DeepSeek's AI Model Disrupts Tech Stocks: Opportunities for Qualcomm

A Chinese artificial intelligence (AI) start-up named DeepSeek has caused a significant stir in the technology sector by introducing a cost-effective AI model. This innovative model was trained using limited resources and can perform tasks on par with more expensive models from prominent companies like OpenAI.

While there are valid concerns about the accuracy of DeepSeek's claims regarding training costs and the quality of chips used, the implications of their breakthrough cannot be overlooked. The company's achievement in creating an AI model that works efficiently without relying heavily on hardware and computing power has changed the conversation in the AI community.

Major firms such as Nvidia, Microsoft, and Meta Platforms have invested heavily in high-performance graphics processing units (GPUs) costing tens of thousands of dollars to support their AI training and data processing needs. However, DeepSeek's revelation that it can create a competent AI model at a fraction of the cost has sent ripples of concern through the investment community, affecting stock prices across the sector.

The sharp decline in stock prices of chip manufacturers like Nvidia came right after DeepSeek's AI assistant surpassed ChatGPT in downloads in the U.S. Apple App Store. However, analysts see potential benefit for certain companies in light of this development.

Recently, Kevin Cassidy, an analyst at Rosenblatt Securities, suggested that companies producing chips for edge AI applications may see a positive impact as a result of DeepSeek's breakthrough. Qualcomm has been identified as one of the prime beneficiaries.

Qualcomm's Role in AI Computing for Edge Devices

Qualcomm has established itself as a leader in creating application processors for smartphones. However, the company also designs chips used in personal computers (PCs), automotive technology, and Internet of Things (IoT) applications. A significant advantage of Qualcomm's products is that many of these devices operate at the edge of networks, allowing them to process data locally rather than relying on cloud servers.

These edge devices are increasingly utilized for AI applications. For example, smartphone manufacturers are introducing models equipped with on-device AI capabilities. Samsung's latest Galaxy S25 line features numerous on-device AI functions, including image generation, text interpretation, content summarization, and personalized user experiences—all powered by Qualcomm's Snapdragon 8 Elite processor.

Qualcomm is also advancing in the AI-enabled PC market with its Snapdragon X Elite processors. These chips allow computers to run large language models (LLMs) with over 13 billion parameters locally.

Moreover, Qualcomm's smartphone processors can handle LLMs with up to 10 billion parameters on device. DeepSeek’s strategy to offer distilled LLMs, which involves simplifying a larger model into smaller, more efficient versions to handle specific tasks, stands to greatly increase the accessibility of generative AI for edge devices.

Traditional LLMs typically require significant hardware resources. However, many edge devices, including smartphones, PCs, cars, drones, and security cameras, have limited capacity for hardware. LLM distillation becomes crucial in this scenario. DeepSeek's R1 LLM, while originally positioned as a large model with 671 billion parameters, is available in smaller versions ranging from 1.5 billion to 70 billion parameters.

These smaller models demand less computational power, making them suitable for edge devices. Lower computational requirements will ultimately result in reduced costs for consumers wishing to leverage AI on their devices. DeepSeek claims its R1 model can be between 20 to 50 times cheaper than OpenAI's o1 model for similar tasks.

At Qualcomm's 2024 Investor Day presentation, the company highlighted how the high costs associated with AI inferencing in the cloud are limiting the scalability of generative AI. DeepSeek's pricing strategy, charging just $0.55 for every million tokens compared to OpenAI's $15, could help alleviate some of these costs.

As more edge devices become capable of running advanced AI models at viable prices, demand for such devices is expected to surge. According to analysts at Fortune Business Insights, the edge AI market is set to grow at an impressive annual rate of 33% through 2032.

Future Growth Potential for Qualcomm

Over the past five years, Qualcomm's stock has increased by 97%, a performance overshadowed by the 165% gain seen in the PHLX Semiconductor Sector index. Qualcomm's stock performance has faced challenges due to sluggish smartphone sales in recent years. However, the advent of generative AI smartphones is on the horizon, with shipments expected to rise at an annual rate of 78% through 2028.

Furthermore, Qualcomm projects that the increasing demand for edge processors in IoT and automotive applications could result in a 2.5 times revenue growth in these sectors, reaching $22 billion by fiscal 2029. This would translate to a robust annual growth rate of 22%. Such developments indicate that Qualcomm's revenue growth potential over the next five years could exceed its historical growth rate of 15% per year.

Currently trading at an appealing valuation of 19 times its trailing earnings, Qualcomm presents an attractive investment opportunity, especially in light of the lucrative possibilities arising from the edge AI sector.

AI, Qualcomm, Investment