Samsung Electronics and Advanced Micro Devices (AMD) have taken a major step toward strengthening their strategic relationship. On March 18, 2026, the two semiconductor giants signed a Memorandum of Understanding (MoU) focused on expanding collaboration in artificial intelligence (AI) memory technology and exploring a possible foundry partnership to manufacture future AMD chips.
This move could have far‑reaching implications for the global AI hardware landscape, influencing supply chains, competition, and performance benchmarks of next‑generation computing systems.
What the Deal Covers
AI Memory Collaboration
Samsung will deepen its role as a key supplier of advanced High Bandwidth Memory (HBM) technology — a critical component that feeds data to AI processors at extremely high speeds. Under the MoU:
-
Samsung will supply its next‑generation HBM4 memory for AMD’s upcoming Instinct MI455X AI accelerators.
-
The agreement also includes optimized DDR5 memory for AMD’s future EPYC server CPUs.
-
Samsung has already been supplying HBM3E chips for AMD’s current AI accelerators, and this deal signals a continuity and expansion of that relationship.
Why HBM Matters for AI
High Bandwidth Memory sits right next to AI processors and moves data faster than traditional memory. AI workloads — especially large language models and data center inference tasks — require memory speeds that can keep up with massive parallel compute units. HBM4’s high throughput and bandwidth are essential to next‑gen performance.
Possible Foundry Partnership on the Horizon
A standout feature of the MoU is that Samsung and AMD will explore a foundry deal. This would bring AMD closer to using Samsung’s fabrication capacity to produce chips — a shift from AMD’s historical dependence on outside contract manufacturers like TSMC.
What This Could Mean
-
Diversified Manufacturing: AMD could secure additional capacity outside Taiwan’s TSMC, reducing geographic concentration risks.
-
Direct Chip Production: Samsung’s advanced nodes, especially in the 3nm to 2nm range, are competitive and could support AMD’s future custom logic products.
-
Stronger Supply Chain Resilience: Contingencies such as geopolitical pressures and supply bottlenecks in AI silicon make multi‑source manufacturing increasingly attractive.
Navigating the Foundry Landscape
Although Samsung’s foundry division has historically lagged behind TSMC, recent advances — including partnerships with NVIDIA and other AI players — have sharpened its competitive edge. Samsung’s advanced process technologies, combined with memory and packaging capabilities under one roof, position it as a potential alternative high‑end chip manufacturer.
Strategic Context: AI Chip Market Dynamics
The semiconductor industry is in the midst of an intense race to serve demand from generative AI, cloud providers, and data centers. Key trends tied to this Samsung‑AMD pact include:
Competitive Memory Market
| Supplier | Approx. HBM Market Share | Notes |
|---|---|---|
| SK Hynix | ~57% | Current leader in HBM supply. |
| Samsung | ~22% | Aims to close gap with HBM4 production. |
| Others | Remaining share | Smaller suppliers competing on niche. |
Samsung’s efforts to scale HBM4 production are designed to reclaim competitiveness in high‑performance memory, which has become a scarce and strategic resource in the AI supply chain.
AI Chip Supply Chain Shifts
Large cloud players and AI integrators — from Meta to OpenAI and Oracle — are diversifying their AI hardware stacks. AMD’s multibillion‑dollar AI chip agreements with major customers have placed pressure on its memory pipelines and fabrication timelines. Expanding collaboration with Samsung helps alleviate some of that strain.
Executive Engagement and Signals
The deal follows high‑level engagement between Samsung Chairman Lee Jae‑yong and AMD CEO Lisa Su, reflecting the strategic weight of this partnership. Su also made a site visit to Samsung’s Pyeongtaek chip manufacturing facility to review production lines and discuss broader cooperation.
These interactions underscore how seriously both companies view the intersection of memory supply and potential foundry collaboration.
What Comes Next
Key Areas to Watch
-
HBM4 Deployment: Volume shipments to AMD’s AI accelerators later in 2026 could shift competitive dynamics against rivals relying on other memory suppliers.
-
Foundry Talks Progress: Official foundry contracts — if they materialize — would mark a watershed moment in AMD’s manufacturing strategy.
-
Industrial Competition: Samsung’s moves may intensify rivalry with SK Hynix in memory and with TSMC in logic semiconductor production.
Broader Industry Impact
The MoU signals an era in which memory, compute, and manufacturing are increasingly interdependent:
-
AI workloads will demand memory architectures tightly integrated with processors.
-
Supply chain diversification is no longer a luxury but a necessity.
-
Strategic alliances like Samsung‑AMD could redefine how cutting‑edge silicon is developed and produced.
Final Take
Samsung and AMD’s MoU represents more than a supplier agreement — it’s a strategic recalibration in response to surging AI demand. By locking in next‑generation memory supplies and opening the door to foundry cooperation, both companies are positioning themselves for a competitive, AI‑driven future where hardware performance and supply resilience matter more than ever.
Read next
Mystery AI model | Nvidia H200 approval | Google AI opt-out | BMG sues Anthropic | Amazon AI revenue | Alibaba AI strategy | UK AI content labels | Tencent AI investment | Softcat profit forecast | Samsung-AMD memory deal
