Smartphone chip specialist Qualcomm<\/strong> (QCOM<\/a> +9.10%<\/span>)<\/span> has underperformed the broader semiconductor sector over the past three years, gaining 58% during this period, compared with the 258% gains clocked by the PHLX Semiconductor Sector<\/strong> index.<\/p>\n Qualcomm’s underperformance isn’t surprising. While several semiconductor stocks have benefited substantially from the booming demand for artificial intelligence (AI) chips, Qualcomm’s reliance on the smartphone market for the majority of its revenue has been a headwind for the stock.<\/p>\n However, Qualcomm stock popped impressively last week<\/a> following the release of the company’s latest quarterly results on April 29. Let’s see why that was the case and check why this semiconductor stock<\/a> is an underrated gem that could be one of the best ways to capitalize on the AI sector’s growth in 2026.<\/p>\n Image source: The Motley Fool<\/p>\n<\/div>\n Qualcomm’s results for its fiscal 2026’s second quarter (ended March 29) weren’t great. Its revenue and adjusted earnings declined year over year. The company saw a significant 13% year-over-year decline in its smartphone revenue to $6 billion, which isn’t surprising given the prevailing softness in this market.<\/p>\n However, Qualcomm management’s comments about a potential breakthrough in the AI data center chip market have boosted investors’ confidence about its prospects. Responding to a question on the latest earnings call<\/a>, Qualcomm CEO Cristiano Amon noted that there is strong interest in the company’s custom application-specific integrated circuits (ASICs).<\/p>\n Qualcomm is engaging with several companies for its custom AI processors. Importantly, a large hyperscaler is reportedly in talks with Qualcomm to develop multiple generations of custom AI processors. It’s worth noting that Qualcomm has been pushing the envelope in the AI chip market for some time now, and it announced a breakthrough in this space last year<\/a>.<\/p>\n It now appears that the growing tilt toward AI inference applications, which don’t require as much computing horsepower as the model training phase, is a tailwind for Qualcomm. The company has already been pushing the envelope in local AI processing in edge devices<\/a>, and the potential gains in custom AI processors and inference-focused accelerators could unlock a whole new growth opportunity for Qualcomm.<\/p>\n Management indicated that it has lined up a few customer wins for its data center chips and AI inference accelerators and will share the details at its Investor Day next month. So there’s a good chance that Qualcomm’s newfound momentum is here to stay.<\/p>\nQualcomm could be on the verge of a breakthrough in AI chips<\/h2>\n