AI infrastructure in focus as markets turn more selective

The recent pullback in equity markets, namely due to geopolitics, has not changed the broader AI story. If anything, it has reinforced where investors believe the strongest long-term demand sits. Institutional positioning has increasingly shifted toward the infrastructure layer of artificial intelligence, particularly semiconductors, networking systems, cloud capacity, and data centre-related businesses, focusing on companies already benefiting from rising compute demand and continued hyperscale spending. 

Why investors are focusing on AI enablers

As the AI theme evolves, markets are becoming increasingly focused on businesses powering the next phase of growth. This includes companies linked to semiconductors, networking equipment, cloud infrastructure, data centres,  and the broader hardware ecosystem required to support rising compute demand. 

Training and deploying large AI models requires significantly higher computing power, greater data storage capacity, and increased energy consumption compared to traditional digital workloads. As a result, infrastructure providers are becoming increasingly central to the AI investment cycle, viewed as key beneficiaries of the significant capital expenditure programmes currently being driven by hyperscale technology firms. 

Hyperscale capital expenditure remains a key driver

Recent earnings releases reinforced this trend, with major technology companies continuing to guide toward elevated AI-related spending. 

According to data tracking Amazon, Alphabet, Microsoft, Meta, and Oracle, hyperscaler capital expenditure as a percentage of operating cash flow has climbed from roughly 15% in 2012 to around 60% in 2025. What is more interesting to observe is the jump in investment since late 2023. This clearly highlights the scale of investment currently being directed toward AI infrastructure and computing capacity. Apart from funding through internally generated cash, hyperscalers have also tapped the bond market to expand their AI investments. 

The consistency of these spending trends continues to support the broader investment case for AI infrastructure, even as parts of the technology sector undergo valuation resets and increased volatility. 

For markets, this matters because ongoing hyperscale investment provides visibility into long-term demand across semiconductors, networking systems, cloud infrastructure, and data centre expansion. 

Large technology firms are competing aggressively to expand AI capabilities, secure computing capacity, and strengthen infrastructure ahead of the next phase of deployment.  

Rather than signalling a slowdown in the theme, the recent market weakness appears to have encouraged more selective positioning within it. 

Valuations are becoming more important again 

The recent pullback has re-introduced a stronger focus on valuations. Following the sharp AI-driven rally, parts of the market had begun trading on increasingly elevated expectations, and the retracement has helped moderate some of that positioning.  

This shift suggests markets are gradually moving toward a more disciplined assessment of which businesses are most directly positioned to eliminate the risk of monetisation, a fear which has hit selective names including some from the big tech. 

Why it matters 

“The current phase of the AI cycle is becoming increasingly focused on infrastructure deployment, capital expenditure visibility, and businesses positioned closest to rising computational demand. With monetisation concerns still present across parts of the AI landscape, companies benefiting directly from AI-related spending through the infrastructure and enabling layer may offer a comparatively lower-risk way to gain exposure to the theme” says Jordan Portelli, Chief Investment Officer at Calamatta Cuschieri Moneybase. 

FAQ 

Which specific sectors or company types are most exposed to the AI infrastructure theme? 

The most directly exposed businesses sit within semiconductors (chips required for model training and inference), networking equipment (high-speed interconnects between servers), hyperscale cloud providers (who both spend on and monetise compute capacity), data centre operators and REITs, and energy infrastructure companies given the significant power demands of AI workloads. 

How can a retail investor get exposure to AI infrastructure? 

Options include direct equity positions in semiconductor or data centre companies, thematic ETFs focused on AI or technology infrastructure, or diversified technology funds with meaningful infrastructure weightings. Investors should assess their risk tolerance carefully, as many of these names carry above-average volatility. 

How long is this capex cycle expected to last? 

Most market estimates suggest the current AI infrastructure build-out has a multi-year runway, driven by the scale of model development, enterprise AI adoption, and sovereign AI initiatives globally. However, the pace of spending is likely to become more uneven as companies shift focus toward demonstrating returns on that investment. 

What is the difference between AI enablers and AI adopters? 

Enablers are businesses that provide foundational infrastructure like chips, servers, networking, and cloud capacity that makes AI possible. Adopters are businesses integrating AI into their products or operations to drive efficiency or revenue. Enablers tend to benefit regardless of which AI applications ultimately win, making them a broader and arguably less risky way to gain exposure to the theme at this stage of the cycle. 

This information is issued by Calamatta Cuschieri Investment Services Ltd (“CCIS”) of Ewropa Business Centre, Triq Dun Karm, Birkirkara BKR 9034, Malta (C13729). CCIS is licensed to conduct Investment Services under the Investment Services Act in Malta by the Malta Financial Services Authority. The value of the investment may go down as well as up and may be affected by changes in currency. Any performance figures quoted refer to the past and past performance is not a guarantee of future performance nor a reliable guide to future performance.  

This information is being provided solely for information purposes and should not be deemed or construed as investment advice, advice concerning particular investments, advice concerning investment decisions, tax, legal, or any other ancillary regulatory advice. There is no guarantee that any forecast or opinion will be realized. The information presented does not take into account your personal circumstances and is provided to You on the express basis that it is not advice, and you may not rely upon it in making any investment decision. Investments in any financial instruments involve risks, you should make your own research before making any investment decisions and should seek the assistance of a financial advisor if in doubt. No person should act upon any opinion and/or information in this document without first obtaining professional advice. CCIS does not accept liability for actions, proceedings, costs, demands, expenses, damages, and losses suffered by persons as a result of information, views, or opinions appearing on this document.