[Edaily Reporter JAEMIN SONG ] As Taiwan’s TSMC, the world’s leading foundry company, moves to raise prices for its cutting-edge processes, SamsungElectronics’ foundry business is likely to benefit. Analysts suggest that with TSMC’s production capacity nearing its limits due to surging demand for artificial intelligence (AI) semiconductors, any further increase in cost could accelerate efforts by global Big Tech companies to diversify their supply chains.

According to Taiwanese IT media outlet Culpium and industry sources on the 26th, TSMC has reportedly notified its customers of plans to raise prices for advanced processes of 7 nanometers (nm) and below by 5–10%. It is reported that the price hike has been expanded to cover all processes of 7 nm and below, not just the 3-nm process as initially reported. Advanced processes account for approximately 75% of TSMC’s wafer sales.

The industry views this price hike as a measure reflecting the burden of large-scale capital investment driven by surging demand for AI semiconductors. TSMC has been reviewing price increase plans since early this year and is accelerating its expansion of global production bases in Arizona, Japan, and Germany, as well as its investment in 2-nanometer mass production.

If the price hike increases the burden on customers, efforts to diversify supply chains are expected to intensify. This is because major customers such as Apple, NVIDIA, and AMD are increasingly likely to consider alternative foundries in response to rising production costs.

In fact, global Big Tech companies are already expanding their “dual-vendor” strategies. Tesla has diversified its supply chain by having its next-generation AI chip, the A16, produced by SamsungElectronics, while its successor, the A16.5, will be manufactured by TSMC. Google is also reported to have entrusted part of its next-generation TPU production to Intel Foundry, given TSMC’s limited production capacity.

In response, SamsungElectronics is accelerating efforts to expand production capacity and improve yield rates. The company plans to begin initial operations at its 2-nanometer plant in Taylor, Texas, by the end of this year and start mass production for major clients starting next year. At a recent global strategy meeting of the DS Division, strategies for improving yields in advanced processes and securing new customers were discussed. SamsungElectronics is responsible for producing Tesla’s AI6 chip and Groq inference chips for NVIDIA platforms, and Jeon Young-hyun, Head of the DS Division (Vice Chairman), recently revealed that he met with Jensen Huang, CEO of NVIDIA, to continue discussions on collaboration for next-generation AI chips.

Industry observers note that as this expansion of orders aligns with increased production capacity, expectations are growing for an improvement in the performance of the foundry business, which has been suffering from prolonged losses. Han Jin-man, President of SamsungElectronics’ DS Division Foundry Business, reportedly stated at a recent briefing on the company’s business status for employees, “There is a high likelihood of achieving profitability by 2028.”