Yangtze Memory Technologies
Integrated device manufacturer: designs, fabricates, and packages 3D NAND flash wafers and sells NAND components, embedded storage (UFS/eMMC), and consumer/enterprise SSDs (Zhitai brand); capital intensity funded by state investors
YMTC raises via state-directed capital injections at parent Changjiang Holdings rather than conventional VC rounds; only the Apr 2025 raise carries a reported post-money valuation. The 2023 point uses post-injection registered capital as a stated proxy, not a market-priced mark. Founding (2016) and interim registered-capital steps carry no reported valuations and are omitted rather than estimated.
Earnings, margins, COGS & capex
YMTC is pre-IPO and publishes no audited financials; the first full financial picture will arrive with its STAR Market prospectus. Tutoring filing was accepted by the Hubei CSRC branch on May 19, 2026 (tutoring institutions CITIC Securities and CSC Financial; auditor Deloitte), with the listing application expected as early as mid-June 2026 - no confirmed application acceptance as of early Jul 2026. What is externally observable: global NAND share climbed from 8% (Q1 2025) to 13% (Q1 2026), tying Micron and SanDisk, during a quarter when industry revenue surged 3.5x YoY to a record $46B - implying roughly $6B of quarterly revenue and ~445% YoY growth per Counterpoint Research. Funding history: founded 2016 with Big Fund and Hubei/Wuhan state backing; ~$7B injection from state-backed investors including Big Fund II in Mar 2023 (SCMP); and an Apr 2025 parent-level raise of ~RMB 9.42B (~$1.3B) at RMB 161.6B (~$22B) post-money that brought in 16 investors including units of Agricultural Bank of China, China Construction Bank, Bank of Communications, and Bank of China plus Quanhong Investment (Yangyuan) - classic pre-IPO cornerstone positioning.
Revenue trend
Margins
Likely rising sharply with supercycle NAND pricing (contract prices up ~55-60% QoQ in Q1 2026 per TrendForce-tracked estimates); unverifiable until prospectus
Historically pressured by sanctions-driven equipment workarounds and yield costs on domestic toolchain
Structurally negative-to-thin for a scaling memory IDM in expansion phase
COGS structure
Wafer fabrication dominates: 3D NAND fab equipment depreciation (the largest line for any memory IDM), silicon wafers, specialty gases/chemicals, packaging/test. Sanctions add a localization tax - domestic tools (Naura, AMEC) replacing blocked US equipment (Lam, Applied, KLA) carry yield and throughput penalties that inflate unit cost versus Samsung/SK hynix/Micron.
Capex
Aggressive expansion: ~$3B Phase III venture at Wuhan (est. Sep 5, 2025; YMTC 50.2% / state-owned Hubei Changsheng Phase III Investment 49.8%), ~130k WSPM currently (~8% of global 3D NAND) ramping toward 150k, with a stated aim of ~15% of global NAND supply by end-2026. Phase III runs >50% domestic tooling and was pulled forward a year to mass production as early as H2 2026; reports point to two further Wuhan fabs longer-term. The STAR IPO is explicitly a capacity-expansion financing event.
Latest earnings
n/a - no consensus exists
None issued; first formal disclosure will be the STAR Market prospectus
- Global NAND share (Q1 2026)
- 13% (tied 4th with Micron and SanDisk; Counterpoint)
- Last-round valuation
- ~$22B (Apr 2025, RMB 161.6B post-money at parent Changjiang Holdings)
- Wafer capacity
- ~130k WSPM (~8% of global 3D NAND) as of 2025, ramping toward 150k
- Flagship product
- Gen5 TLC NAND X4-9070, 294 total / 232 active layers, Xtacking 4.0 hybrid bonding, 3600 MT/s
Growth drivers
- AI/data-center NAND supercycle — enterprise segment was 43% of total NAND demand in Q1 2026, projected to surpass 60% by year-end (Counterpoint)
- China domestic-substitution mandate — only mainland full-stack 3D NAND IDM, with procurement pull from Chinese OEMs, cloud, and government
- Share gains from 8% to 13% as incumbent supply stays disciplined (major suppliers adding virtually no new capacity in 2026, per TrendForce) and Chinese buyers de-risk from US/Korean suppliers
- 294-layer (232 active) Gen5 TLC NAND (X4-9070, Xtacking 4.0) shipping since early 2025 with bit density above 20 Gb/mm2 - among the highest of any shipping TLC NAND
- STAR Market IPO proceeds earmarked for capacity expansion
Bull & bear
YMTC is the structural winner of China's storage-sovereignty push at the exact moment the AI storage supercycle makes NAND scarce: verified share gains to 13%, a genuinely leading-edge product (294L Xtacking Gen5), captive domestic demand, bottomless state capital, and a STAR Market IPO that will likely price well above the ~$22B last round.
