A huge collaboration is echoing in the technology world. Competition between semiconductor giants is gaining a new dimension as AMD breaks course with Samsung Foundry for its next-generation 2nm manufacturing process. According to information from Daishin Securities and various reliable leak sources, AMD’s processors codenamed Enice and Verano are produced in Samsung’s advanced facilities.

This move is seen as not only a supplier change, but also a strategic step that shakes the balance in the global chip market and the artificial intelligence race. The fact that TSMC’s production capacity will be largely full by 2028 stands out as one of the biggest factors in AMD’s handshake with Samsung.

TSMC’s Full Capacity Benefited Samsung

TSMC, the undisputed leader of the semiconductor world, has been pushing the limits of its production capacity due to the excessive demand for artificial intelligence chips in recent years. At TSMC, where giants such as Apple and Nvidia are primary customers, it is almost impossible to find a gap in advanced node capacity by 2028.

This situation pushes performance-oriented giants such as AMD to look for alternatives. Samsung is emerging as the strongest candidate to fill this gap.

AMD CEO Lisa Su’s personal visit to Samsung’s factory in Pyeongtaek recently gave the signals of this partnership months ago. What Samsung will use in the 2nm process GAA (Gate-All-Around)The transistor architecture fully matches AMD’s goals of high efficiency and low power consumption.

In addition, the fact that Samsung is also a memory (DRAM) manufacturer helps AMD provide easier access to HBM3E and next-generation memory technologies, which are vital for AMD’s high-performance processors.

Venice and Verano: Performance Monsters of the Future

The “Venice” processor, which AMD plans to produce at Samsung facilities, is designed to reinforce the company’s dominance in the data center market. This monster, planned to be released in 2026, is full 256 “Zen 6C” coresowner In this structure built on 8 CCD (Core Complex Dies), each die contains 32 cores. Such a high core count combined with the 2nm manufacturing process could mean a new world record in performance efficiency per Watt on the server side.

On the other hand, the “Verano” project prepared for 2027 has a much more special place. Verano is fully optimized for “Agentic AI” workloads, that is, artificial intelligence that can make decisions on its own and perform complex tasks. This processor, which is expected to come with Zen 7 architecture, serves as the main server unit (host CPU) for AMD’s Instinct MI500 series accelerators. Samsung’s 2nm process provides the thermal budget and overclocking potential this complex architecture requires.

2nm GAA Technology and Yield Problem

GAA architecture, Samsung’s biggest trump card against TSMC, allows transistors to be controlled from all sides, reducing leakage current and increasing performance. However, this technology also brings production efficiency risks. Industry analysts also note that AMD could use Samsung as a “second source” to distribute risks, rather than replacing TSMC entirely.

Nevertheless, this cooperation not only increases Samsung’s prestige in the foundry business, but also makes AMD’s supply chain flexible. If Samsung can achieve the expected efficiency in the 2nm process, AMD’s hand in the competition with Intel and Nvidia in the processor world will become stronger than ever. The year 2026 seems poised to be a real breaking point in processor architectures and production technologies.