[Science World] AI Semiconductor Development Capability as National Competitiveness

[Asia Economy] Artificial Intelligence (AI) operates inside computers. Therefore, the better the semiconductor performance of a computer, the higher the performance of AI. Recently, ‘ChatGPT’ has become a hot topic, and behind it all lies high-performance semiconductor technology. Lately, the competition to develop ‘AI semiconductors’ specially designed to maximize AI performance has intensified. This is because it is recognized as a key strategic task to dominate future society. The global market research firm ‘Gartner’ also forecasted that the AI semiconductor market size will reach $117.9 billion (approximately 150 trillion KRW) by 2030.

What types of AI semiconductors are there? First, many are made by applying computer Graphics Processing Units (GPUs). GPUs process video signals, and this calculation method has many similarities with data AI computations. The American GPU specialist company Nvidia leads this field, accounting for 90% of the AI semiconductor market.

Dedicated AI semiconductors are also being researched and developed. The ‘Tensor Processing Unit (TPU)’ developed by Google is famous as the first-generation AI semiconductor. It adds designs to maximize efficiency on top of the advantages of GPUs. It is about 15 to 30 times faster than GPUs.

‘Neuromorphic semiconductors,’ which mimic the structure of the human brain, are also under research. They are called second-generation AI semiconductors. Intel’s ‘Loihi’ is a representative example. Intel created AI models suitable for this semiconductor and experimented, showing data analysis capabilities more than 3,000 times superior to existing ones. Related research is also underway in Korea. News of related research comes from the Korea Institute of Science and Technology and Samsung Electronics’ Advanced Institute of Technology.

The global semiconductor market is dominated by big tech companies from the United States and others. It seems difficult for AI semiconductors produced by us to stand out as products. Even so, we must not let go of research. AI semiconductor development capabilities are directly linked to national competitiveness. Without scientific and technological prowess in the era of technological hegemony, we cannot secure our place.

Jeon Seung-min, Science and Technology Specialist Writer

This content was produced with the assistance of AI translation services.