In a dual-pronged strategy to cement its leadership in the artificial intelligence (AI) sector, SK hynix is moving closer to its biggest customers—both geographically and technically.
A report from TrendForce indicates that the memory giant has secured a new strategic hub in the Pacific Northwest, just as it enters a critical “paid sampling” phase for its next-generation HBM4 memory with NVIDIA.
Table of Contents:
The New Seattle Hub
SK hynix has reportedly signed a lease for approximately 5,500 square feet of office space at the City Center Bellevue in Washington. This move places the company in the heart of a “Big Tech” corridor, specifically positioned to facilitate immediate collaboration with NVIDIA and Amazon.
Industry analysts suggest the Seattle expansion is a calculated move to support the growing demand for custom AI Application-Specific Integrated Circuits (ASICs). As hyperscalers like Amazon, Google, and Microsoft ramp up their internal chip designs to challenge NVIDIA’s dominance, the need for localized engineering support for High Bandwidth Memory (HBM) has become a competitive necessity.
The HBM4 Commercial Milestone
This physical expansion coincides with a major technical breakthrough. SK hynix and rival Samsung have reportedly transitioned from providing free engineering prototypes to delivering paid HBM4 samples to NVIDIA.
The shift to paid samples is a significant industry signal:
- Commercial Readiness: It indicates that the technology has met NVIDIA’s baseline specifications.
- Contract Finalization: Negotiations are now entering a pre-contract stage, with final volumes and pricing expected to be locked in during the first quarter of 2026.
- Quality Qualification: SK hynix is currently supplying between 20,000 and 30,000 final samples for rigorous testing within NVIDIA’s upcoming Rubin AI accelerator architecture.
A Growing US Ecosystem
The Seattle office serves as a strategic companion to SK hynix’s $3.87 billion advanced packaging facility in West Lafayette, Indiana. While the Indiana plant—slated for mass production in late 2028—will handle the physical manufacturing and packaging of AI chips on U.S. soil, the Seattle hub will focus on the “front-end” of the relationship: design collaboration and client responsiveness.
Competitive Pressure from Samsung
The timing of these moves is critical as the competitive landscape shifts. While SK hynix has long held the “HBM King” title, Samsung is showing signs of a strong recovery. Samsung is reportedly leveraging its in-house 4nm foundry process to gain positive feedback on its own HBM4 samples, recently surpassing performance targets in tests for Broadcom.
The Bottom Line
By establishing a physical presence in the Seattle tech ecosystem while simultaneously hitting commercial milestones with HBM4, SK hynix is building a “defensive moat” around its partnership with NVIDIA. As the industry moves toward the 2026 contract finalization, SK hynix is betting that being “in the room” with the world’s leading AI architects will ensure they remain the primary supplier for the next generation of AI infrastructure.
