Samsung Foundry may have secured a significant order for 2-nanometer chips from a North American fabless customer. The most widely circulated name is AMD, although there is no official confirmation from either party at this time. The information comes from a note by Daishin Securities cited by the leakster Jukan and reported by specialized media, mentioning a supposed 2 nm CPU order for laptops. If confirmed, this would be an important signal for Samsung, which has been trying for years to regain credibility in advanced manufacturing against TSMC’s dominance.
It is important to read this news with caution. AMD already announced in April 2025 that its next-generation EPYC processors, codenamed Venice, had reached the silicon milestone on TSMC’s N2 node. AMD described Venice as the first HPC product made using TSMC’s advanced 2 nm process, making TSMC the confirmed partner for that server family. Therefore, the alleged order to Samsung does not necessarily mean AMD will abandon TSMC or that Venice will switch foundries. It could be a second source, a specific variant, client chips, mobile APUs, or a strategy to diversify capacity.
The interesting aspect is not just the potential customer name, but the context. The industry is entering a phase where access to advanced nodes has become a strategic bottleneck. Server CPUs, AI accelerators, GPUs, premium laptop chips, and custom SoCs are competing for capacity in highly limited fabrication facilities. In this scenario, AMD has clear incentives not to depend solely on one supplier if it wants to sustain growth in data centers, AI, and PCs.
Samsung needs a major customer to validate its 2 nm process
For Samsung, an agreement with AMD would be more than just a contract. The South Korean company has invested for years in its foundry business, but its position relative to TSMC has weakened due to performance issues, process maturity, and a lack of major external customers in advanced nodes. Its GAA technology was ahead of TSMC’s schedule, but that was not enough to shift the market balance.
Samsung reported in its Q4 2025 results that its foundry business had begun mass production of first-generation 2 nm products and planned to increase production of second-generation 2 nm products in 2026. It also noted expectations of order growth from HPC and mobile customers, as well as strengthening solutions that combine logic, memory, and advanced packaging.
This combination is significant for AMD. The company not only needs advanced wafers for CPUs but also competes in AI accelerators, servers, complete platforms, and increasingly in systems where memory, packaging, and connectivity are as important as the manufacturing node. Samsung can offer an integrated proposal with foundry, memory, and packaging technologies—aligning with current pressures to secure supply of HBM, interposers, and advanced capacity.
Samsung’s roadmap positions its SF2 process as an evolution based on GAA with improvements in performance, efficiency, and area compared to SF3. In 2023, the company explained it would start mobile applications with 2 nm in 2025, extend to HPC in 2026, and automotive in 2027. Simply put: first volume mobile, then more demanding workloads.
AMD cannot afford to depend solely on TSMC
AMD is now in a much different position than a decade ago. It is no longer just a competitor to Intel in consumer CPUs; it is a key player in the server market, competing in AI accelerators with the Instinct family and designing complete platforms for large training and inference workloads. This ambition requires sufficient and predictable manufacturing capacity.
TSMC remains AMD’s main partner for the most advanced nodes. Venice, based on Zen 6 cores, is officially associated with TSMC’s N2 node. Moreover, specialized media have positioned Venice as a next-generation EPYC family with up to 256 cores in some variants, aimed at high-density servers and AI workloads.
However, relying on TSMC has costs. Advanced capacity faces enormous demand from Apple, NVIDIA, AMD, Qualcomm, large AI clients, and custom designs from hyperscalers. Even if AMD keeps TSMC as its main supplier for critical products, exploring a second source with Samsung could reduce risk, improve negotiation leverage, and secure more volume.
The question is which specific product Samsung would produce. The original rumor mentions 2 nm CPUs for laptops from a North American fabless customer. Some analysts and leakers point to AMD, but this doesn’t directly align with the Venice and Verano narrative if interpreted as server products. It’s also possible it involves a client family or a less-publicized portion of the roadmap.
Verano adds another layer of uncertainty. AMD’s roadmap for AI has linked this name to future platforms in 2027 alongside the Instinct MI500 GPUs, as a next step after the Helios and Venice ecosystems. Yet, public details remain limited, and many specifications are based on leaks or tech presentations. It’s prudent to treat these as plans or rumors rather than confirmed facts.
A race no longer won solely by the best node
The potential involvement of Samsung in AMD orders signals a fundamental shift: the semiconductor race is no longer decided solely by who has the most advanced node on paper. Manufacturing yields, available capacity, cost per wafer, packaging, memory, lead times, geographic diversification, and trust in execution now play critical roles.
TSMC remains the dominant player. Its advantages in customer base, process maturity, and volume are hard to match. But the market needs alternatives. If Samsung can demonstrate that its 2 nm process is competitive for demanding products, it could regain ground with external clients and reduce global reliance on a single leading-edge foundry.
For AMD, even limited initial volume makes sense. Using Samsung as a second source would allow testing real performance, learning about the process, reserving future capacity, and gaining flexibility. If results are positive, the collaboration could expand. If yields aren’t sufficient, TSMC would remain the primary partner.
For Samsung, the margin for error is narrower. A customer like AMD would validate its technology in the market but also impose high standards. Manufacturing its own mobile chips or internal designs differs from producing high-performance CPUs for a competitor in servers, PCs, and AI. Process quality, energy efficiency, and supply stability will be critical.
The final takeaway is cautious but significant. No official confirmation exists from AMD or Samsung, and the information originated in analyses and leaks. Nonetheless, the rumor reflects a real trend: chip designers seek more capacity, alternative sources, and less dependency on a single foundry. In the AI boom era, strategic value lies not only in designing the best processor but also in manufacturing it on time and at volume.
Frequently Asked Questions
Has AMD confirmed it will manufacture 2 nm chips at Samsung?
No. So far, this is unconfirmed information based on a note from Daishin Securities and leaks reported by specialized media.
Does this mean AMD is abandoning TSMC for Venice?
No. AMD officially announced Venice was made using TSMC’s N2 node. If Samsung is involved, it might serve as a second source, a different product family, or a specific variant.
Why would this be important for Samsung Foundry?
Because an order from AMD would validate its 2 nm process for the market and strengthen its position against TSMC in advanced manufacturing.
What does AMD gain from a second foundry?
Increased capacity, reduced dependence on TSMC, better supply diversification, and leverage to negotiate costs and lead times in a high-demand market for advanced nodes.