Packaging Is the BottleneckNarrow moat

SK hynix (SKHY) — moat facet

The constraint on HBM is not the wafers but the assembly, which is why SK hynix is committing about 19 trillion won to a single packaging plant.

Most discussions of memory capacity are about wafers, and for HBM that is the wrong unit. The constraint is packaging: the advanced bonding and stacking capacity that turns good dies into a good twelve-high stack.

Announced Korean plant investment, ₩ trillion~₩15TM15X fab~₩19TP&T7 packaging₩21.6TYongin fab 1P&T7 was raised by a further ₩7,093.1bn on 22 July 2026. Packaging, not wafers, is the limit.
The company is spending more on one packaging plant than on the fab feeding it.

This is where the real barrier sits. A competitor can add DRAM wafer capacity with money and time. Adding the capability to stack those wafers at commercial yield is a different problem, because the failure modes are mechanical and thermal rather than lithographic, and the learning is specific to the equipment set and the process. SK hynix is building an advanced packaging plant alongside its other capacity precisely because this is the gate1.

The economics are unforgiving in a way that favours the incumbent. Scrapping a stack destroys twelve good dies at once, so a producer with a few points less yield does not earn a few points less margin — it can be unprofitable at a price where the leader is comfortable.

The limitation is that packaging capacity is a physical asset with a build time, which means SK hynix's own growth is gated by it too. A shortage protects the leader's price and caps the leader's volume simultaneously.

Watch HBM shipment growth against DRAM bit growth. If HBM volume lags, packaging is the reason, and it is limiting the company as much as it is limiting the competition.

Moat trajectory: Widening

The board approved a further ₩7,093.1 billion for the P&T7 advanced packaging plant in July 2026, on a project already sized near ₩19 trillion. The constraint on HBM is assembly, and SK hynix is spending directly on it.

The number that tests this moat
Reported
Property, plant and equipment
₩88.9tn at June 2026, from ₩77.5tn in December

Packaging plants and fabs show up here before they show up in revenue. The balance flattening while demand holds would mean capacity, including packaging, is not being added.

Source: SK hynix semi-annual business report, January-June 2026 (Form 6-K, August 2026) ↗
⚠ Threats to the moat
⚠ THREAT 01
The Bottleneck Is Being Financed Away
A shortage that everyone is spending to eliminate is a shortage with a scheduled end date.
References
  1. ReportedAn advanced packaging plant, P&T7, is under construction in Cheongju alongside the company's other capacity.
    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 ↗
Sources
Generated September 23, 2026