Is It Time to Invest in Samsung? Samsung Electronics Foundry Business Enters Recovery Period
AI Podcast
Samsung's foundry business is poised for recovery, driven by increasing AI chip demand and significant client partnerships with Nvidia, AMD, Tesla, and Qualcomm. Nvidia's use of Samsung's 4nm process for next-gen AI chips highlights their advanced capabilities. Agreements with AMD for HBM4 and foundry collaboration, along with Tesla chip production and Qualcomm's 2nm technology discussions, signal a shift towards longer-term customer relationships and stable capacity utilization. The market anticipates profitability in Q4, with break-even achievable next year, as yield improvements and sustained client trust become key performance indicators.
TradingKey - Against the backdrop of continuously rising demand for AI chips, Samsung Electronics' foundry business is entering a long-awaited recovery window. As news of partnerships with clients such as Nvidia ( NVDA ), AMD, Tesla ( TSLA) and Qualcomm ( QCOM) gradually materializes, the market has begun to reassess Samsung's competitiveness in advanced process nodes. Consequently, whether the foundry business can turn a profit in the fourth quarter of this year has become a focal point of external attention.
Recently, the market has seen a steady stream of reports regarding Samsung securing orders or collaborations with Nvidia, AMD, Tesla, and Qualcomm. It is widely believed that this will provide more stable capacity utilization for Samsung's wafer foundry business, offering a turnaround window for the logic chip business that has been suppressed for a long time. Based on these developments, South Korean media has also judged that Samsung's foundry business is expected to turn a profit in the fourth quarter of this year, or at the very least, the inflection point for earnings improvement is now clearer than before.
What is truly prompting the market to re-examine Samsung is its AI-related business. At this month's GTC conference, Nvidia publicly showcased a next-generation AI chip manufactured using Samsung's 4nm process. This move by Nvidia directly brought Samsung's advanced process capabilities to the forefront, leading the market to reassess its competitiveness in high-end foundry services.
Simultaneously, Samsung and AMD have signed a memorandum of understanding (MOU) to continue advancing cooperation around HBM4 supply, EPYC platform memory solutions, and potential future foundry collaborations. For Samsung, this is no longer just a matter of securing another order; it signifies that its position in the AI supply chain is gradually shifting from a peripheral collaborator to a more central role.
More notably, Samsung is not just focusing on individual projects this time, but is seeking ways to foster longer-term customer relationships. Senior executives recently mentioned that the chip business will increasingly adopt multi-year contracts of three to five years in the future, with the clear objective of minimizing uncertainty caused by market volatility. On the Tesla side, Samsung has confirmed it will produce related chips at its Texas plant in the U.S., with mass production expected to begin in the second half of 2027; on the Qualcomm side, discussions are also underway with Samsung for foundry production using 2nm technology. For a foundry that has long faced skepticism regarding customer stickiness and the depth of its advanced process order book, this change in contract structure carries more weight than any single large order.
From an operational perspective, Samsung's foundry business is not far from breaking even. The market expects that this business will have the opportunity to achieve break-even next year, as confidence in its advanced process yields, capacity utilization, and ability to onboard major clients is gradually returning. If orders from Nvidia, AMD, Tesla, and Qualcomm continue to materialize, the pace of narrowing losses in the foundry business could be faster than originally anticipated.
For investors, the narrative for Samsung Foundry is no longer the old question of "can it catch up with TSMC," but rather "when can it truly achieve stable profitability." Demand for AI computing power continues to drive intensive procurement of high-end chips and HBM. If Samsung can simultaneously capture recovery dividends from both its memory and foundry lines, the elasticity of its earnings improvement will be more pronounced than looking at either business in isolation. Moving forward, observers will be watching closely not just the volume of orders, but also whether yields can continue to rise, whether the delivery pace remains stable, and whether customers are willing to entrust more long-term projects to Samsung.
