Qualcomm built its empire selling chips to smartphone makers. Now it wants to sell chips to the companies building AI supercomputers.
At its Investor Day on June 24, 2026, the company laid out a target to grow data center revenue from roughly $0.3 billion in fiscal year 2026 to more than $15 billion by fiscal year 2029. That is a 50x increase in three years.
What Qualcomm actually announced
Qualcomm has a multi-generation agreement with Meta for data center CPUs, which gives the target some structural grounding beyond aspirational math.
The company also disclosed progress on custom silicon deals with other hyperscalers, including an accelerator developed under a collaboration called the HUMAIN project. Qualcomm is pitching a product line called the Dragonfly portfolio, built for what it describes as agentic AI data centers. The roadmap includes the AI200 inference accelerator in 2026 and the AI250 following in 2027, with power efficiency as the primary selling point over raw performance.
The automotive numbers got their own headline. Qualcomm’s auto design-win pipeline expanded to $65 billion, with a $10 billion revenue target in that segment by FY29.
Zooming out, Qualcomm raised its overall non-handset QCT segment revenue target to $40 billion by FY29, roughly double its prior target. Mobile handsets are now expected to account for approximately one-third of QCT revenue going forward.
CEO Cristiano Amon framed all of this around the concept of a compute continuum, the idea that AI processing now spans from cloud data centers down through vehicles, industrial equipment, and edge devices, and that Qualcomm’s architecture is designed to operate across all of it.
Why this matters beyond the revenue targets
Qualcomm is staking a claim to a combined total addressable market it estimates at approximately $1.7 trillion by 2030, spanning data centers, automotive, robotics, industrial AI, and edge AI.
Qualcomm’s non-GAAP EPS target exceeds $18 by FY29, which gives analysts a concrete profitability benchmark to model against.
For investors watching the stock, the key variable is execution cadence. The AI200 lands in 2026, which means the first real data point on whether Qualcomm can compete in this market arrives within the current fiscal year.
The automotive pipeline is arguably the lower-risk leg of the thesis. At $65 billion in design wins, even modest conversion rates translate into material revenue. Vehicles have long development cycles, which means the pipeline is sticky in a way that software deals are not.