Micron vs SK Hynix

HBM Market Leadership and Share

SK Hynix is the undisputed leader in HBM production, holding approximately 60–62% of the global market as of mid-2025. The company's dominance stems from being the earliest mover in stacked HBM technology and its deep, long-standing relationship with NVIDIA as the primary memory supplier for successive generations of AI GPUs. UBS projects SK Hynix will capture roughly 70% of HBM4 volume destined for NVIDIA's Vera Rubin platform in 2026.

Micron has made significant gains, overtaking Samsung to become the second-largest HBM supplier at around 21% share. While still a distant second, Micron's trajectory is notable — the company has locked in its entire 2026 HBM supply through forward pricing agreements, signaling strong customer confidence. The gap is narrowing, but SK Hynix's structural advantage in volume and NVIDIA allocation remains substantial.

HBM4 Technology and Performance

SK Hynix completed HBM4 development first, announcing the milestone in September 2025 with a 12-layer, 36 GB product featuring a 2,048-bit I/O interface — double the previous generation. The company demonstrated speeds of 11.7 Gbps and claims a 40% boost in power efficiency, with projections that HBM4 will improve AI service performance by up to 69%.

Micron's HBM4 counters with bandwidth that exceeds 2.8 TB/s per stack — a 2.3x improvement over its HBM3E — and speeds above 11 Gbps. Micron emphasizes a 20% power efficiency gain over HBM3E. Both companies began volume shipments of HBM4 for NVIDIA's Vera Rubin in early 2026, though SK Hynix holds the larger allocation. The performance differences between the two are close enough that customer relationships and supply reliability may matter more than raw specs.

Capacity and Capital Investment

Both companies are investing aggressively to meet surging demand. Micron has announced approximately $20 billion in fiscal 2026 capital expenditure, primarily to expand HBM production capacity to 15,000 wafers per month. The company forecasts an annualized HBM revenue run-rate of approximately $8 billion and has pre-sold its entire 2026 output.

SK Hynix operates the world's largest HBM production capacity and continues to expand, though it has been less specific about exact capex figures. The company is investing in both its Icheon and Cheongju fabs in South Korea. The key differentiator is that SK Hynix can allocate existing high-volume DRAM lines to HBM conversion more quickly, giving it a structural capacity advantage that Micron is working to close.

NVIDIA Partnership and Supply Chain Position

SK Hynix occupies a privileged position as NVIDIA's primary HBM supplier. This relationship, built over multiple GPU generations from the A100 through the H100, H200, and now Vera Rubin, gives SK Hynix first-mover advantage on qualification and volume allocation for each new platform. For the Vera Rubin generation, SK Hynix is expected to supply the majority of HBM4.

Micron has diversified its customer base beyond NVIDIA, securing significant commitments from hyperscale data center operators and other AI chip designers. This broader customer mix provides some insulation against over-reliance on any single platform. Both companies are also positioning for next-generation platforms beyond Vera Rubin, where the competitive dynamics may shift as HBM4E enters the picture.

Next-Generation Roadmap: HBM4E and Beyond

SK Hynix showcased a 16-layer, 48 GB HBM4 module at CES 2026, signaling its intent to lead in higher-density stacking. The company's roadmap extends to HBM4E and potentially HBM5, maintaining its strategy of being first to market with each new generation. SK Hynix's close coordination with NVIDIA on future GPU architectures gives it valuable lead time in development.

Micron is developing HBM4E with a targeted 60% capacity increase over HBM3E and further power efficiency gains. The company's US-based manufacturing also positions it favorably for government-backed AI infrastructure projects that prioritize domestic supply chains. As AI infrastructure spending continues to accelerate, both companies are racing to ensure their roadmaps keep pace with the exponential growth in memory bandwidth requirements.

Geopolitical and Supply Chain Considerations

Micron's headquarters in Boise, Idaho, and its expanding US manufacturing footprint give it a strategic advantage amid growing emphasis on domestic semiconductor production. The CHIPS Act and related policies favor US-based memory production for defense and critical AI infrastructure applications.

SK Hynix, headquartered in South Korea, benefits from proximity to the broader Asian semiconductor ecosystem, including packaging and testing facilities. However, the company faces potential risks from US-China trade tensions, particularly regarding its operations in China. Both companies must navigate an increasingly complex geopolitical landscape where memory chips are recognized as strategically important technology.