Talk of AI dominated discussions at the Taiwan External Trade Development Council (TAITRA)-hosted COMPUTEX 2025, one of the world’s most extensive IT exhibits and a premier annual event for the global tech industry, and its attending exhibit for startup companies, InnoVEX 2025. Almost everything, in some form, was connected to the concept, from EdTech, medicine, online analyses, to EV charging, edge AI servers, and CPUs.
The recent development of ChatGPT and other AI software, and world governments’ race to build data centers, further hyped up AI. But what is AI, and does it truly hold the key to the continued advancement of technology? How does a set of programming programs on a machine aim to accomplish such goals?
Consider that AI, at its base, is a set of programming code running on a machine. To function, it requires a significant amount of data to learn from, or Large Language Models, and would, in turn, output an astronomical amount of data, thereby necessitating the need for storage. Consider, too, that the machine in which the program is housed would require robust processing power to allow the code to process the data, which would require a staggeringly high amount of power to maintain and run.
These requirements in and of themselves fuel the development of peripheral and next-gen technology as we continue our quest to develop AI. Seagate, known for its data storage products, has answered the challenge by developing the Heat Assisted Magnetic Recording (HAMR) technology to ensure the increased data density stored on each storage disk, resulting in the same number of discs on a drive, but increased total data.
Intel, the CPU giant, acknowledged that AI is a continuing trend that will see further development over recent years. When asked what it was doing to not just be a passive trend follower, but what it could do to define, or set, the trend, Intel said it is doing its best in developing hardware that would be ready for the public when AI use is more widely accepted.
As a producer of both CPUs and GPUs, Intel recognizes that the edge GPUs have over CPUs in AI processing won’t be going away soon, and stated instead that the company considers the main challenge ahead to be how to harness the processing power of both CPUs and GPUs to process AI algorithms.
Similarly, MSI is developing rack-level edge AI, or the direct deployment of AI models on local edge devices for real-time data processing and analysis without reliance on cloud servers, with a focus on industrial applications, which again highlights how the development of AI technology drives companies and other technologies to keep pace.
