Companies are following the ‘Invest in AI now, ask the ROI question later’ – Susquehanna’s Mehdi Hosseini

[Robotic hand pointing at a candlestick chart of stock price movement in dark background. Illustration of the concept of AI tools for picking and screening profitable securities and maximizing ROI]
Dragon Claws

Mehdi Hosseini, senior equity research analyst at Susquehanna, believes companies are following a “spend now, ask ROI later” approach in artificial intelligence investments, suggesting this trend could continue longer than previously anticipated.

In an interview with CNBC, Hosseini noted that questions about return on investment for massive AI spending are being pushed out to “sometime in '27” rather than 2026 as he had initially predicted.

Unlike previous cycles that were driven by distributors and OEMs (original equipment manufacturer), the current AI boom features major tech companies like NVIDIA (NVDA [https://seekingalpha.com/symbol/NVDA]), AMD (AMD [https://seekingalpha.com/symbol/AMD]), and Broadcom (AVGO [https://seekingalpha.com/symbol/AVGO]) as the primary buyers.

“This is why this cycle is sustainable throughout [2025-2026], and it may even sustain into 2027,” Hosseini said, highlighting the unusual durability of the current market conditions.

The concentration of buying power among major AI companies is creating unique market dynamics throughout the supply chain. When questioned about whether this customer concentration makes the cycle more vulnerable, Hosseini countered that if NVIDIA (NVDA [https://seekingalpha.com/symbol/NVDA]) isn’t being penalized for its own customer concentration, then “Micron (MU [https://seekingalpha.com/symbol/MU]) shouldn’t be penalized for that because that’s the downside risk to the overall AI spending.”

For memory manufacturer Micron (MU [https://seekingalpha.com/symbol/MU]) specifically, Hosseini sees significant upside potential. “What investors haven’t realized is not only supply is going to remain tight, the next generation of AI chip – Rubin (a new class of GPU purpose-built for massive-context processing) – is going to require even more advanced DRAM (dynamic random-access memory), and Sanjay [Mehrotra, Micron’s CEO] is going to be able to extend margins towards 55% or high 50s,” he said, projecting Micron Technology (MU [https://seekingalpha.com/symbol/MU]) could reach $200 per share.

Despite his optimism, Hosseini acknowledged potential risks with the vendor financing model emerging in the AI space.

“Vendor financing is very concerning… I think at some point, this is going to backfire,” he cautioned, drawing parallels to similar patterns he observed in the solar industry in 2008-2009.

However, he believes any significant downturn has been delayed: “Spend is spent, and then you ask the ROI question later.”

MORE ON AI:

* AIEQ: AI's Unfulfilled Promise [https://seekingalpha.com/article/4821225-aieq-ais-unfulfilled-promise]
* Capturing AI's Upside: ARTY's Concentrated Play On Tomorrow's Winners [https://seekingalpha.com/article/4818062-capturing-ai-upside-arty-concentrated-play-on-tomorrows-winners]
* ARTY: A Generational Investing Theme [https://seekingalpha.com/article/4810858-arty-etf-generational-investing-theme]
* Goldman Sachs sees potential GDP growth accelerating at least 2.1% due to AI productivity [https://seekingalpha.com/news/4497967-goldman-sachs-sees-potential-gdp-growth-accelerating-at-least-2_1-percent-due-to-ai]
* Only a few AI stocks will succeed in 20 years – HOLT’s Michel Lerner [https://seekingalpha.com/news/4491215-only-a-few-ai-stocks-will-succeed-in-20-years-holt-s-michel-lerner]