AI-linked stocks remain undervalued despite strong gains and heavy infrastructure spending, according to investment analysis provider Morningstar.
At the Morningstar Investment Conference UK 2026 held yesterday, Mike Field, chief European market strategist and Kenneth Lamont, principal, manager research, Emea, Morningstar, said that concerns over an AI bubble were overstated, with research suggesting valuations remain supported by revenue growth and earnings visibility.
Lamonth and Field pointed to rapid changes since AI was first discussed as an emerging investment theme two years ago, citing mass adoption of the technology, significant investment into AI infrastructure and growing clarity over how AI could reshape industries and portfolios.
A ke theme of the session was whether the AI rally still had room to run. Morningstar’s research has shown an index tracking the most frequently held AI-focused stocks globally was trading at around a 20% discount to fair value despite recent gains.
The discussion highlighted that the current AI cycle differed from previous technology booms because companies were already generating substantial revenues linked to AI adoption. Large-scale infrastructure spending — estimated at around half a trillion dollars this year — was described as significant but increasingly supported by visible earnings growth.
Research showed increases in projected data centre spending by major AI infrastructure firms including Nvidia, Broadcom, Marvell Technology, Advanced Micro Devices and Intel through 2026. Revenue growth forecasts for several companies exceeded 50%, with Broadcom projected at 93.1% and Nvidia at 73.8%.
Research covering 120 companies found around 20 had their economic moat ratings downgraded because of AI disruption. Areas such as cybersecurity, IT services, enterprise software and financial data were among sectors identified as facing pressure.
At the same time, disruption also created opportunities for firms with stronger exposure to core AI infrastructure and software.
A Morningstar framework ranked companies according to their “thematic exposure” to AI based on projected future revenues linked to the technology. Companies expected to derive more than half of their revenues from AI within five years received the highest scores. Among the highest-scoring firms highlighted were Snowflake, Oracle, Microsoft, Nvidia, Taiwan Semiconductor Manufacturing, Broadcom, MongoDB and Advanced Micro Devices.
AI’s “black box” funds and energy consumption raise concerns
Companies with the greatest exposure to AI were also among the most undervalued. Slides shown during the discussion indicated that 67% of “core AI” companies were categorised as undervalued, while none were considered overvalued.
The session also explored differences between equity research analysis and fund manager positioning. A consensus-based approach examining holdings across AI-focused funds globally showed Nvidia, Microsoft, Alphabet, Broadcom and Taiwan Semiconductor among the most commonly held stocks.
However, the analysis also found discrepancies between companies viewed by analysts as core AI businesses and those widely owned by AI funds. Some companies were described as potentially overlooked despite strong AI exposure, while others frequently held in AI portfolios were considered less directly tied to AI revenues.
Nearly all of the most commonly held AI stocks discussed during the session were US-listed companies, with limited European representation beyond firms such as ASML and SAP.
The presenters concluded that companies with the highest direct exposure to AI were also among the most attractively valued. They also said AI-linked equities could continue outperforming broader markets, citing continued infrastructure investment, earnings growth and improving revenue visibility across the sector.
