Google Quietly Acquires Qualcomm: Mobile Chip Giant Targets Data Centers, Aiming for Over $15 Billion Annual Revenue

Google Quietly Buys Qualcomm: The Mobile Chip Giant Sets Its Sights on Data Centers, Aiming for Over $15 Billion in Annual Revenue

On November 6 local time, Alphabet, Google’s parent company, disclosed that as of the end of Q3 2026, the company holds approximately 2.977 million shares of chip giant Qualcomm, with a market value of about $548 million. Qualcomm was not listed in the documents from the end of June.

After the news was released, Qualcomm’s pre‑market stock price briefly rose more than 3%. Since Qualcomm is trying to shed its reliance on the mobile‑chip business, this new stake was immediately linked by the market to Qualcomm’s data‑center business:Qualcomm is attempting to move from mobile chips to AI infrastructure, and Google is also expanding compute capacity to meet growing AI demand.

In the latest Q4 earnings report, Qualcomm CEO Cristiano Amon reiterated the company’s plan to achieve $5 billion in data‑center revenue in fiscal year 2027; according to previously announced targets, this figure will exceed $15 billion by fiscal year 2029.

However, as more documents are uncovered, Alphabet’s holding of Qualcomm stock also suggests another explanation.

According to The Information, the stake may be related to Qualcomm’s acquisition of U.S. AI software startup Modular. Alphabet’s venture arm GV previously invested in Modular, and when Qualcomm acquired the company this year, it primarily used its own stock as consideration. In Qualcomm’s subsequent SEC filing, two GV‑affiliated funds were listed as selling shareholders of Qualcomm stock, collectively holding about 3.196 million shares.

However, this number does not match the 2.977 million shares disclosed by Alphabet, and the share counts in the two documents have not been fully reconciled.

Qualcomm is seeking its next growth driver

For a long period, Qualcomm’s business model has been tightly linked to the smartphone market. As one of the world’s major mobile chip suppliers, its Snapdragon processors are widely used in Android phones, and its baseband chips and communications patent licensing businesses have also provided the company with long‑term revenue.

However, as the smartphone market matures, this business structure is beginning to face new pressures. On one hand, global smartphone shipments are slowing, and manufacturers are placing greater emphasis on cost control; on the other hand, Apple’s continued push for its own baseband chips also puts Qualcomm at risk of losing long‑term customers. Qualcomm’s latest disclosed Q4 FY2026 results already show this pressure:Due to the weakening smartphone business and other factors, its mobile chip revenue for the quarter fell about 26% year‑over‑year, dropping to $5.14 billion.

(Source: Qualcomm)

In recent years, Qualcomm has gradually increased its investments in automotive, IoT, personal computing, and industrial sectors, aiming to transfer the low‑power computing, communications, and chip‑design expertise it has built in the mobile chip arena to a broader range of scenarios.

But with the rise of generative AI, Qualcomm sees an even larger market:data centers.

In the past, AI computing in data centers was primarily dominated by NVIDIA GPUs. However, as AI application scales grow, computing demands are also shifting. Especially as AI moves from simple Q&A to agents, models need to continuously invoke tools, execute tasks, and handle more requests, increasing inference workloads. For cloud service providers, beyond chip performance, they must also consider computing costs, energy consumption, and overall data center operational efficiency.

This gives Qualcomm an entry point. Its long‑term accumulation of low‑power computing technology, as well as the collaborative design capability of different compute units such as CPU, GPU, and NPU, can theoretically provide advantages in certain data‑center workloads. Especially in the AI inference market, as model deployment scales continue to grow, customers are increasingly focused on balancing performance, power consumption, and cost.

In June 2026, Qualcomm officially unveiled its data‑center business roadmap at the Investor Day, launching the Dragonwing product portfolio for data centers, which includes the C1000 CPU, AI300 inference accelerator, and High‑Bandwidth Compute (HBC) technology, among others. Unlike the traditional GPU approach, Qualcomm aims to deliver a compute solution covering various data‑center segments through a combination of CPU, AI accelerator, memory, and interconnect technologies. The company also projected that by fiscal year 2029, data‑center revenue will exceed $15 billion, while non‑mobile revenue will reach $40 billion.

