Mark your calendars for June 24th—this was the day Qualcomm decided to stop playing it safe. In a packed event in New York, CEO Cristiano Amon didn’t just update us on the usual smartphone numbers; he laid out a massive, dual-engine strategy for AI data centers and automotive tech. The headline? They’re officially launching the Dragonfly series of data center CPUs and AI accelerators, and they’ve just forked over nearly $3.9 billion to snap up AI infrastructure software giant Modular. Let’s be real: this is the biggest strategic shake-up in Qualcomm’s entire history.
During the investor day presentation, NUPIAO got a first look at the Dragonfly family. We’re talking about the Dragonfly C1000 data center CPU, the next-gen AI300 accelerator, and some serious high-speed interconnect solutions. Basically, they’ve built a full-stack product matrix that covers everything from general computing to AI acceleration and networking. They aren’t just dipping a toe in the water anymore; they are diving straight into the deep end of the AI compute race dominated by NVIDIA.
Here’s the cool part about the C1000 CPU: it’s built on the RISC-V architecture they acquired from Ventana. It’s all about energy efficiency. If you’re running AI inference or cloud-native workloads, this chip delivers way more compute density than traditional x86 architectures for the same power bill. Then there’s the AI300 accelerator. It’s designed specifically for inference tasks using a “near-memory” computing architecture. This means you don’t need those expensive, high-bandwidth graphics cards. For large language model (LLM) inference and multi-modal processing, the performance-per-watt advantage is just insane.
Cristiano Amon put it perfectly during his speech. He pointed out that right now, the biggest headache for building AI data centers isn’t even the cost of the chips—it’s power. Cloud giants are hitting a wall with electricity supply. But here’s where Qualcomm shines: after two decades of squeezing every milliwatt out of smartphones, they’ve mastered extreme power control. That experience? It’s becoming their secret weapon in the data center market.
But wait, there’s more. Qualcomm also bumped up their long-term growth forecast for their automotive division. The numbers are staggering. Design orders for their car business have ballooned to $65 billion, climbing steadily from last year. By fiscal year 2029, they’re aiming for $10 billion in revenue from cars alone. Their Snapdragon Digital Chassis solution now covers everything from connectivity to smart cockpits and advanced driver-assist systems. Almost every major car manufacturer has a model on the road right now running Qualcomm tech. It’s their fastest-growing segment, racking up a 43.8% compound annual growth rate between 2020 and 2025, with revenue hitting close to $4 billion in FY2025.
Amon drove the point home: Cars and Data Centers are the twin pillars holding up Qualcomm as they move away from relying solely on phones. Thanks to this explosive growth, they hiked their non-phone revenue guidance for FY2029 from $22 billion to a whopping $40 billion—an increase of nearly 91%. Looking further ahead, they believe the total addressable market could hit $900 billion by 2030.
Of course, you can’t just talk big; you need customers. Qualcomm announced some serious partnerships to back up their new products. Mark Zuckerberg himself jumped on a video call to confirm a multi-generation strategic deal. Qualcomm is set to become the core CPU supplier for Meta’s future data centers, working together to optimize energy efficiency. Meanwhile, Microsoft Azure has already started deploying Qualcomm’s custom high-bandwidth compute chips. Plus, two other global hyperscale cloud providers have signed contracts for custom chips, with revenue starting to flow in before the end of 2026.
The financial outlook is aggressive but promising. According to their guidance, the data center business could start churning out tens of millions in revenue as early as FY2027, with actual custom chip sales kicking off in Q1 of that year. The mid-to-long-term goal? Shattering $15 billion in revenue for this segment by FY2029. Honestly, that crushes what most analysts were expecting.
To round out the puzzle and fix their software ecosystem gap, Qualcomm made a huge move: acquiring Modular. They’ve finalized an agreement to buy the AI software company. The deal involves issuing up to 19.2 million shares to Modular shareholders, valuing the whole transaction at roughly $3.92 billion. That figure includes a $300 million incentive plan for employees. Barring any regulatory hiccups, the deal should close in the second half of 2026.
For those who don’t know, Modular was founded in 2022 by Chris Lattner, the genius behind LLVM and Swift. Their flagship product, Mojo, along with their MAX inference platform and AI compiler toolchain, allows AI apps to run efficiently across CPUs, GPUs, and NPUs without rewriting code. Developers build once, deploy everywhere. Industry insiders are calling it the most promising challenger to NVIDIA’s CUDA ecosystem. Qualcomm says once this acquisition closes, Modular’s stack will be the backbone of their “edge-to-cloud” strategy, breaking down barriers between device-side and cloud development.
The market loved what they heard. Right after the shareholder meeting, Qualcomm’s stock surged over 12% in after-hours trading. Even though it dipped 3.29% during regular hours to close at $197.41, the market cap still sits around a cool $208 billion. Wall Street analysts are buzzing, noting that this strategic upgrade is reshaping Qualcomm’s long-term value. JPMorgan recently noted that Qualcomm is evolving from a cyclical phone chip vendor into a stable, growing platform company.
