Key Takeaways
- • Data-center exposure in closed-end private-capital funds has reached USD 122 billion in aggregate valuation as of Q3 2025, representing a significant institutional allocation across infrastructure, real-estate, and private-equity strategies.
- • Data-center investments in private markets have delivered annualized returns of 23.8% since 2011, substantially outperforming public and private-market benchmarks over the 15-year period.
- • Global data-center construction starts surged from approximately USD 60 billion in estimated completed value in early 2020 to USD 340 billion by 2025, reflecting explosive sector expansion driven by artificial intelligence and computational demand.
- • Thematic concentration risk persists despite apparent asset-class diversification: data-center holdings span hyperscalers, colocation operators, cloud providers, network infrastructure, and software services, but returns remain structurally dependent on a single variable—computational demand.
- • Asset owners, consultants, and general partners must reassess portfolio construction methodology, as traditional asset-class diversification may obscure hidden concentration exposure in the data-center ecosystem.
Summary
Data centers have emerged as a material allocation strategy across private markets, commanding USD 122 billion in aggregate valuation within closed-end funds as of Q3 2025, according to MSCI's analysis of 584 data-center companies and assets held across the MSCI Private Capital Universe. The sector encompasses five operational categories: hyperscalers, colocation and wholesale operators, cloud providers, network-infrastructure companies, and supporting software and service firms. This diversification across multiple investment structures—infrastructure funds, real-estate vehicles, and private-equity closed-end funds—has created the appearance of thematic diversification while concentrating systematic risk around a single demand driver.
Performance metrics underscore the asset class's institutional appeal. Data-center investments tracked by a broad index have generated annualized returns of 23.8% since 2011, significantly outpacing comparable public and private-market benchmarks over the 15-year window. Valuation acceleration intensified after 2020, coinciding with OpenAI's May 2020 launch of GPT-3 and November 2022 release of ChatGPT. This performance trajectory reflects both structural tailwinds from rising computational demand and cyclical momentum from artificial intelligence spending cycles.
The underlying market fundamentals reveal extraordinary capital deployment. Global data-center construction starts—measured by estimated completed value—escalated from roughly USD 60 billion in early 2020 to approximately USD 340 billion by 2025, demonstrating a five-fold expansion over a five-year period. This acceleration reflects acute scarcity in operational capacity relative to demand from hyperscalers, enterprise cloud migration, and emerging AI infrastructure requirements.
However, MSCI's analysis identifies a critical structural vulnerability: apparent diversification across asset classes obscures thematic concentration risk. While data-center exposure is distributed across infrastructure funds, real-estate vehicles, and private-equity strategies, all positions ultimately derive returns from a single underlying variable—computational demand. A sustained contraction in AI capex cycles, shifts in architectural preferences favoring distributed computing, or saturation of data-center capacity could simultaneously impair performance across nominally uncorrelated fund vehicles. Asset owners, consultants, and general partners are advised to explicitly stress-test portfolio exposure to data-center demand scenarios rather than rely on traditional diversification metrics.
Performance metrics underscore the asset class's institutional appeal. Data-center investments tracked by a broad index have generated annualized returns of 23.8% since 2011, significantly outpacing comparable public and private-market benchmarks over the 15-year window. Valuation acceleration intensified after 2020, coinciding with OpenAI's May 2020 launch of GPT-3 and November 2022 release of ChatGPT. This performance trajectory reflects both structural tailwinds from rising computational demand and cyclical momentum from artificial intelligence spending cycles.
The underlying market fundamentals reveal extraordinary capital deployment. Global data-center construction starts—measured by estimated completed value—escalated from roughly USD 60 billion in early 2020 to approximately USD 340 billion by 2025, demonstrating a five-fold expansion over a five-year period. This acceleration reflects acute scarcity in operational capacity relative to demand from hyperscalers, enterprise cloud migration, and emerging AI infrastructure requirements.
However, MSCI's analysis identifies a critical structural vulnerability: apparent diversification across asset classes obscures thematic concentration risk. While data-center exposure is distributed across infrastructure funds, real-estate vehicles, and private-equity strategies, all positions ultimately derive returns from a single underlying variable—computational demand. A sustained contraction in AI capex cycles, shifts in architectural preferences favoring distributed computing, or saturation of data-center capacity could simultaneously impair performance across nominally uncorrelated fund vehicles. Asset owners, consultants, and general partners are advised to explicitly stress-test portfolio exposure to data-center demand scenarios rather than rely on traditional diversification metrics.
Go deeper than the headline
Research every firm in this story, track the ones that matter and get the daily briefing — free for 7 days.