This means that Qualcomm is not satisfied with being merely a chip supplier, but hopes to extend the hardware‑software co‑design capabilities it built in the mobile sector to cloud computing infrastructure.

The acquisition of Modular precisely fills the software gap. This startup focuses on AI software infrastructure, with core products including the Mojo programming language and the MAX AI inference platform, aiming to solve software compatibility across different chip architectures and make it easier for developers to deploy and optimize AI models on diverse hardware.

This is especially important for Qualcomm. In the past, its advantage was mainly in mobile devices, while the developer ecosystem and software toolchain were the parts that needed to be bolstered to enter data centers. After acquiring Modular, Qualcomm can integrate its software platform with its own chips, lowering the barrier for developers to use Qualcomm hardware.

The Chip Battle in Data Centers

Qualcomm’s entry into data centers is not an isolated business move. Over the past two years, AI compute demand has continued to grow, and competition in data centers has expanded from simply buying GPUs to encompassing chip design, memory, network interconnects, and software ecosystems. Large tech companies are increasingly getting involved in chip design, hoping to reduce costs, improve energy efficiency, and lessen reliance on a single supplier through customized solutions.

NVIDIA still holds the most important position. Its advantage lies not only in the GPUs themselves but also in the CUDA software ecosystem, network interconnect technology, and a complete solution from chip to full system. However, the high procurement cost and supply constraints also drive large cloud providers to look for alternative solutions.

AMD is vying for the GPU market with its Instinct series accelerators; Broadcom focuses on designing custom AI chips (ASICs) for large tech companies. Compared with general‑purpose GPUs, ASICs can be optimized around specific workloads, creating differentiation in energy efficiency, cost, and performance. Chip companies such as Marvell are also seeking opportunities in network interconnect and custom compute.

At the same time, cloud computing giants have also started designing their own chips. Google has the TPU, Amazon offers Trainium and Inferentia, Microsoft launched Maia, and Meta continues to push MTIA. For these cloud providers, developing in‑house chips does not conflict with external procurement; the key is to find the most suitable hardware solutions for different compute tasks.

Qualcomm’s entry is happening amid this round of industry division of labor changes.

After unveiling its data‑center roadmap in June, Qualcomm began accelerating efforts to partner with large tech companies. In September, the company announced a multi‑generation custom‑chip collaboration with Amazon Web Services (AWS), focusing on AI inference computing and data‑center optical interconnects. The partnership shows that Qualcomm has started engaging customers with massive compute demand. Qualcomm is also advancing its data‑center chip business with Microsoft, Meta, and other firms.

(Source: Qualcomm)

It is noteworthy that Qualcomm’s partnership with Amazon has moved beyond a traditional chip procurement relationship. According to Reuters, the agreement allows Amazon to purchase up to roughly $60 billion worth of Qualcomm AI data‑center chips and related products in the future, and Qualcomm has granted Amazon warrants to buy its own stock, covering up to 25 million shares. This means that Amazon’s equity arrangement is linked to the actual procurement business.

However, winning a customer order is only the first step into the data center market; delivering on time is an equally real test. Qualcomm CEO Cristiano Amon said in an interview that current AI infrastructure construction faces supply chain tightness, and capacity for key segments such as advanced‑process chips remains limited. For Qualcomm, which has just entered this market, it must not only win customers but also ensure chip manufacturing and product delivery can keep up with demand.

At the same time, Qualcomm also has to contend with Nvidia, AMD, and Broadcom and Marvell, which are already deeply involved in the supply chains of major cloud providers. It must prove not only chip performance but also software compatibility, system integration capability, and long‑term supply reliability.

Google’s Compute Power Landscape