Korean Media: Samsung Follows TSMC in Raising 4nm and 5nm Wafer Foundry Prices by ~15%

Following market reports that TSMC (2330-TW)(TSM-US) had increased pricing for its advanced process technologies, Korean media reported that Samsung Electronics (005930-KS) has also raised its wafer foundry quotes for new customers. Prices for 4nm, 5nm, and select automotive 8nm processes have been increased by about 15%, indicating that the AI wave is reshaping the pricing mechanisms in the semiconductor foundry industry.

The AI investment boom is rewriting the competitive rules of the global semiconductor foundry sector. As demand for advanced processes continues to outstrip supply, foundry providers are gradually shifting from a historically 'customer-driven' pricing model to a 'supplier-driven' one.

According to industry sources cited by Chosun Biz, Samsung’s price adjustment is not a blanket increase but targeted at process nodes with strong demand from AI chips and automotive semiconductors, specifically for new customers.

Analysts believe this move is not merely a reflection of rising costs but a strategic rebalancing of pricing to better align with current market supply and demand dynamics. It also suggests that Samsung’s pricing power in certain process nodes has improved compared to the past.

Earlier, market rumors indicated that TSMC had notified major clients including NVIDIA (NVDA-US), Apple (AAPL-US), and AMD (AMD-US) of planned price increases for advanced processes such as 3nm, 5nm, and 7nm, with hikes ranging from 5% to 10%.

If the Korean media reports are accurate, this pricing adjustment extends beyond the most cutting-edge nodes to include more mature advanced processes as well.

What makes this round of price hikes more significant than previous ones is not the magnitude of the increase, but the fundamental shift in the pricing logic of the foundry industry. Historically, new process nodes commanded higher prices during early production due to low yields and high capital costs. Once yields improved and capacity expanded, prices for the same node typically remained stable or even declined over time, with price increases rare unless transitioning to the next-generation node.

However, rapid expansion in AI infrastructure investment has significantly altered market dynamics. Continued high-volume wafer starts by AI chipmakers like NVIDIA (NVDA-US), coupled with cloud service providers (CSPs) aggressively building AI data centers, have kept advanced process capacity in a prolonged state of shortage.

Meanwhile, development costs for next-generation nodes like 2nm continue to rise, and capital expenditures on extreme ultraviolet (EUV) equipment and other tools are increasing. As a result, foundry pricing is becoming increasingly influenced by market supply-demand conditions and investment costs, rather than solely by process maturity.

Analysts note that supply tightness driven by AI investments has granted foundries greater pricing authority. While premium pricing was previously based on technological leadership, current supply constraints are further strengthening overall negotiation power. In the future, advanced process pricing may become more flexible, with market participants likely to see more frequent price adjustments for the same node in response to shifting demand and supply conditions.

Industry observers warn that if foundry prices continue to rise, combined with increasing high-bandwidth memory (HBM) costs and persistent bottlenecks in advanced packaging capacity, the manufacturing cost of AI chips will continue to climb. This could lead to higher AI server deployment costs and eventually feed into higher prices for end-user devices such as smartphones and personal computers. As long as the AI investment cycle continues, the upward pricing trend centered on advanced processes is expected to persist.