Arm, a British chip designer, is currently preparing for an initial public offering (IPO) that has caught the attention of investors in the semiconductor and artificial intelligence (AI) industries. The IPO aims to raise nearly $5 billion, valuing Arm at over $50 billion. While Nvidia has experienced a significant rally this year, driven by high demand in these sectors, analysts believe that Arm’s prospects may be different due to its unique position in the market.
One of the main drivers behind the increased interest in semiconductor companies, like Arm and Nvidia, is the growing prominence of AI technology. OpenAI’s ChatGPT has brought generative AI into the spotlight, showcasing the ability of AI models to generate answers based on user prompts. The training of these AI models requires substantial computing power, which is where companies like Nvidia come in. Nvidia specializes in designing graphics processing units (GPUs) that are used in data centers for training and running AI models.
In contrast, Arm focuses on designing blueprints or “architectures” for semiconductors. These architectures serve as the foundation for building chips, including central processing units (CPUs) that are used in smartphones and other devices. Arm’s CPUs are present in 99% of the world’s smartphones, including those produced by major players like Apple. While Arm CPUs are also used in data centers, they are often coupled with GPUs for training data. Arm primarily generates revenue through royalties and licensing its architecture, with more than 50% of its revenue coming from smartphones and consumer electronics.
According to analysts, Arm’s near-term growth is not primarily driven by AI but by the mobile market and increasing royalties. The company is trying to shift investors’ focus toward AI’s potential in the edge and data centers in the long term. However, this is not currently a significant part of Arm’s exposure. The company’s revenue boost is not expected to come from the demand for chips used in training big data AI models.
Despite its limited involvement in AI training, Arm is likely to play a significant role in AI on the “edge.” The term refers to AI processes being carried out on devices, such as smartphones, rather than in the cloud. To enable AI applications on the edge, devices will require low-power yet high-performance chips. Arm is currently designing the architecture for these chips, as it aims to optimize the model to run locally on end-devices.
Arm’s processors are already capable of running AI workloads, such as voice recognition and image filtering, efficiently on smartphones. However, the benefits of AI in terms of revenue are not expected to accrue to Arm for at least three to five years. While Arm has been positioned as an AI company, similar to Nvidia, by its owner SoftBank, the actual revenue from AI applications on end-devices is not currently flowing to Arm.
Arm’s upcoming IPO has garnered significant attention from investors due to the surge in interest in both semiconductors and AI. However, Arm’s potential in AI differs from that of Nvidia. While Nvidia focuses on GPUs for AI training in data centers, Arm’s future lies in designing low-power, high-performance chips for AI applications on the edge. Although Arm has already optimized its processors for running AI workloads on smartphones, the revenue boost from AI is not expected to be immediate. Arm’s positioning as an AI company in the long term aligns with the increasing demand for AI on end-devices, but it will take time for this transition to materialize in terms of revenue. As the semiconductor and AI industries continue to evolve, Arm’s unique role and potential in AI on the edge deserve careful consideration and analysis.