Qualcomm’s 5% Pop Isn’t An AI Win—It’s A Costly Escape From The Smartphone Trap
(SeaPRwire) –
By: Reginald Vance
For three straight years, Qualcomm traded at a sector discount. Investors wrote it off as a pure smartphone cyclical play. It carried two persistent, value-crushing overhangs. First, slowing handset upgrade cycles capped core revenue growth. Second, it had no credible foothold in high-growth AI compute. That market is dominated almost entirely by Nvidia. Qualcomm shares moved in lockstep with Android shipment numbers. Any hint of Apple cutting modem orders sent shares tumbling. Analysts repeatedly warned the company had no durable second act. That long-held market panic shifted hard this week. The shift came right after Qualcomm’s annual Investor Day presentation.
The company dropped a string of long-awaited announcements. It named two marquee hyperscale customers for custom silicon. Meta will deploy its Dragonfly C1000 CPUs starting in 2028. Microsoft Azure will adopt its HBC chip architecture by mid-2027. CFO Akash Palkhiwala noted HBC fuses logic and memory. That design delivers high bandwidth at far lower power draw. Palkhiwala acknowledged the company is late to the data center market. He argued Qualcomm brings unique technical advantages to solve customer pain points. Each customer deal is projected to top $1 billion in revenue by fiscal 2027. Qualcomm nearly doubled its long-term non-handset revenue target. It now projects $40 billion in non-handset revenue by fiscal 2029. That is up sharply from the prior $22 billion guidance. Data center sales alone are expected to hit $15 billion that year. Handsets made up 72% of fiscal 2025 revenue. That share is expected to drop to one third by 2029. The company is putting cash behind that pivot. It announced a $3.9 billion acquisition of AI software firm Modular. Modular built an AI programming language positioned to rival Nvidia’s CUDA. Palkhiwala framed the resulting stack as fully open source. It will run on Qualcomm chips, as well as chips from competing vendors. Qualcomm also expanded its partnership with Hugging Face. The collaboration targets AI development across data center storage.
The market reaction was immediate and sharp. Shares jumped 12% in after-hours trading Tuesday. They rose more than 10% in premarket trade Thursday. Early trading gains settled at 5.3% as analysts weighed in. Price targets now span a massive $110 gap across firms. Benchmark leads the bull camp with a $300 target and Buy rating. It argues Qualcomm successfully shifted investor focus to AI infrastructure. Morgan Stanley upgraded shares from Underweight to Equalweight. It set a $231 target on the back of 2027 data center guidance. Skeptics remain firmly on the sidelines, even with higher targets. Susquehanna set a $190 target with a Neutral rating, citing mobile headwinds. Bank of America kept an Underperform rating with a $220 target. It notes current share prices already price in major data center wins. It flags unproven custom silicon ramp risks as a key threat. It also calls out Apple QTL renewal risk as a high-margin overhang. Cantor Fitzgerald set a $220 target with a Neutral rating. KeyBanc kept a Sector Weight rating, noting the roadmap is still early. The math here does not pencil out for a quick Nvidia upset. Qualcomm’s first major data center revenue is still three years out. Its CUDA rival will require years of developer buy-in to gain traction. Hyperscalers routinely spread custom chip orders across multiple vendors. None of these announced deals lock in exclusive long-term supply. The company is spending nearly $4 billion just to build a software layer. It will face brutal price competition from AMD, Intel, and AWS Graviton. Any delay to 2027 or 2028 product launches will crater current targets. Investors buying the pop today are paying for 2029 execution today. Trim positions on rallies until 2026 silicon samples ship to customers.
Author bio: Reginald Vance, venture partner focused on semiconductor valuation, with 15 years of experience assessing advanced compute hardware and materials supply chains.