Qualcomm in advanced talks to acquire Modular for about $4 billion

Reuters reported that the acquisition is an all-stock deal valued at nearly $4 billion, and calculated the transaction value at $3.92 billion using Qualcomm’s last closing price.[1] Qualcomm expects to issue up to 19.2 million common shares to Modular’s equity holders.[1] The companies expect the transaction to close in the second half of 2026, subject to customary closing conditions and regulatory approvals.[2] Bloomberg’s earlier report described advanced discussions but said talks could still fall through or deal details could change.[4] For planning purposes, the transaction should therefore be treated as pending rather than as a completed change in ownership.

Reuters reported that Modular’s software lets AI models run across different chips without developers writing code for each processor, and that the software primarily serves inference—the running of AI models.[1] Reuters also reported support for chips from Nvidia, AMD and other vendors.[1] Modular says its platform spans CPU, GPU, NPU and custom ASIC architectures without requiring rewrites for each accelerator.[2] The company says developers can build once and deploy across environments, with lower total cost of ownership.[2] Those are Modular’s stated platform capabilities, not evidence here of customer results or measured savings.

Reuters reported that Qualcomm is pursuing the acquisition as it seeks a larger data-center presence amid rising generative-AI demand.[1] Its smartphone chips generate the bulk of its revenue, and Reuters said the company is trying to reduce reliance on that business.[1] Qualcomm already targets data centers with processors and other AI chips, with shipments planned by year-end, according to Reuters.[1] Qualcomm says Modular would strengthen its software foundation for generative and agentic AI in data-center and edge environments.[2] Reuters also reported that Qualcomm was in talks to acquire AI-chip startup Tenstorrent for between $8 billion and $10 billion.[1]

Qualcomm says the acquisition would help it offer a silicon-agnostic compute layer across devices, edge systems and data centers.[2] It says the combined platform is intended to improve performance per watt, increase hardware flexibility and help customers deploy AI more efficiently.[2] Modular describes its developer community as open, industry-friendly and vendor-neutral.[2] Reuters noted that Nvidia’s CUDA has helped underpin its AI dominance by tying millions of developers to its chips.[1] Emarketer analyst Jacob Bourne said Qualcomm is betting that software to make inference more efficient could help it establish a data-center position.[1] The evidence does not establish whether the proposed combination will deliver those intended benefits.