Michel Lerner Predicts Only a Few AI Stocks Will Succeed in 20 Years

Michel Lerner, Head of HOLT, predicts that only a few AI stocks will succeed in 20 years. He estimates that only two out of ten value creators among AI companies will make it, highlighting the need for investors to be selective when investing in this field. Lerner believes that only a selected few technology and AI companies will be able to sustain their value in the long term.

Michel Lerner, Head of HOLT at UBS, has shared his insights on the future of AI stocks, predicting that only a few will succeed in the next 20 years. During a CNBC interview, Lerner emphasized that only a select few technology and AI companies will be able to sustain their value in the long term [1].

Lerner noted that while AI stock valuations have reached significantly more extreme levels compared to the dot-com peak on a price-to-book basis, today's companies demonstrate substantially higher earnings power than their predecessors. The technology sector is currently trading at approximately 40x earnings, the highest level since 2004 and about 1.5 standard deviations above the long-term median [1].

The increasing difficulty for companies to meet market expectations is evident in the recent earnings seasons. While Microsoft (MSFT) demonstrated successful execution by leveraging increased cloud demand driven by AI adoption, NVIDIA (NVDA) faced heightened scrutiny of its data center business, illustrating the elevated performance standards these companies must meet [1].

Lerner described the current market environment as a "wheat and chaff exercise," predicting a significant divergence between successful and unsuccessful AI investments. "We all know that there are AI stocks that are going to make it, are going to be magnificent, even if the valuation needs a reset. But we also know there’s going to be AI stocks that are just going to go by the wayside because their business model wasn’t robust enough," he said [1].

Historical data suggests investors should be highly selective when choosing AI investments for the long term. Lerner noted, "If we look back at our database, at any point in time, the top 10 value creators – and roll the clock forward 10 years – you will only find four of them still making the list. Roll the clock forward 20 years, and you only have two of them" [1].

The U.S. government is also making significant moves in the AI sector. The White House has been buying equity stakes in key AI supply chain companies, such as Intel (INTC) and MP Materials (MP), to secure strategic chokepoints and anoint national champions in AI [2]. This government intervention could create massive upside opportunities for investors who can identify the next companies to receive similar backing.

Microsoft (MSFT) has also signed a US$3 billion AI deal with the U.S. General Services Administration (GSA) to provide AI services at reduced prices, aiming to modernize the federal government's technology infrastructure and accelerate the adoption of AI tools [3]. This initiative is expected to generate up to US$3 billion in cost savings in its first year.

In conclusion, Michel Lerner's prediction of a few AI stocks succeeding in the next 20 years underscores the need for investors to be highly selective. The increasing valuations and elevated performance standards make it crucial for investors to focus on companies with robust business models. The U.S. government's strategic investments in the AI sector also present new opportunities for investors.

References:
[1] https://seekingalpha.com/news/4491215-only-a-few-ai-stocks-will-succeed-in-20-years-holt-s-michel-lerner
[2] https://investorplace.com/hypergrowthinvesting/2025/09/the-white-house-premium-how-washington-is-creating-ai-stock-rockets/
[3] https://mexicobusiness.news/cloudanddata/news/microsoft-signs-us3-billion-ai-deal-modernize-us-government