Over the summer a group of our analysts and portfolio managers spent a week in San Francisco. The trip was an integral part of our ongoing research into artificial intelligence (AI) and related investment opportunities. Our scale as investors affords us exceptional corporate access and during our time in San Francisco we met senior leaders from around a dozen different technology companies, from heavyweights like NVIDIA and Broadcom to names operating throughout the data supply chain.
The week-long tour reinforced our bullish outlook on AI investments, and we see scope for strong returns within the AI supply chain.
The AI runway is long and wide
- Limited AI literacy – there is acknowledgment that senior management teams within businesses often have a limited understanding of use cases for AI, as well as concerns around adoption costs. Salesforce mentioned such challenges in our meeting with them, and it is likely that AI providers will have to work closely with customers on driving implementation.
- Untested pricing models – both Salesforce and ServiceNow discussed having to adjust how clients are charged for the incremental costs associated with new AI products. Many new models are anchored around consumption-based pricing, but customers can be less comfortable with a ‘pay as you go’ approach due to a lack of up-front clarity around overall costs.
- Power constraints – energy demand has emerged as the primary bottleneck in scaling AI technologies. This in turn creates challenges for expanding data centers and implementing AI solutions at scale. As a result, adoption may stall as companies search for more efficient computing solutions, including the development of better power infrastructure and more efficient technologies.
Sector spotlights
The implications of AI’s rollout range far beyond directly related sectors and businesses. As a result, our research project involves sector specialists ranging from utilities to industrials, energy, real estate, materials and more.
Utilities: AI’s shifting of energy dynamics
Sean Lenahan, Senior Equities Analyst, and Mary Titler, Senior Fixed Income Analyst
AI is reshaping power demand dynamics, with data centers emerging as a major driver of electricity consumption. Our forecasts suggest that by 2030 data centers could account for more than 20% of total US power use, with AI-specific facilities surpassing non-AI centres by 2027.
This surge presents both challenges and opportunities for the utilities sector. Utilities stand to benefit from incremental load growth, and we see several names that are well positioned. However, the sector must also navigate infrastructure constraints and long lead times for new generation and transmission, as well as growing local opposition to data centre siting. We see compelling investment opportunities in select vertically-integrated utilities, which provide increased certainty to data centers looking to procure power. The ability to efficiently manage peak demand and secure long-term contracts with hyperscale customers will be critical to unlocking value in this evolving landscape.
Industrials: Cooling demand heats up with AI infrastructure
Courtney Yakavonis, Senior Equity Analyst and Niranjan Aiyagari, Equity Research Analyst
The management of thermal loads is key within AI-driven data centers. This is because the compute power required for developing and running AI models generates substantial heat. Cooling demand is expected to grow from $10 billion in 2024 to $60 billion by 2030. As rack densities rise – reaching 300kW–500kW by 2028 – traditional air cooling becomes insufficient, which is prompting a shift to liquid cooling. Liquid cooling not only improves energy efficiency but also enables higher compute density, making it central to AI infrastructure scalability and sustainability. As such, the liquid cooling market, currently valued at $2-3 billion, is forecast to expand 10-fold, capturing 40% of data center cooling spend by 2028. Technologies such as direct-to-chip and immersion cooling are gaining traction, with coolant distribution units (CDUs) emerging as critical components. We see several plays with advanced solutions and M&A is increasingly a theme as leading names seek to broaden their portfolios.
Looking ahead: The AI investment landscape
After a week in San Francisco meeting management teams, our trip reinforced our bullish assessment on AI as a theme set to drive innovation and growth across a range of industries. The company visits helped enrich our investment thesis with more nuance. The potential of AI was evident, and we saw many examples of how technologies are rapidly evolving business practices. We also explored developments like agentic AI – an evolution of autonomous capabilities with potentially transformative impacts within company operations.
At the same time, we learned more about the potential challenges associated with AI’s rollout –including power consumption, pricing models and organizational readiness. However, given the breadth and depth of AI applications witnessed and discussed on our trip, we would suggest that these factors represent growth hurdles as opposed to structural barriers.
Significant investment in related infrastructure, as well as rapid data growth, reinforce our conviction around momentum in AI. From an investment perspective we see select opportunities in technology leaders such as NVIDIA and Microsoft, but also across the broader ecosystem of enablers and solutions providers.
