing alone cannot sustain the pace of build‑outs required. Public financial markets, internal cash flows, and commercial banking sources can only partially address these needs. As a result, institutional capital from sovereign funds, pension assets, and infrastructure debt markets plays an increasingly central role. By integrating compute infrastructure into mainstream capital markets, investors are repositioning it alongside assets such as long‑haul transportation, power transmission, and communication networks, long regarded as foundational to economic development. This repositioning reflects compute infrastructure’s dual identity: a provider of essential services for the digital economy and a provider of stable long‑term cash flows desirable to institutional capital.

Compute demand, particularly from AI workloads, will shape how capital markets price infrastructure risk and performance. AI workload capacity could grow by approximately 3.5 times between 2025 and 2030, and AI workloads may constitute around 70% of total data center capacity demand by 2030 [1]. This unprecedented growth trajectory creates opportunities, but also introduces risk. Project sponsors and capital providers must navigate uncertainties around demand forecasts, technology shifts (e.g., more efficient AI models), and operational cost structures (e.g., power and cooling costs). These uncertainties can widen risk premiums or make long‑term contracts more attractive to institutional investors seeking predictable returns.

Institutional investors, particularly those with long investment horizons, are attracted to compute infrastructure because it aligns broadly with infrastructure asset characteristics: durable physical capital, higher entry costs, and long useful lives. But compute also presents unique risk factors, including:

1. Technology obsolescence — rapid changes in compute hardware or software can shorten asset lifespans;

2. Energy dependencies — data centers require reliable, often heavy, energy infrastructure;

3. Regulatory uncertainties — data sovereignty, privacy laws, and cross‑border data flow regulations vary widely;

4. Demand uncertainties — while AI demand is strong, long‑term patterns and elasticity remain imprecise.

Institutional capital spreads these risks across larger pools, for example, through diversified portfolios of data center assets or phased financing structures, and often pairs capital with operational expertise from experienced partners in hyperscale computing.

Capital markets are actively repricing compute by treating data centers and related digital infrastructure as long‑duration, infrastructure‑like assets. This reclassification reflects not only the enormous cumulative capital required to support AI and cloud compute demand but also structural shifts in how institutional investors allocate capital. Sovereign wealth funds, infrastructure debt providers, and long‑duration equity pools are increasingly participating in compute financing, attracted by stable contractual cash flows, long asset lifespans, and essential service characteristics.

However, these investment flows are concentrated in jurisdictions with supportive regulatory frameworks and substantial infrastructure readiness, leaving significant global inequities. Bridging these gaps will require coordinated policy action, targeted risk mitigation mechanisms, and blended finance approaches that align public objectives with private capital incentives. As compute becomes central to global competitiveness and economic growth, understanding how capital markets price and finance compute infrastructure will be essential for investors, policymakers, and industry leaders alike.

1. The Cost of Compute: A $7 Trillion Race to Scale Data Centers | McKinsey & Company (2025)
   https://www.mckinsey.com/industries/technology-media-and-telecommunications/our-insights/the-cost-of-compute-a-7-trillion-dollar-race-to-scale-data-centers 

2. Data Center Demands and Capacity Forecasts | McKinsey & Company (2025)
   https://www.mckinsey.com/featured-insights/week-in-charts/data-center-demands 

3. World Investment Report 2025: International Investment in the Digital Economy | UNCTAD (2025)
   https://investmentpolicy.unctad.org/publications/1310/world-investment-report-2025-international-investment-in-the-digital-economy 

4. Global Investment in the Digital Economy Surges but Remains Uneven | UNCTAD (2025)
   https://unctad.org/news/global-investment-digital-economy-surges-remains-uneven 

5. The Investment Dimension of Digital Strategies | UNCTAD (2025)
   https://unctad.org/publication/investment-dimension-digital-strategies