en.Wedoany.com Reported - Qualcomm reported revenue of $44.3 billion and operating profit of $12.4 billion in its last fiscal year. The company, whose core business revolves around patent licensing and chip sales, is attempting to extend its influence from mobile communications into the artificial intelligence and data center markets. CEO Cristiano Amon spent nearly two hours in the company's latest investor update before mentioning 6G, and barely touched on network-related content, with his focus more on the faster-growing areas of AI and data centers.
Qualcomm faces multiple pressures in the smartphone sector. According to Omdia data, overall device shipments in the recent first quarter grew only 1% year-over-year, with AI demand driving up component prices and prompting consumers to delay upgrades. Meanwhile, Qualcomm's reliance on Chinese customers has deepened, with 46% of sales in the last fiscal year coming from customers headquartered in China, compared to 37% two years ago. Tensions with major U.S. customer Apple are also rising, as the latter has shifted to in-house chips and may replace Qualcomm products with new modems in future devices.
Qualcomm's strategy in the radio access network (RAN) sector is being adjusted. According to Light Reading, Qualcomm has begun informing the market that it plans to stop selling its FSM100 and FSM200 small cell base station chip platforms to new customers, offering only a single software update to existing customers going forward. At the same time, Qualcomm's participation in the RAN market is expected to increase. Its distributed unit (DU) chips and software products under the X100 brand have already partnered with Viettel, which plans to deploy 5,000 sites by the end of 2026. Caleb Banke, employee manager in Qualcomm's product marketing department, stated that in real networks, Qualcomm's technology delivers up to 24% better energy efficiency under high traffic loads compared to traditional network vendors.
Looking ahead to the 6G era, Qualcomm plans to offer a proprietary solution. Sunil Patil, Vice President of RAN Product Management at Qualcomm, said the goal is to have a central processing unit (CPU) that efficiently handles RAN workloads, while using a neural processing unit (NPU) to handle AI workloads. Qualcomm's NPU technology may play a dual role similar to Nvidia's graphics processing unit (GPU) in RAN—both handling L1 RAN workloads and using spare capacity for AI inference. Qualcomm has introduced Hexagon technology into its upcoming AI200 and AI250 chips, application-specific integrated circuits (ASICs) designed specifically for AI inference.
In the data center sector, Qualcomm is attempting to challenge Nvidia. By acquiring Modular for $4 billion, Qualcomm aims to create an abstraction layer that allows developers to write code that can run on multiple hardware platforms, including CUDA. Cristiano Amon stated that 6G infrastructure is no longer dedicated communications equipment and requires computing power. Any competitor with higher energy efficiency may find the market receptive.
