BlackRock: Investors Are Betting Big on Data Centre Energy

AI remains central to investor thinking, but BlackRock says capital is shifting away from the technology firms building AI platforms and towards the energy infrastructure required to power data centres at scale.

In its latest Investment Directions report, the asset manager surveyed 732 of its EMEA-based clients to gauge priorities for 2026. 

While AI growth is still widely expected, enthusiasm for large US technology firms has cooled. Only one fifth of respondents consider big tech a compelling opportunity for the year ahead, signalling a recalibration after the investment surge of 2025.

The findings point to a growing recognition that the most durable returns linked to AI may sit beneath the software layer, in the power generation and grid assets that support hyperscale data centres.

Energy demand reshapes data centre investment

More than half of respondents identified data centre energy as an attractive investment theme, with 37% preferring energy infrastructure over exposure to large technology companies. The shift reflects mounting awareness of how capital intensive AI has become.

Estimates from Gartner suggest around US$1.5tn was invested in AI during 2025, much of it concentrated among a small group of firms. For investors, this concentration raises concerns about risk and return, particularly as borrowing costs remain elevated.

Data centres sit at the centre of this debate. AI workloads require facilities that operate continuously, drawing vast amounts of electricity for compute and cooling. As clusters grow larger and denser, power availability has become a gating factor for expansion in several markets.

“It’s increasingly important to risk-manage megacap and AI exposure while also capturing differentiated upside opportunities,” says Ibrahim Kanan, BlackRock’s Head of Core US Equity.

This analysis frames energy not as a supporting asset class but as a primary beneficiary of AI-led digital infrastructure growth.

Data centres elevate energy providers

The scale of electricity required to support AI training and inference has transformed the role of utilities, renewable developers and grid operators. For data centre operators, securing long-term power contracts is now as strategic as acquiring land or network connectivity.

BlackRock expects AI-related electricity demand to become a structural component of global power consumption rather than a temporary spike. This outlook has pushed energy assets higher up the investment agenda, particularly those linked to reliable baseload and clean generation.

Renewable energy developers are seen as especially well-positioned as operators pursue sustainability targets alongside price stability. Long-term power purchase agreements allow data centre owners to manage cost volatility while addressing emissions obligations tied to increasingly stringent reporting standards.

Infrastructure constraints come into focus

Beyond generation, investors are also looking closely at the physical infrastructure that enables data centre growth. Transmission networks, substations and energy storage are emerging as critical bottlenecks in several regions.

BlackRock, Microsoft and NVIDIA have already announced plans for a US$100bn investment focused on AI data centres and power infrastructure, underscoring the scale of capital required. 

However, permitting delays and grid connection queues are slowing deployment in parts of Europe and North America.

For data centre developers, these constraints translate into longer lead times and higher upfront costs. From an investment perspective, they also highlight opportunities in grid modernisation and capacity expansion that sit outside traditional technology portfolios.

Big tech still matters, but priorities shift

Despite the pivot, large technology firms have not been written off. Companies such as Google, Meta, NVIDIA, OpenAI and Microsoft continue to feature prominently in BlackRock-managed funds due to their central role in AI software, chips and platforms.

However, the survey suggests investors are increasingly cautious about the heavy capital expenditure required to sustain AI leadership. 

The construction of data centres carries significant near-term emissions and cost burdens through steel, cement and energy-intensive manufacturing, while revenue gains may take longer to materialise.

This timing gap has implications for both financial performance and climate strategies, particularly for investors with shorter return horizons.

A more diversified AI investment thesis

Only 7% of respondents believe AI represents a market bubble, indicating confidence in the technology’s long-term impact. 

The shift towards energy and infrastructure reflects a more mature view of the AI ecosystem, one that extends beyond software and chips to the foundations that keep data centres running.

Private markets are expected to play a growing role in funding grid upgrades and energy projects that are difficult to access through public equities. 

For data centre operators and investors alike, the message from BlackRock is clear – the future of AI investment is increasingly tied to the power that makes large-scale compute possible.