Qualcomm is seeking to take advantage of the momentum surrounding generative AI to motivate car manufacturers to acquire more of its chips. To that purpose, it unveiled an example of how its chips could be put to use for furnishing cars with ChatGPT-like aides, without the need of a phone. As of now, only around 3% of Qualcomm's earnings comes from automotive.
Qualcomm is best known for its chips and modems inside Android phones, but it is now hoping to leverage generative AI to further its presence in the automotive industry. The company has sold its hardware package, Snapdragon Digital Chassis, to automakers like GM, Hyundai, and Volvo, and is expecting more sales over the next few years. It estimates that automotive revenue will reach $9 billion by 2031. The company makes between $200-$3,000 per car that uses its chips and software, and $5 per car with 5G capabilities. For example, GM's new $130,000 Cadillac Escalade IQ SUV, equipped with Qualcomm chips and software, enables features such as 55-inch dashboard display, lane-keeping and hands-free driving. However, Qualcomm is facing stiff competition from other chipmakers, including Intel and Nvidia, as well as traditional auto suppliers such as Continental, NXP Semiconductors, and Bosch.
At this week's demonstration, Qualcomm displayed potential upcoming uses of its chips in collaboration with comprehensive language models and generative AI. One scenario was of a car assistant that could search for a chicken enchilada recipe and add the ingredients to the shopping list. The second example showcased Stable Diffusion, an AI model, to deliver an AI-made birthday greeting card to the driver's brother. All the demos were performed on the car's processor, not on a cell phone. Its CEO, Cristiano Amon, views cars as a "new computing platform," he stated in a blog post. In terms of AI, Qualcomm is competing with Nvidia's cloud GPUs that function in applications like ChatGPT. Although the AI trend has aggravated Nvidia's stock prices by three times this year, Qualcomm has only increased by 5%. A near-term aim for the language models is to design a sharp user guide that has been programmed and optimized to the long automotive user handbooks, indicated Nakul Duggal, Qualcomm's automotive senior vice president. Additionally, driver monitoring is a sector for the company, with artificial intelligence used to assess whether the driver is dozing off or not attentive. Duggal added that they are working with automakers to upgrade the manual in the glove compartment to an understanding car. A further possibility for Qualcomm's platform is having features that could be utilized to freshen up the software of the car with new self-driving workings, which could be a new source of income for car manufacturers. He concluded, "Things like where are you driving, do you have hassle parking, would you like to subscribe to an automated parking facility that you have not yet bought, but can be upgraded via the air? Or would you like to get a free trial for a period of time? There are so many options once you have the context."
Comments