Qualcomm (NASDAQ: QCOM) has surged roughly 50% since mid-April, signaling a potential turning point in how the market values its core technology advantages.
For three years, the AI investment rally concentrated almost entirely on data center chipmakers, leaving Qualcomm to trade as a smartphone company tied to a slowing upgrade cycle.
Nvidia soared, AMD rallied, and data center hardware captured practically every dollar of AI-driven capital flows, while Qualcomm’s broader potential went largely unrecognized by investors.
That dynamic now appears to be shifting, as AI workloads begin migrating from centralized cloud infrastructure onto individual devices at the network edge.
Power efficiency, connectivity, and low-latency compute, the exact capabilities Qualcomm has spent decades refining, are quickly becoming the defining requirements of this next AI phase.
Today’s AI ecosystem runs on centralized compute concentrated in data centers, but routing every inference through the cloud is costly, slow, and power-hungry at scale.
Local inference, running AI models directly on devices rather than in the cloud, is faster, more private, and does not depend on a live network connection.
Qualcomm’s Snapdragon platforms already sit inside smartphones, PCs, and a growing number of vehicles, and its automotive business alone represents a $45 billion design-win pipeline.
Newer platforms like Dragonwing, alongside partnerships across ecosystems including Arduino, are extending Qualcomm’s reach into robotics and industrial AI applications.
Qualcomm is effectively evolving from a handset component supplier into a broader compute platform for connected, intelligent devices across multiple industries.
AI spending itself is also shifting from training to inference, the process of running trained models in production environments where efficiency matters more than raw performance.
Inference is projected to reach two-thirds of total AI compute demand by 2029 and represent between 80 and 90% of an AI system’s lifetime operating cost.
Qualcomm’s chips are built around power efficiency and cost per query, a design philosophy that is growing more valuable as AI infrastructure runs into real-world power grid constraints.
A large data center can be constructed in 12 to 24 months, yet securing a high-capacity grid connection in key U.S. markets can take between 36 and 84 months.
Most AI accelerators from Nvidia and AMD rely on a specialized packaging technology called CoWoS, which combines computing chips with high-bandwidth memory in extremely short supply.
Nvidia has reportedly secured more than half of TSMC’s CoWoS capacity through 2026, leaving competitors to compete aggressively for what remains of the constrained supply chain.
Qualcomm’s AI200 chip uses LPDDR5X memory instead of high-bandwidth memory, allowing it to operate on a different supply chain and sidestep one of the industry’s most critical bottlenecks.
On the financial side, Qualcomm generated approximately $44 billion in revenue in fiscal year 2025, with consensus pointing toward around $42.6 billion in fiscal year 2026 due to memory shortages and Apple’s modem transition.
If revenue accelerates at roughly 15% annually, driven by AI, CPU chips, and automotive, sales could reach $65 billion by 2029 under a scenario that assumes continued execution on these growth vectors.
Holding net margins at approximately 25%, just above trailing twelve-month levels, would produce around $16 billion in annual net income if those revenue targets are achieved.
With share repurchases likely to continue, the share count could trend down from approximately 1.07 billion to around 950 million by 2029, pushing earnings per share toward roughly $17.
The broader semiconductor sector currently trades at more than 35 times forward earnings, while Qualcomm trades at just approximately 17 times forward earnings, representing a significant valuation discount.
Even a modest 22 times multiple, which would reflect Qualcomm’s AI growth potential while still discounting the stock relative to high-flyers, implies a share price of approximately $370.
That target would represent nearly a doubling of current price levels, driven by a combination of earnings growth and multiple expansion as the edge AI narrative gains wider market acceptance.