- Share trajectory is real and third-party-verified: 8% to 13% in four quarters, now tied with Micron and SanDisk for fourth globally (Counterpoint Research, Q1 2026)
- Product parity achieved despite sanctions: Gen5 294-layer TLC (X4-9070) with bit density above 20 Gb/mm2, among the highest of any shipping TLC - Xtacking hybrid bonding is an architectural moat incumbents only later adopted
- Demand tailwind is secular, not just cyclical: enterprise segment heading past 60% of NAND demand by end-2026 (Counterpoint); Chinese hyperscalers and OEMs have policy reasons to buy YMTC first
- Capital is not a constraint: ~$7B state injection (2023), ~$1.3B pre-IPO round (2025) with major state banks' investment arms, ~$3B Phase III fab venture, and IPO proceeds ahead - YMTC can outspend the cycle in a way no private-market peer can
- Phase III's >50%-domestic toolchain hitting mass production as early as H2 2026 would neutralize the bear case's central chokepoint
- A-share semis trade at large scarcity premiums; even the sober China Daily-reported forecast of ~RMB 300B (~$44B) implies roughly 2x the Apr 2025 round for pre-IPO holders
YMTC is an opaque, state-directed NAND pure-play whose current estimates are inflated by a once-in-a-decade price supercycle, whose cost structure carries a permanent sanctions tax, and whose speculated IPO valuations (up to RMB 2-3T floated in local press) are unmoored from the ~$22B round priced only 14 months ago - with no audited financials to arbitrate.
- The entire revenue narrative rests on third-party estimates during an anomalous quarter (industry revenue up 3.5x YoY to $46B); normalize NAND pricing and YMTC's implied ~$6B quarter shrinks dramatically - NAND downcycles are brutal for the highest-cost producer
- Sanctions tax is structural: replacing Lam/AMAT/KLA tools with Naura/AMEC alternatives costs yield and throughput; YMTC plausibly has the worst unit economics of the top six, and no prospectus yet proves otherwise
- No meaningful DRAM/HBM exposure means YMTC misses the highest-margin AI memory pool while Samsung, SK hynix, and Micron harvest HBM profits to fund NAND price wars; its reported TSV-based HBM effort (TrendForce, Sep 2025) is early-stage
- Valuation gap is extreme: the Apr 2025 priced round says ~$22B; China Daily-reported forecasts say ~RMB 300B (~$44B); one outlet floats RMB 2-3T (~$280-420B). Only the first is a real mark, and the top of that range implies 12-18x in 14 months
- Governance: no controlling shareholder and a register dominated by state entities (Hubei Changsheng Development 26.54% largest, Hubei Ziguang Guoqi, Big Fund I+II); capital allocation serves national strategy (capacity at any cost), which historically destroys memory-industry returns
- US investors largely cannot own it anyway (A-share listing, Entity List name), while its success directly erodes Micron's and SanDisk's China and global pricing - it is a threat to hold-list names, not an opportunity
What it is worth
Last-priced-round anchor + peer EV/Sales cross-check + A-share scarcity-premium scenario (no DCF possible without disclosed financials)
~$20-25B
NAND cycle rolls over before listing, export controls tighten, IPO delayed or priced near the Apr 2025 round mark of ~$22B
~$45-70B
prospectus confirms scaled revenue but sanctions-taxed margins; 2-3x EV/S on a normalized ~$20B revenue base with a moderate A-share premium - consistent with the ~RMB 300B ($44B) reported market forecast at the low end
~$120-180B
supercycle revenue confirmed by prospectus, domestic toolchain yields prove out, STAR Market prices at the scarcity premium A-share semis command (still well below the RMB 2-3T chatter)
The only real mark is the Apr 2025 round: RMB 161.6B (~$22B) post-money at parent Changjiang Holdings. Cross-check: if the Q1 2026 implied ~$6B quarter annualizes to ~$20-24B revenue, $22B is around 1x EV/S versus NAND peers (Kioxia, SanDisk, Micron) at roughly 2-4x mid-cycle - so the last round already looks stale-cheap IF the supercycle revenue is real and durable, which is unproven. Forward markers: China Daily-reported market forecasts of ~RMB 300B (~$44B) for the IPO, versus one local outlet (YT Finance) floating RMB 2-3T (~$280-420B) - the latter implies 12-18x the priced round in 14 months and prices in both peak-cycle earnings and an extreme A-share scarcity premium; treat as sentiment, not valuation. US investors get exposure only via the read-through to MU/SNDK/WDC competitive pressure, not via ownership.
