Could Qualcomm Stock Be the Best Data Center Investment Here?

With a dividend yield of 2.6%, Qualcomm (QCOM) may be one of the most undervalued artificial intelligence stocks to buy today. All as it pushes into the AI data center market, focusing heavily on inferencing (running AI models), directly challenging Nvidia’s dominance. Fueling further upside potential, the company’s acquisitions of Alphawave Semi and Ventana Micro Systems are seen as strongly supporting Qualcomm’s market competitiveness.

All of which helps expose QCOM to a potential $100 billion AI inference market opportunity, say analysts at Wells Fargo. Yet, despite the news, Qualcomm is trading at 25% below its recent high. It’s also trading at a discount to its peers in the semiconductor market. Plus, analysts at Wells Fargo argue that a potential $100 billion opportunity is not priced into Qualcomm.

Wells Fargo also argues that Qualcomm could capture $5 billion to $7 billion in annual revenue by 2027, based on its participation in that market. It’s why Wells Fargo now has an “Equal-Weight” rating on the QCOM stock with a price target of $150. “We remain focused on Qualcomm’s evolving data center strategy with the integration of its recent acquisitions, Alphawave Semi and Ventana, which we see as supportive of its AI-series solution strategy, as well as expanding its ability to offer custom solutions/IP to hyperscale customers,” said Wells Fargo analysts, as quoted by Seeking Alpha.

In short, we have a massive opportunity that’s not being priced into the stock yet, plus incredible undervaluation as compared to peers. All screaming, “Buy me.”

In an attempt to take market share from Nvidia, Qualcomm will release its AI200 this year and its AI250 chip-based acceleratory cards by 2027. Both are focused on inference, or running AI models, and not training, which teaches a model using large data sets.

In addition, as noted by CNBC, “Qualcomm said its AI chips have advantages over other accelerators in terms of power consumption, cost of ownership, and a new approach to the way memory is handled. It said its AI cards support 768 gigabytes of memory, which is higher than offerings from Nvidia and AMD.”