DeepSeek's Breakthrough: How Qualcomm May Benefit from Affordable AI Models

A Chinese artificial intelligence (AI) startup named DeepSeek has recently caused a stir in the technology sector by introducing an AI model that is highly cost-effective. This new model, trained with limited resources, claims to perform comparably to more powerful and expensive models from established companies like OpenAI. The announcement has caused shares of major tech firms to decline, raising questions about the investment worthiness of powerful hardware previously considered essential for AI development.

The reactions from investors stem from skepticism regarding DeepSeek's assertions about its training costs and the types of chips utilized. Nonetheless, the development signifies a potential shift in the AI landscape, as it shows that efficiency can be achieved without extensive computing power.

Major firms like Nvidia, Microsoft, and Meta Platforms have invested heavily in high-performance graphics processing units (GPUs), which can cost tens of thousands of dollars, to build data centers adept at training AI models. With DeepSeek's revelation, there are concerns among investors that such massive financial commitments may no longer be necessary.

In the wake of DeepSeek's success, some companies stand to gain from this new approach. Analyst Kevin Cassidy from Rosenblatt Securities suggests that firms producing chips for AI applications that operate at the edge may see an uptick due to this breakthrough. One of Cassidy’s notable recommendations is Qualcomm.

Qualcomm's Role in Edge AI

Qualcomm is well-known for making application processors primarily for smartphones. Additionally, the company designs chips that are utilized in personal computers, automobiles, and Internet of Things (IoT) devices. A significant advantage of these devices is their ability to process data locally at the edge of a network, which minimizes the need for cloud data processing.

AI applications are increasingly being run on these edge devices. For instance, smartphone manufacturers are integrating on-device AI features. The latest Galaxy S25 smartphones from Samsung will showcase multiple on-device AI functionalities like image generation, text interpretation, and content summarization, all powered by Qualcomm's Snapdragon 8 Elite processor.

Qualcomm is also expanding into the AI-enabled PC market with its Snapdragon X Elite processors, which can support large language models (LLMs) with over 13 billion parameters operated locally. Meanwhile, Qualcomm’s smartphone processors can handle LLMs of up to 10 billion parameters directly on the device. This trend towards distilling larger models into smaller, task-specific models aligns well with the growing push for more efficient edge AI capabilities.

LLMs generally necessitate significant computing resources to operate. However, edge devices like smartphones, PCs, and drones have limitations in terms of hardware due to size and energy constraints. This is where DeepSeek's figure, which offers distilled versions of its R1 LLM that range from 1.5 billion to 70 billion parameters, can play a crucial role. These lighter models can effectively function on devices with less processing power.

The reduced computational costs mean consumers will benefit financially, with DeepSeek claiming its R1 model can be 20 to 50 times less expensive to use than OpenAI's more traditional models. At its recent Investor Day presentation, Qualcomm noted that the substantial costs associated with AI inferencing in the cloud hinder widespread scalability. With DeepSeek's more affordable approach, consumers could experience significant savings.

The demand for edge devices capable of running AI models at lower costs is expected to rise significantly. According to forecasts by Fortune Business Insights, the market for edge AI could grow at an annual rate of 33% up until 2032.

Future Growth for Qualcomm

Over the past five years, Qualcomm's stock has experienced a 97% increase, which is less than the 165% growth seen in the PHLX Semiconductor Sector index during the same timeframe. Qualcomm's progress has been affected by stagnant smartphone sales in recent years. Yet, the surge in demand for generative AI smartphones could mark a turning point, with shipments projected to grow by 78% annually through 2028.

Moreover, Qualcomm anticipates that the rising need for edge processors in IoT and automotive sectors could result in a revenue increase of 2.5 times, potentially reaching $22 billion by fiscal 2029. This suggests a 22% annual growth rate ahead, outperforming Qualcomm’s previous sales growth of 15% per year over the last five years.

Given that Qualcomm's stock trades at a reasonable 19 times its trailing earnings, it presents as an attractive AI investment opportunity, especially given the forthcoming edge AI prospects.

AI, Qualcomm, DeepSeek