The AI trade is getting bumpy as investors question whether the massive spending on data centers and chips will generate sufficient returns. There is only so much capital companies can raise, and cracks are beginning to appear. Bonds issued by AI cloud provider CoreWeave have come under pressure, while shares of the company and other AI-related names, including Dell Technologies, Micron and Advanced Micro Devices, have suffered sharp declines amid concerns about rising borrowing costs and the sustainability of AI demand.
Yet some prominent investors believe the market is underestimating how much further the AI boom can run.
Speaking to CNBC earlier this week, Altimeter Capital founder and Chief Executive Brad Gerstner said investors have spent the past two years worrying about an eventual glut of AI computing capacity, only to miss substantial gains in semiconductor stocks. He believes the boom still has room to run because physical constraints, including chip production capacity and electricity supply, are limiting how quickly companies can expand AI infrastructure.
Gerstner is one of the loudest AI bulls on Wall Street, and his fund keeps buying AI stocks. Check his top 10 stock picks here. His latest filings show new additions too. Read his two new AI picks here.
“I actually think you have a natural constraint here,” Gerstner said on CNBC. “There’s only so fast in terms of production of wafers in this country. It’s only so fast that we can bring power on.”
HSBC’s Willem Sels also believes investors are underestimating the AI opportunity. In a September interview with Bloomberg Television, the bank’s global chief investment officer said US stock valuations did not fully reflect the potential for AI-driven productivity gains and earnings growth. He particularly highlighted semiconductor companies, whose valuations he believes reflect skepticism about future earnings forecasts despite the possibility of continued strong demand.
“The US is not expensive. The markets are questioning the sustainability of earnings growth, but that’s in the price because that gap has closed,” Sels told Bloomberg.
So, what are the best undervalued AI stocks to buy? We answered that question in an article titled “10 Best Stocks To Buy for High Returns“. You can find out what other media companies think below.
Methodology
We based our research on the most undervalued AI stock on at least 13 major finance sites to see what AI stocks they consider undervalued and analyzed the reasons they gave for their claims.
The ranking itself uses one measure: how many times a stock was called undervalued in the pieces we reviewed. The more often a stock came up, the higher it ranks. We counted each article once.
#10 Marvell Technology, Inc. (NASDAQ:MRVL)
Marvell Technology, Inc. designs custom AI chips and networking silicon for data centers. A notable research site says the stock trades 37% below its fair value estimate and expects the company to keep competing at the cutting edge against Broadcom and Nvidia. Other research from earlier this year argued that the market priced Marvell like a legacy connectivity company and ignored its custom AI chip business.
One community-driven valuation site puts fair value about 42% above the share price, and another site rates the stock a buy with about 44% upside to its target. Sites do not agree on how cheap it is. One piece noted a forward P/E of 51 after a 130% gain this year, and another flagged softer hyperscaler spending as the biggest threat. Marvell’s earnings calls have a history of moving the stock, and one of them looked like it could go badly. Check what management said to calm investors here.
