Qualcomm Acquires Modular to Strengthen AI Software

JAKARTA — Qualcomm acquires Modular on Wednesday to strengthen its artificial intelligence software capabilities, especially in the inference and data center markets that are being built out at scale.

The move brings together the U.S. chipmaker and Modular, an AI infrastructure startup known for cross-chip software. Qualcomm did not disclose the deal value, but Reuters estimated it at about US$3.92 billion, while Bloomberg first reported talks were near US$4 billion.

The transaction is expected to close in the second half of 2026. Qualcomm said the move is aimed at helping companies run AI more efficiently as token usage costs rise and budget pressure intensifies across corporate IT teams.

Qualcomm acquires Modular to chase the AI inference market

For Qualcomm, the goal is not just to add an asset. The company is trying to build a stronger position in the inference market, the stage where AI models are used for real tasks, from answering questions to analyzing large-scale data.

At this point, efficiency is everything. Many companies can build AI models, but not all can operate them cheaply and reliably. That is because token costs — the small units language models use to process prompts — keep rising and squeezing technology budgets.

Modular fits that gap. The startup builds a software platform that lets developers run AI across different chips without rewriting code for each architecture. That approach reduces dependence on a single hardware ecosystem.

“We believe the future is horizontal platforms that are developer-friendly, can run across different computing environments, and give customers real choice in how and where they run AI,” Qualcomm CEO Cristiano Amon said in a statement.

Competition with Nvidia also looms

The acquisition also carries strategic weight. Qualcomm is not talking about software alone. The company wants a deeper role in data centers, a market that has become a spending magnet for tech giants because of surging demand for generative AI.

At the same time, Modular brings Qualcomm closer to Nvidia’s software stack, which has long dominated the AI ecosystem. Wired said Modular’s platform and programming language are designed so developers can write once and run code on many chips. That may sound like a small challenge, but it could grow if adoption spreads.

Qualcomm has long been better known for smartphone chips. Its biggest revenue still comes from the mobile segment. Still, the company has been pushing beyond that, including chips for data centers and other AI devices that are expected to begin shipping this year, according to CNA.

Bringing in Modular gives Qualcomm a smoother path forward: not just selling silicon, but also the software layer that makes those chips easier to use. The AI market is moving in that direction. Strong hardware alone is no longer enough.

Deal value and what it means for the AI business

Qualcomm did not give an official transaction amount. Even so, two separate reports provide a fairly close picture. Bloomberg said negotiations were around US$4 billion, while Reuters put the figure at US$3.92 billion based on Qualcomm’s latest closing share price.

The structure matters too. Wired reported that Qualcomm will issue up to 19.2 million common shares to Modular equity holders. In other words, the acquisition is being paid for with stock, not all cash. For Qualcomm, that helps preserve liquidity while it spends heavily on AI expansion.

For Modular, the deal marks the end of a fast-rising startup phase. Wired noted that the company raised US$250 million at a US$1.6 billion valuation about nine months ago. The jump in value shows just how hot investor demand has become for AI infrastructure software.

For companies using AI, the impact could show up in operating costs. If Qualcomm-Modal software helps run inference more cheaply, companies can reduce computing bills, speed up model deployment, and rely less on expensive infrastructure.

Why the Modular acquisition matters for business and tech readers

The deal points to a new direction in AI competition: it is not just about who has the fastest chip, but who can make AI cheapest to use. That matters for business operators, startups, and technology teams trying to calculate the cost of AI adoption.

If inference costs fall, AI adoption could spread across more sectors. A retail company could use models for customer service. A bank could process documents faster. A factory could analyze sensor data without driving up computing bills.

Investors will also read the transaction as a signal that Qualcomm does not want to stay limited to the phone market. Data centers, software stacks, and AI optimization now sit high on the priority list. The market is large, and the risks are real. Competition with Nvidia and other players will not be easy.

Qualcomm is set to discuss its business direction further at its investor day on Wednesday. At that moment, the market will likely wait for one simple answer: how far is the company willing to go in turning itself from a mobile chip maker into a more complete AI infrastructure player?

“We want to give customers real choice in how and where they run AI,” Amon said. It sounds like a strategy. But in an AI market that keeps getting more expensive, it also sounds like a business bet.

Quick summary

Qualcomm acquires Modular to strengthen AI software and expand its push into data centers. The transaction is estimated at about US$3.92 billion and is targeted to close in the second half of 2026. The main focus is more efficient, more flexible, and cheaper AI inference for companies.

Short FAQ: What is Modular? An AI infrastructure startup that builds cross-chip software. Why does it matter? It can help lower the cost of running AI models. Who is most affected? Companies building AI services at scale.