The AI security equation: Power, resources and public trust

AI’s economic engine meets strategic imperatives

Three years after the launch of ChatGPT in November 2022, its transformative potential for businesses and the economy itself is becoming clear. Markets reflect this: AI-related stocks have since accounted for 65%-75% of S&P 500 returns, profits, and capital spending.¹

At present, AI’s energy needs far outstrip supplies, but the race to power AI is no longer just an economic story: It’s a national security imperative. Assets tied to AI’s growth are strategic priorities in the contest for global competitiveness.

As the power supply gap has garnered attention, clean energy has come back into focus, and solar and nuclear power companies have outperformed. The next wave of AI investment will be shaped not just by technological innovation, but by security and sustainability considerations—where reliable power, access to critical minerals, and public trust are both catalysts and constraints. 

The “speed to power” push will benefit renewables

As shown by headlines, and our own recent outlook, AI data centers have massive power needs as they scale up. The Department of Energy estimates an additional 100 gigawatts (GW) of peak capacity is needed by 2030,² half of which is attributable to data centers. However, the U.S. already faces aging infrastructure and strained power markets, risking blackouts at times of high demand.

“Speed to power” is now an urgent priority. We think markets will follow an “all of the above” approach, securing power from multiple sources. As this chart shows, each renewable source has distinct benefits and drawbacks based on cost, how quickly it can be ramped up, carbon emissions and consistent power generation. In the near-term, suppliers who can quickly add power generation and accelerate grid connections are positioned to secure premium, long-term contracts.

The market is rewarding all power sources. Nuclear power stole the spotlight with billion-dollar hyperscaler deals and executive orders to reinvigorate the industry, as it is well-suited for 24/7 data center workloads. However, we see some nuclear companies as overbought: There are a limited number of existing reactors that can be restarted, new builds are often plagued by cost overruns and delays, while next-gen reactors are far from scaling.

Meanwhile, solar companies outperformed the broad global index by over 22%,³ supported by the need for power and improving company fundamentals. We have more positive views on commercial-scale renewables with relatively lower valuations, boosted by catalysts like policy clarity, project deployment timelines and cost-competitiveness.

The “power couple:” Renewables and natural gas

U.S. political headlines may say that clean energy is dead, but markets say otherwise: An active clean energy fund manager has outperformed their broad global index by over 6% through October 2025.⁴ Valuations, recovering from multi-year declines, and project fundamentals drive the upside.

Natural gas and renewables pair well together: Solar and wind projects are the quickest and cheapest to develop because of improving technologies and economies of scale, but at present, that power is only available when the sun is shining or the wind is blowing. In the U.S., natural gas is cost-effective⁵ and can ramp up when renewables are unavailable. Extended delivery times and rising costs for gas equipment may push developers towards other solutions, with battery adoption gaining due to falling costs⁶. The economic advantage is clear – in the U.S., renewables and batteries are expected to make up 99% of all new power added in 2026, with battery additions 50% higher than last year.⁷

Given these market dynamics, we see renewables, battery storage, and grid solutions well-positioned over the next 12 months. We think U.S. and European utilities and power infrastructure funds are the best way to access this theme in both public and private markets.

Utilities are raising their earnings per share (EPS) growth guidance, reflecting higher capex and data center deals, while valuations look inexpensive relative to the sector’s history and the current market.

Private infrastructure funds are another beneficiary of AI-induced demand. Surging investment in transmission upgrades, power assets, and energy storage is driving secular growth in an asset class offering stable income and inflation protection.

Copper and water, face supply constraints

Critical minerals (e.g., copper) are key to building power and grid assets, part of the contest for technological and economic leadership. Copper is needed for transmission and renewables, lithium for batteries, and rare earths for energy and defense. We think copper is particularly critical in an energy security context, which benefits miners, especially those exposed to North American and South American production.

Copper is the linchpin of electrification, but supplies are projected to fall well short of demand. This should keep copper prices elevated and incentivize investment activity. Miners would have to invest around $350 billion between now and 2040 to meet rising copper demand.⁸ This might be a big ask, however, as the industry faces limited new discoveries and constraints on expanding existing sites. Instead, we’re monitoring potential M&A activity between miners with strategic copper resources.

Water, too, is rising in strategic value: It’s essential for both chip manufacturing and data center cooling, yet aging infrastructure, community opposition, and climate-driven scarcity pose risks to project execution. These constraints should drive adoption of water-efficient innovations, but also reward developers with a proven record of community engagement.

Spending on data center cooling is poised to grow 18% annually, and projects to become a $16.9 billion market by 2028.⁹ Industrials selling efficient cooling solutions, reducing the electricity burden of AI infrastructure, are well-positioned.

Affordability

Data centers use vast amounts of power, straining the grid, and some consumers are already seeing this in their electricity bills.¹⁰ Data center servers now consume more than one-fifth of Ireland’s power, and as of October 2025, energy prices had spiked 26% in one year—the second-largest increase across EU countries.¹¹

In Virginia, home to “Data Center Alley,” electricity prices are up 13% from last year and about 30% from 2021 levels. Some governments are taking steps to protect retail consumers, such as large load tariffs for hyperscalers. Still, these interventions may not cover the full costs of additional generation and transmission, and may struggle to reverse the public’s negative perception of data centers.¹²

Communities have been pushing back, citing concerns about higher electricity costs and increased water usage. This pressure led Google to cancel a project in Indianapolis in 2025, while Amazon pulled out of a proposed data center in Tucson. We expect to see AI-linked affordability debates in the 2026 midterm elections, and for leading tech companies to try to secure public trust through community engagement.

Corporate adoption and the impact on jobs

AI’s impact on the workforce seems both empowering and uncertain: It could revolutionize access to information, but may also disrupt career paths. We expect job rotations and enhancements over the near term, but long-term impacts on job creation will likely take decades to materialize. Still, we think investors should focus on emerging trends that could be leading indicators of change.

It’s early days, but we’re monitoring increasing unemployment in AI-influenced roles (e.g., software developers), particularly amongst younger workers. Workers under 30 in technology-related occupations have seen unemployment climb 3% just this year, strongly outpacing that of older industry colleagues as well as young peers in other occupations.¹³

We see opportunity for education technology platforms, as they are positioning themselves to personalize learning, improve outcomes and match candidates to jobs—a market forecasted to more than double by 2030 and reach $348 billion.¹⁴

The need for cybersecurity

AI is making an impact on cybersecurity—both good and bad. Generative AI tools are improving efficiency for bug fixes and threat detection, while AI agents and model governance create business segments for enterprises. While benefits abound, these tools also open up a new threat landscape. Anthropic recently reported the first instance of an AI-orchestrated cyber espionage campaign¹⁵. Moreover, the broader social impacts of deepfakes and misinformation are still being assessed. Threats to public trust and critical infrastructure only increase the value of transparency and protection. While the software sector faces uncertainty, cybersecurity spend should increase as AI penetration deepens.

Our conclusion

For AI investors, these findings present potential risks, but also open opportunities for first movers. Companies that address concerns head-on with responsible governance will likely win market share as public support for transparency, ethics and regulation gains momentum.¹⁶

Interest in AI and adoption of the technology have grown rapidly in three years, and we strongly expect that trend to continue. However, the AI infrastructure boom is a double-edged sword, driving economic growth while exposing potential bottlenecks— investment areas to watch. We think investors should focus on power and critical resources as strategic assets, while societal impacts are increasingly important to the growth of AI infrastructure.