Summary of Key Points
The CEOs of Arm and NVIDIA discussed the surge in CPU demand driven by the explosion of AI agents at the Taipei Computer Show. AI agents require CPUs for tasks such as distribution and management, leading to demand that far exceeds expectations. The Arm architecture has become highly sought after due to its energy efficiency advantages, expanding its customer base. The use of PCs has diversified into two categories: portable devices and high-performance models for AI applications. Operating systems have not been marginalized by AI agents but have become even more crucial. The scale of the computer industry could grow by 10 times due to AI, and the stock prices of related companies have risen accordingly.
1. AI Agents are Driving CPU Demand Faster than Expected
The proliferation of AI agents, such as those that can automate tasks, has led to a dramatic increase in CPU demand. The Arm CEO noted that while predictions previously suggested a four-fold increase in the number of CPU cores per watt of power consumption, current estimates are for eight- or even tenfold increases. The total CPU market was once thought to exceed $120 billion in five years; now some predict it will double. This is because AI agents need CPUs to distribute tasks and manage data, transporting the smallest units of processing (e.g., “tokens”) to their intended destinations. Every CPU manufacturer is witnessing this surge in demand, though the exact multiplier remains uncertain, but it is clearly on the rise.
2. Arm Architecture CPUs are in High Demand Due to Energy Efficiency and Customer Appeal
Arm’s CPU architecture (as opposed to Intel and AMD’s x86) is highly efficient in power consumption. For example, Arm’s AGI CPUs deliver twice the performance per watt compared to x86 chips. In large data centers, using Arm architecture can save billions of dollars in energy costs. More and more companies are adopting Arm: Google’s ninth-generation TPU uses Arm technology, Amazon’s Graviton is gaining momentum, and NVIDIA’s Vera CPUs and RTX Spark chips also utilize Arm architecture. Companies like Meta, OpenAI, Oracle, and ByteDance are all Arm customers. The first generation of AGI CPUs has been mass-produced, with the second generation in development and the third generation on the way.
3. Diversified Use Cases for PCs
The demand for PCs has split into two categories: portable devices with good battery life and connectivity for everyday use, and high-performance models designed for AI applications. NVIDIA’s RTX Spark chips are aimed at the latter category, enabling large models to be run directly on the PC, turning it into an “always-running AI agent.” This allows for convenient local processing without relying on cloud services.
4. Operating Systems Are More Important Than Ever
Some worry that operating systems will become obsolete with the rise of AI agents. However, Jensen Huang (NVIDIA’s CEO) asserts this is a misconception. AI agents rely on various tools (such as office and design software), which all run on operating systems, making them even more essential. For instance, Windows allows AI agents to utilize 100% of the software’s functionality—although users may only use a small portion of it—and therefore will continue to be in demand for a long time.
5. AI Could Double the Size of the Industry, Driving Stock Price Increases
Jensen Huang predicts that the computer industry’s user base could grow from one billion to ten billion due to the adoption of AI agents, robots, and autonomous vehicles. After NVIDIA’s release of the RTX Spark, Arm’s stock price rose by 15%, and NVIDIA’s stock price increased by 6%. He also joked that Arm’s stock price always rises when NVIDIA releases new products, noting that he had previously sold some of his shares to fund cash needs. Although NVIDIA has planned its supply chain, demand may exceed expectations, while Arm does not have to worry about supply chain issues since it only licenses its technology.
In summary, the news highlights how the explosion of AI agents has led to a surge in CPU demand, giving an advantage to Arm’s energy-efficient architecture. This has transformed the PC and operating system landscape, benefiting the entire industry and related companies. For the general public, this means that CPUs are becoming increasingly important in the AI era, and companies like Arm and NVIDIA are seizing these opportunities.