SWOT
Strengths
- Only mainland-China full-stack 3D NAND IDM (design + fab + packaging) - unmatched strategic position in the world's largest chip-consuming market
- Xtacking hybrid-bonding architecture is genuinely differentiated — separates CMOS logic from memory array, enabling TLC bit density above 20 Gb/mm2 (294L Gen5 shipping since early 2025)
- Deep-pocketed state backing — Big Fund I+II and Hubei/Wuhan vehicles among top shareholders, plus major state banks' investment arms in the Apr 2025 round - capital access does not depend on markets
- Rapid, third-party-verified share gains (8% to 13% in a year, Counterpoint) into a record $46B/quarter NAND market
Weaknesses
- US Entity List (Dec 2022) blocks access to leading US fab tools (Lam, Applied Materials, KLA), with aligned Japanese restrictions since 2023 - domestic-equipment substitution carries yield, throughput, and cost penalties
- No public audited financials — profitability, leverage, and true unit economics unverifiable until the prospectus
- NAND is the most commoditized memory segment — price-taker exposure in downcycles, and YMTC has no meaningful DRAM/HBM hedge today (only early-stage TSV-based HBM exploration reported by TrendForce, Sep 2025), unlike Samsung/SK hynix/Micron
- Governance opacity and state-directed capital allocation may prioritize national objectives over shareholder returns
Opportunities
- STAR Market IPO (tutoring accepted May 19, 2026 — application expected from mid-June 2026, CITIC Securities lead) unlocks large-scale public capital for expansion
- Enterprise/AI SSD mix shift (43% of NAND demand in Q1 2026, heading past 60% per Counterpoint) plays to high-layer TLC density where YMTC is competitive
- Phase III fab running >50% domestic tooling reaching mass production as early as H2 2026 would de-risk the sanctions chokehold and validate China's tool ecosystem
- Chinese cloud and smartphone OEMs actively qualifying domestic NAND for supply security - a captive demand pool incumbents cannot serve on equal political footing
Threats
- Further US export-control tightening (or enforcement against loopholes) could stall next-node scaling or capacity additions
- NAND supercycle reversal — the same 3.5x YoY industry revenue surge that flatters current estimates can unwind violently; rising memory costs are already dampening smartphone/PC demand (2026 PC shipments forecast -11.3%, IDC)
- Samsung, SK hynix, Kioxia, Micron, SanDisk all racing past 200+ layers - technology parity is perishable
- Geopolitical escalation (Taiwan contingency, sanctions on customers) could sever remaining foreign inputs or export channels
Moats, dependencies & bottlenecks
Moats
Separating CMOS periphery from the memory array gave YMTC density leadership at the TLC edge (294L/232 active Gen5, >20 Gb/mm2); incumbents are converging on similar bonding approaches, so the edge is real but perishable
Only mainland full-stack 3D NAND IDM; policy-driven procurement preference from Chinese OEMs, cloud, and government is not replicable by foreign rivals
Big Fund I+II, Hubei/Wuhan vehicles, major state banks' investment arms - survives downcycles that would bankrupt a private pure-play
~130k WSPM (~8% of global 3D NAND) is real scale but still trails Samsung and the Kioxia-SanDisk JV fab base; sanctions-era toolchain dilutes the cost benefit of scale
Dependencies
AMEC 688012.SS, ACM Research ACMR) Post-Entity List, next-node scaling depends on a maturing Chinese toolchain; Phase III's >50%-domestic flow reaching mass production (as early as H2 2026) is the critical test
Hubei Changsheng, state banks' investment arms) Funding is abundant but comes with state-directed priorities; IPO diversifies but does not remove this
Pure-play NAND exposure; current estimates ride a record $46B industry quarter with contract prices up ~55-60% QoQ
Entity List since Dec 2022; any tightening of foreign-direct-product rules or third-country tool loopholes hits the roadmap directly
Lenovo 0992.HK, Alibaba BABA, Tencent TCEHY, Huawei) Concentrated, policy-aligned domestic demand base; export channels are narrower due to sanctions stigma
Advantages
