Surging demand for artificial intelligence chips is also benefiting Samsung Electronics' (005930) foundry business. With AI chip production heavily concentrated at TSMC, a growing number of customers are seeking to spread their orders, helping Samsung Electronics hold onto second place in the global foundry market.
Samsung Electronics took about 4% of the global foundry market in the second quarter, ranking second behind TSMC, according to research firm Counterpoint Research on the 30th. Demand for the company's SF4 and SF5 nodes — its 4-nanometer and 5-nanometer processes — along with rising wafer prices supported the results.
The boost from expanding AI investment is spreading quickly to foundry players beyond TSMC. Second-quarter revenue at pure-play foundries excluding TSMC rose 18% from a year earlier, roughly double the 9% growth logged in the first quarter. The gain reflected stronger demand not only for AI accelerators but also for server central processing units, networking chips and power management integrated circuits.
The "Foundry 2.0" market, which combines foundry services with back-end processes such as packaging, is also growing. The market was worth $96.6 billion (about 131 trillion won) in the second quarter, up 25% from a year earlier. Pure-play foundry revenue within that total climbed 29% to $54.9 billion (about 74 trillion won).
The gap with TSMC nevertheless remains wide. TSMC posted second-quarter revenue of $40.2 billion (about 54 trillion won), up 34% from a year earlier, accounting for roughly 42% of the entire Foundry 2.0 market. Counterpoint said Samsung Electronics and Intel are absorbing demand from customers looking for additional suppliers, but that they remain difficult substitutes for TSMC in terms of advanced-node capacity, yields and advanced packaging capabilities.
The center of gravity in the competition is expected to widen from wafer production to advanced packaging. Counterpoint projected that the Foundry 2.0 market will sustain double-digit growth in the second half, with packaging, substrate and test capacity emerging as the key variables in expanding AI chip supply.
