China Electronic Specialty Gases 2026 — The Semiconductor Localization Breakthrough

Industrial Research Institute | June 18, 2026

Chapter 1　Industry Overview and Definitions

1. From "Industrial Gas" to "Electronic Specialty Gas": The Value Transformation

Industrial gases are often underestimated in manufacturing due to their invisibility. But when ordinary elements enter a semiconductor fab, a fundamental value transformation occurs. Nitrogen purified to 99.99999% (7N) becomes a strategic material priced at tens of thousands of yuan per ton, while tungsten hexafluoride (WF₆) reached 5 million yuan per ton at its 2026 peak — multiples of gold by weight.

Electronic specialty gases (ESGs) are defined by two simultaneous requirements: application in semiconductor, panel, photovoltaic, LED, or optical fiber manufacturing; and purity at 5N (99.999%) or higher, with key grades reaching 7N (99.99999%), controlling impurities at ppb (10⁻⁹) to ppt (10⁻¹²) levels.

2. Five Dimensions Separating ESGs from Bulk Industrial Gases

First: Purity magnitude. Industrial N₂ (2N, 99%) vs. ultra-high purity electronic N₂ (6N, 99.9999%) differs in oxygen content by 10,000x. The engineering challenge of achieving 7N is not incremental — it is exponentially more difficult at each "N" step.

Second: Product complexity. Bulk gases number fewer than ten species; ESGs span over 300 varieties — fluorides (NF₃, WF₆, HF, CF₄, C₄F₆), hydrides (SiH₄, AsH₃, PH₃, GeH₄), nitrogen compounds (NH₃, N₂O), and rare gases (Kr, Xe, Ne, He at high purity). A mid-size 12-inch fab uses over 100 specialty gas types simultaneously.

Third: Certification barriers and stickiness. Introducing a new gas supplier at a fab takes 1–3 years of testing, making switching costs extremely high. Once qualified, a specialty gas supplier enjoys near-permanent customer retention.

Fourth: Packaging rigor. ESGs rely almost entirely on cylinder delivery with stringent internal surface treatment, valve specifications, and cleanroom filling procedures. High-value gases like WF₆ require nickel-alloy cylinders with precision passivation; any corrosion product contaminates the gas.

Fifth: Price differential of four orders of magnitude. Industrial liquid N₂ runs ~800 yuan/ton; ultra-high purity NF₃ runs 300,000–600,000 yuan/ton; 7N WF₆ reached 5 million yuan/ton in 2026 — a 4–5 order of magnitude premium encoding manufacturing difficulty, certification barriers, scarcity, and strategic value.

3. ESG Classification by Process Application

Etch and clean gases: NF₃ (CVD chamber dry clean), HF (wet etch), CF₄, C₄F₆, C₄F₈ (plasma dry etch of silicon and dielectrics), SF₆ (deep silicon etch), Cl₂ (metal etch). NF₃ alone accounts for ~37% of the global ESG market by volume.

Deposition gases (CVD/ALD): SiH₄ (silicon source for poly-Si/SiO₂/SiN), WF₆ (tungsten CVD for plugs and interconnects), NH₃ (nitrogen source for SiN), N₂O (oxygen source for SiO₂).

Dopant gases (ion implantation): AsH₃ (arsine, n-type), PH₃ (phosphine, n-type), BF₃ (p-type), GeH₄ (germanane, SiGe channel).

Carrier and purge gases: Ultra-high purity Ar, N₂, He, H₂ — used throughout the manufacturing process as carriers and protective atmospheres.

Photolithography laser gases: ArF excimer laser gas (Ar/F₂ mix, 193 nm), KrF (Kr/F₂, 248 nm) — the most technically demanding category with <5% domestic supply rate.

Calibration mixture gases: Precision blends for instrument calibration in fabs — highest margin subcategory, with gross margins up to 80%.

4. Strategic Value and "Chokepoint" Characteristics

ESGs directly represent 8–15% of wafer production cost, but their strategic weight far exceeds this fraction. A single month of WF₆ supply disruption in a fab producing 100,000 wafers/month would lose production valued at 200–300 million yuan — multiples of the entire monthly gas procurement budget. This "small-spend, high-dependency, large-loss-multiplier" characteristic elevates ESGs to a strategic tier far above their cost ratio.

Chapter 2　Global Landscape and China's Position

1. The Five Oligarchs and Their Moats

The global ESG market is dominated by five multinationals: Air Liquide (France), Linde plc (Germany/US), Air Products (US), Taiyo Nippon Sanso (Japan, Mitsubishi Chemical subsidiary), and Air Water (Japan). Together, they control over 90% of the global market — a dominance rooted in 50+ years of co-development with Intel, TI, Samsung, and TSMC.

Three mutually-reinforcing moats: decades of process know-how and formula databases; deep customer lock-in through qualification systems with very high switching costs; and global follow-manufacturing capability (building supply infrastructure alongside new fabs worldwide).

2. Global Market Size and Growth Structure

2025 global ESG market: $5.1 billion (narrow definition) or ~$110–125 billion (broad, including bulk electronic-grade gases), forecast to reach $193 billion by 2034 at CAGR of ~7.5%. The Asia-Pacific region accounts for ~69% of global consumption. China mainland grows fastest (8–10% annually), well above the global average.

3. China's Path from Consumer to Producer

Before 2018, China's domestic supply was negligible — national localization rate ~10%, with ~55% of imports from Japan and ~30% from Western companies. The pivot came from two directions: policy (National IC Fund Phases 1 & 2 including materials; "14th Five-Year" new materials plan) and market (U.S. export controls and supply-chain security consciousness forcing fabs to accelerate domestic supplier qualification).

