You're offline - Playing from downloaded podcasts
Back to All Episodes
Podcast Episode

Qualcomm Buys AI Startup Modular for $3.9 Billion to Challenge Nvidia's CUDA

June 25, 2026

0:00
5:10
Podcast Thumbnail

Qualcomm has agreed to acquire AI software startup Modular in an all-stock deal worth roughly $3.9 billion. Modular's Mojo language and MAX inference engine let AI models run across chips from Nvidia, AMD, Intel and Qualcomm, directly targeting Nvidia's CUDA software lock-in. The deal is expected to close in the second half of 2026.

Qualcomm Makes Its Biggest AI Software Bet

Qualcomm has confirmed it will acquire Modular, an AI infrastructure software startup, in an all-stock transaction valued at approximately $3.9 billion. Announced on June 24, 2026, the deal sees Qualcomm issue up to 19.2 million shares to Modular's shareholders, with completion expected in the second half of 2026 once regulatory and shareholder approvals are secured. CEO Cristiano Amon described the acquisition as a pivotal moment for the AI industry and a clear signal that Qualcomm intends to compete well beyond its traditional mobile chip business.

Software, Not Silicon

Unlike most AI acquisitions, this one is not about chips. Modular, founded in 2022 by Chris Lattner and Tim Davis, builds software that lets AI models run efficiently across many different kinds of hardware. Its flagship technologies are the Mojo programming language and the MAX inference engine, a hardware-agnostic stack that allows developers to write AI code once and deploy it across CPUs, GPUs, NPUs and custom ASICs from Nvidia, AMD, Intel and Qualcomm, without rewriting for each platform.

Chris Lattner is a heavyweight in the field. He created Apple's Swift programming language and the LLVM compiler infrastructure that underpins much of modern computing, and he previously led Tesla's Autopilot software team. His co-founder Tim Davis also came from Google, where the pair worked on machine learning infrastructure.

Aiming at Nvidia's Moat

The strategic logic centres on CUDA, Nvidia's software platform that has effectively locked a generation of AI developers into Nvidia hardware. CUDA is widely seen as Nvidia's strongest competitive advantage, far more durable than any single chip. By owning a hardware-agnostic software layer, Qualcomm hopes to make running AI workloads on its own chips easier and more attractive, while loosening Nvidia's grip on the wider market.

Qualcomm is expected to bring in Modular's entire team of around 150 employees, including both founders. The acquisition complements Qualcomm's broader push into the data centre, where it has been building out server-class AI accelerators and needs a credible software story to make them usable for developers.

Why It Matters

The move reflects a growing industry consensus that the battle for AI dominance will be won or lost on software, not just raw silicon. If Modular's open, cross-hardware approach gains traction, it could lower costs for companies running AI, increase competition among chipmakers, and reduce the industry's heavy dependence on a single vendor. For Qualcomm, it is a bold step toward becoming a serious force in the trillion-dollar race to power artificial intelligence.

Published June 25, 2026 at 6:39am

More Recent Episodes