The upcoming Computex is expected to prominently feature GenAI.
Taiwan's Ministry of Finance recently released preliminary statistics for March 2024's customs import and export trade, highlighting the market dominance of AI servers, with Taiwanese manufacturers securing 90% of the global market share. According to data from the Foreign Trade Association, the number of exhibitors has exceeded 1,500.
The United States has imposed high tariffs on servers and network communication products manufactured in China. Such measures have led to a mass Sino-exodus of server and network communication factories, with server manufacturers choosing to return to Taiwan.
Network communication factories, on the other hand, are expanding into Taiwan and Southeast Asia. The emergence of ChatGPT has stimulated demand for GPUs and AI servers, coinciding with Taiwan's peak performance, leading to an influx of orders. Taiwan's customs also notably feel the heat of this thriving trend.
The latest statistics show that Taiwan's export value of information and communication products and audiovisual products has grown by over 70% to both China and Europe, 1.4-fold to the United States, 14 times to Southeast Asia, and 40% to Japan.
The Foreign Trade Association pointed out that Arm CEO Rene Haas will attend the COMPUTEX exhibition in Taipei and deliver a speech on "Accelerating AI Innovation from Cloud to Edge." It aims to address how AI can unleash unprecedented capabilities for human society and how AI operates at global computing touchpoints from the cloud to the edge.
The Taipei Computer Association stated that Computex 2024, themed "Connecting AI, Co-Creating the Future," will feature up to 4,500 booths. Computex 2024 will showcase a wide array of solutions for GenAI computing, including AI servers, AI PCs, and Edge AI devices.
Nvidia CEO Jensen Huang previously remarked that GenAI is the defining technology of this era, and the company's proposed Blackwell architecture is the GPU engine driving the new industrial revolution. The AI chips of the Blackwell architecture primarily include B100, B200, and GB200, with the GB200 chip being four times faster in AI training speed and 30 times stronger in inference capability than the previous generation H100 chip while consuming 25 times less energy.
IDC forecasts that global enterprise investment in Generative AI across all aspects will reach US$143 billion by 2027, with a Compound Annual Growth Rate (CAGR) of 73.3% from 2023 to 2027, 13 times the CAGR of global IT spending during the same period.
