AI Demand Boosts CPU Importance; Intel CEO: 'Many CEOs Calling for More CPUs'

With the rise of AI, the importance of Central Processing Units (CPUs) has been elevated once again. Intel CEO Pat Gelsinger stated that CPU demand is increasing, while supply is constrained. He noted that in the past four weeks, many CEOs have called him asking for more CPUs, which represents an opportunity for Intel. Gelsinger spoke to global media after his COMPUTEX keynote today. Regarding CPU trends, he pointed out that AI requires new architectures, and CPUs are crucial for reinforcement learning and orchestration, leading to growing demand. Gelsinger said that supply is indeed constrained, and Intel is very focused on supply to ensure customer needs are met and customers are satisfied. In the past four weeks, many CEOs have called him asking for more CPUs, and he has asked his executives to help, hoping to keep up with customer schedules. He stated that Taiwan is very strong in personal computers and server rack systems, which is why Intel has established long-term cooperative relationships with many Taiwanese companies and is investing more resources to expand these relationships. Regarding product production strategy, Intel pointed out that it adopts a hybrid model of in-house manufacturing and outsourcing, which gives Intel maximum flexibility and is a business advantage. In addition, unique packaging technology allows Intel to choose different foundries without being tied to a specific one. From a data center perspective, all products are within the scope of Intel's wafer fabs, which is key to achieving differentiation and building partnerships. However, Intel also cooperates with multiple wafer foundries. Intel pointed out that it will adopt a hybrid model, sometimes needing its own fabs and sometimes needing external cooperation. The engineering team can flexibly choose partners as needed, especially since Intel possesses unique packaging technology, which is being used to advance business development and address some supply constraints.