Manufacturing and YieldNarrow moat
SK hynix (SKHY) — moat facet
Yields are not a margin item in HBM; they decide whether a given price is profitable or ruinous, and they cannot be bought.
In a business where every competitor sells the same specification at the same published price, the entire difference between a good year and a catastrophic one is what it costs you to make the thing. That cost is mostly yield: the proportion of what comes off the line that is good enough to sell.
For ordinary DRAM this is hard. For HBM it is brutal. A twelve-high HBM stack requires bonding twelve dies together with thousands of vias running through them, and a single failed connection ruins a package that already contains twelve good chips. The scrap is not one die, it is twelve. This is why HBM capacity cannot be conjured by anyone with a chequebook, and why a rival can announce a competitive product long before it can ship one profitably.
SK hynix's own account of what it competes on is a short and unglamorous list: pricing, manufacturing costs, yields and product availability, and product performance1. Note what is not on it — brand, relationships, ecosystem. This is a company that wins by making more good chips per wafer than the other two.
That advantage is real and it is also perishable. It is a lead measured in quarters, not decades, because the physics are public and the equipment is sold to everyone by the same handful of suppliers. What keeps it is continuous reinvestment, which is why the fabs never stop being built.
The number to watch is gross margin against the other two through a down quarter. In an up-cycle everybody's yields look adequate because the price covers everything. Cost leadership is only visible when the price stops covering everything, and SK hynix has not been tested that way since 2023.
M15X took its first wafers in the first quarter of 2026, Yongin's first cleanroom is due in the first quarter of 2027, and P&T7 is being built alongside. The manufacturing base is being enlarged faster than at any point in the company's history.
Yield leadership has to be rebuilt in every new fab. Spending falling while three sites are under construction would mean the pipeline is slipping.
Source: SK hynix semi-annual business report, January-June 2026 (Form 6-K, August 2026) ↗- ReportedThe competitive factors SK hynix lists are pricing; manufacturing costs, yields and product availability; and product performance, quality and reliability.SK hynix Inc., Form 424B4 prospectus for its Nasdaq offering (SEC, CIK 2120882) — FY2025 revenue W 97,147bn, cost of sales W 38,456bn, gross profit W 58,691bn, S&A W 5,019bn, R&D W 6,466bn and profit for the year W 42,948bn, against FY2024 revenue W 66,193bn / profit W 19,797bn and FY2023 revenue W 32,766bn / cost of sales W 33,299bn / gross loss W 533bn / loss for the year W 9,138bn; Q1 2026 revenue W 52,576bn, gross profit W 41,679bn, profit W 40,346bn. IDC market shares for Q1 2026: second in DRAM at 29.1%, first in HBM at 56.4%, second in NAND at 18.5%, with the three DRAM producers together above 90% of revenue. DRAM products were 77.1% of total sales in 2025 and 77.3% in Q1 2026; NAND 21.3% and 22.0%. Competitors named: Samsung Electronics, Micron Technology and CXMT in DRAM; Samsung Electronics, Kioxia, Micron Technology and Sandisk in NAND; competitive factors listed as 'pricing; manufacturing costs, yields and product availability; product performance, quality and reliability'. First to develop HBM using TSV packaging, commercialised HBM3E in 2024 and developed HBM4 in 2025; HBM carried a per-gigabyte price premium of more than five times traditional DRAM in 2025. M15X cleanroom opened October 2025 with wafer input from Q1 2026; Yongin construction began February 2025 with the first fab's phase-one cleanroom expected Q1 2027; an advanced packaging plant (P&T7) is under construction in Cheongju and one is planned in Indiana for the second half of 2028. Intel's NAND business cost US$6.6bn in December 2021 plus US$2.2bn in March 2025, operated as Solidigm; NAND mass production is transitioning from 176-layer to 238- and 321-layer technologies. Gartner forecasts DRAM revenue of US$143bn in 2025 rising to US$401bn in 2027, HBM US$33bn to US$86bn and NAND US$68bn to US$341bn. A substantial portion of sales is attributable to a limited number of customers located in the United States and China; the two largest were 14.8% and 12.4% of revenue in Q1 2026 and the largest was 23.9% of revenue in 2025. The offering was 177,900,000 ADSs at US$149.00, each ADS one-tenth of a common share, representing 17,790,000 shares against 712,702,365 outstanding; Baillie Gifford, Coatue and Situational Awareness Partners indicated non-binding interest of up to US$7bn. — FY2023-FY2025 and Q1 2026 · publ. 2026-07-10 · source ↗
- SK hynix Form 424B4 prospectus for the Nasdaq offering (SEC EDGAR)
- SK hynix — new facility investment, P&T7 advanced packaging plant (SEC Form 6-K)
- SK hynix — semi-annual business report to 30 June 2026 (SEC Form 6-K)