Qualcomm acquires Modular for $3.9B to challenge Nvidia's AI dominance
Qualcomm has announced its intent to acquire AI software startup Modular to modernize its AI development pipeline and debuted two new data center chips, the Dragonfly C1000 and AI300. Meta has signed a multi-year agreement to implement the C1000 CPU in its infrastructure, contributing to a significant upward revision in Qualcomm's projected non-handset revenue.
Key Takeaways
- Qualcomm will issue 19.2 million shares to acquire Modular, valuing the startup at more than double its last private funding round.
- Meta signed a multi-year, multi-generation agreement to deploy the 250-core Dragonfly C1000 CPU in its server fleet starting in H2 2028.
- The Dragonfly AI300 machine learning accelerator was unveiled as a power-efficient alternative for data center racks, compatible with liquid cooling.
- Non-handset revenue projections for fiscal 2029 were raised by $18 billion, with the data center segment alone targeted to exceed $15 billion.
- Adjusted earnings per share guidance for 2029 was increased to $18, surpassing analyst expectations of $15.26.
Why It Matters
Qualcomm is making an aggressive play to erode Nvidia's CUDA-driven dominance by acquiring Modular’s hardware-agnostic software layer. This enables high-performance video and AI workloads to run across diverse silicon without costly code rewrites, addressing a major bottleneck for streaming platforms scaling generative AI and personalized recommendation engines. By securing Meta as a cornerstone customer for its custom server CPUs, Qualcomm proves it can compete for hyperscale infrastructure budgets beyond its core mobile business. Industry observers should track whether the 2028 production timeline for the C1000 allows competitors like AMD or internal hyperscaler silicon projects to close the efficiency gap first.
Additional Context
The acquisition of Modular follows a period of rapid growth for the startup, which was founded by former Apple and Google engineers behind the Mojo programming language. Per Reuters and Sacra in late 2025, Modular had achieved a $1.6 billion valuation after a $250 million Series C round led by the U.S. Innovative Technology Fund, with participation from GV and Greylock. Its central value proposition—a unified compute layer designed to abstract away hardware-specific optimizations—is a direct challenge to the proprietary software ecosystems that have historically forced developers to commit to specific GPU architectures. Qualcomm's broader strategic pivot towards the data center has been building for several years. Per ServeTheHome and Investing.com reporting from June 2024 to 2026, the company's "Dragonfly" portfolio was first teased as an answer to the surging power demands of agentic AI. This diversification is critical as global smartphone shipments face significant annual contractions in 2026, per Bloomberg. By expanding its automotive design-win pipeline to $65 billion and targeting $10 billion in automotive revenue by 2029, Qualcomm aims to shift its business composition so that handsets represent only one-third of its semiconductor revenue by the end of the decade. Concurrent with the Modular deal, rival infrastructure moves are intensifying. Per reports in June 2026, OpenAI and Broadcom recently unveiled the "Jalapeño" inference chip, while Microsoft continues to expand deployments of its own Maia AI accelerators. Qualcomm’s success in this environment depends heavily on its ability to deliver the promised performance-per-watt benefits of the C1000 and AI300, particularly as hyperscalers increasingly move toward disaggregated, multi-vendor architectures to manage the high costs of scaling large-scale AI production environments.
Read full article at siliconangle.com
Get this in your inbox → Subscribe
Enjoy our coverage?
Add StreamingMeme as a preferred source on Google to see more of our streaming news at the top of your Search results.
Add as preferred source