- Density leadership at the TLC leading edge (294L Gen5 X4-9070, >20 Gb/mm2, shipping since early 2025)
- Captive, policy-preferred domestic demand pool no foreign rival can access on equal terms
- Counter-cyclical state funding lets it expand into downturns when rivals cut capex
- IPO timing into a memory supercycle with AI storage demand at record highs
Weaknesses
- Structurally higher unit costs from sanctions-driven equipment substitution
- No meaningful DRAM/HBM revenue - largely locked out of the richest AI memory profit pool
- Financial black box until the prospectus lands; every revenue figure is an outside estimate
- Effectively uninvestable for most US investors (A-share venue, Entity List), capping its global capital pool
Bottlenecks
- Leading-edge fab equipment access — the binding constraint on layer-count scaling and capacity additions until domestic tools prove out
- Yield and throughput on the localized toolchain versus incumbents running full US/Japan tool flows
- No public-market currency until the STAR listing completes — the multi-stage CSRC process (application, inquiry, listing-committee review, registration) is still ahead as of Jul 2026
- Advanced packaging/test capacity for enterprise SSD controllers as mix shifts to data center
Top signals & trends
Top signals
First audited look at revenue, margins, and leverage - will confirm or destroy the third-party estimates
If yields are commercial, the sanctions-chokehold thesis weakens materially
8% to 13% in a year; watch whether share holds as YMTC ramps toward a stated ~15% supply goal while incumbents add virtually no 2026 capacity
Tightening third-country tool or subsystem loopholes is the highest-impact regulatory risk
Enterprise was 43% of NAND demand in Q1 2026, projected past 60% by year-end (Counterpoint); a rollover would deflate the whole cohort
Sets the A-share valuation benchmark for Chinese memory scarcity premiums ahead of YMTC's own pricing
Trends
Record $46B industry quarter (Q1 2026, +3.5x YoY); enterprise share of NAND demand shifting past 60% (Counterpoint)
YMTC is the NAND pillar of the national program; procurement, capital, and now capital-markets access all flow from it
Entity List constrains tools and export markets; escalation risk is permanent
YMTC at parity today (294L), but Samsung/SK hynix/Kioxia/Micron/SanDisk roadmaps compress any lead within quarters
Smartphone/PC NAND demand softening (2026 PC shipments forecast -11.3%, IDC) even as enterprise booms - YMTC's China OEM base skews consumer
Ecosystem & competitor graph
Suppliers feed the company; customers pull from it. Line thickness shows the strength of each tie (supply-chain dependency, customer earnings contribution). Hover to isolate a tie.
Lead domestic etch/deposition tool supplier for the localized flow
Domestic etch tools central to the sanctions-workaround toolchain
US-listed, Shanghai-operating wafer-cleaning tools used across China fabs
Former key etch/dep supplier - shipments halted by Dec 2022 Entity List; installed base remains
Former supplier, now blocked; a proxy for what YMTC lost
Japanese tools restricted under aligned Japan export controls since 2023
Chinese smartphone OEM qualifying domestic UFS/eMMC NAND
PC/server OEM using domestic SSDs in China-market SKUs
Chinese cloud - enterprise SSD demand pool under data-sovereignty procurement
Chinese hyperscaler NAND buyer
Chinese memory-module house packaging YMTC wafers into retail/embedded products
Private; strategic domestic customer for storage in phones, servers, and telecom gear
NAND share leader (29%, Q1 2026, Counterpoint); full memory portfolio incl. HBM; Korea-listed (005930.KS)
18% NAND share plus Solidigm; HBM profit engine funds NAND resilience; Korea-listed (000660.KS)
13% NAND share - now tied with YMTC; the US name most directly exposed to YMTC's China share gains after Beijing's 2023 Micron ban
14% NAND share; Tokyo-listed (285A.T); NAND pure-play peer and the closest business-model comp
13% NAND share post-WDC split; shares the Kioxia JV fab base; direct consumer/enterprise SSD rival
China's DRAM champion, further along in its own STAR IPO (listing-committee review since late May 2026) - sibling national champion, not a NAND competitor