By 2025, China's overall ESG localization rate reached 30–40%, with a clear three-tier structure: bulk electronic-grade gases >80%; mid-range (NF₃, WF₆, SiH₄, NH₃) at 25–65%; high-end (excimer photolithography gases, dopants) still below 10%.

4. NF₃ and WF₆: China's Two Leading Cards

NF₃: Global capacity ~40,000–45,000 MT/year; China accounts for ~60–65% (CNOOC Special Gas + Nanda Optoelectronics + Huahua + Jinhong). China is the world's #1 NF₃ producer nation, underpinned by fluorite resource advantage (China ~60% of global fluorite production) and domestic panel-market demand.

WF₆: Global effective capacity ~~8,000–9,000 MT/year in 2025; China's ~3,700 MT/year (~~40–45%). After Japan's Kanto Denka Kogyo and Central Glass permanently exit in July 2026 (combined ~2,200 MT/year), China's share surges past 60% — completing the most thorough single-category structural repositioning in Chinese ESG history.

5. Deep Structural Changes in the Competitive Landscape

Foreign majors have not retreated but chosen "deep localization" — building domestic capacity while maintaining technology moats in premium categories. Foreign ESG share in China has already fallen from ~90% in 2018 to ~60–70% in 2025 (across mid-range categories), with no reversal in sight. The more significant shift: Chinese fabs' buying power is growing. In 2025, major fabs can source multiple qualified domestic suppliers for most mid-range categories — a structural transition from "unilateral dependency" to "competitive multi-sourcing."

Chapter 3　Core Technology Systems

1. Purity Architecture: The Challenge Hierarchy from 3N to 7N

Each "N" step represents an order-of-magnitude reduction in impurity. 5N is the baseline ESG threshold; 6N is mainstream IC requirements; 7N is required for advanced nodes. Achieving 7N means allowing ≤0.001 grams of total impurities per ton of gas — and holding this precision in continuous industrial production for months without deviation. At 5 nm and below, trace metal requirements (Fe, Ni, Cr) fall below 0.1 ppb, a sub-ppb target that requires dedicated CRDS or ICP-MS instrumentation just to verify.

2. Synthesis Routes by Category

NF₃: Electrolysis of ammonium fluoride/hydrogen fluoride molten salt; the anode generates F₂ which reacts with NH₃ to form NF₃, followed by multi-stage purification. Power consumption is the dominant cost (~12,000–15,000 kWh/MT), driving producers toward low-electricity-price inland regions.

WF₆: Direct fluorination of high-purity tungsten powder (≥4N) with F₂ at ~350–400°C in sealed Ni-alloy reactors. Tungsten purity and reactor sealing determine the achievable purity ceiling.

SiH₄: Trichlorosilane (SiHCl₃) disproportionation — multiple staged reactions and distillation, ending in cryogenic purification.

NH₃ (ultra-high purity): Starting from industrial ammonia, multi-stage molecular sieve adsorption (dehydration) + cryogenic distillation + catalytic deoxygenation achieves 6N+.

3. Purification Technologies

Key purification methods include: cryogenic distillation (exploiting boiling-point differences through high-plate-count columns); multi-stage adsorption (PSA/TSA with molecular sieves for ppb-level water/O₂ removal); catalytic purification (converting residual impurities); and membrane separation. The combination design — which methods, in what sequence, with what parameters — is the proprietary engineering core of every leading ESG manufacturer.

4. Analysis and Metrology: The Eyes of Quality Control

Analytical capability directly determines quality assurance capability. Key instruments: GC (ppm-level routine analysis), ICP-MS (sub-ppb metal trace analysis, 300,000–1 million yuan/unit), CRDS laser spectroscopy (sub-ppb moisture in real-time), optical particle counters (online particulate monitoring). CRDS and high-end ICP-MS remain predominantly imported — a subtle dependency where even "proving the quality of one's own product" relies partly on foreign tools.

Chapter 4　Supply Chain Decomposition

1. Upstream Raw Materials: Strategic Resources

Fluorine-containing ESGs (NF₃, WF₆, HF, CF₄) trace to fluorite (CaF₂) → HF. China is the world's largest fluorite producer (~60% of global output), providing inherent cost advantage in the largest ESG sub-family (fluorine-based gases account for >60% of ESG market value). Tungsten (WF₆ precursor): China holds ~60% of global reserves and ~85% of production. The 2026 tungsten export controls that drove Japan's WF₆ exit demonstrated for the first time that China's mineral position can be deployed to reshape electronic materials supply chains directly.

2. Midstream Manufacturing: Value Architecture

The value model is "precision premium": NF₃ at 3N industrial grade is ~8,000 yuan/ton; at 5N electronic grade ~50,000 yuan/ton; at 7N ~100,000+. Each purity tier at minimum doubles value. Beyond manufacturing, gas companies add value through on-site services (installation, emergency response, waste gas treatment) and cylinder leasing (hardware-level customer lock-in).

3. Domestic Supply Chain Localization Assessment

Supply Chain Segment	Localization	Key Gaps	Outlook
Mineral raw materials	High (70–90%)	Minor high-spec imports	Essentially self-sufficient
Chemical intermediates (HF/W powder)	Medium-high (60–80%)	7N-grade W powder	80%+ by 2026–2028
Purification equipment	Medium (40–60%)	High-end multi-stage distillation	Gradual improvement
Analytical instruments	Low-medium (20–40%)	CRDS, high-end ICP-MS	Hard to break through before 2028
Premium cylinders & valves	Medium (50–70%)	Ni-alloy cylinders, ultra-low-leak valves	Further improvement possible
Gas manufacturing	Medium-high (30–65%, tiered)	Excimer gases, dopant gases	Continuous progress

The two core bottlenecks — analytical instruments and excimer photolithography gases — are where the industry's remaining structural gaps are concentrated. All other links have transitioned from "completely absent" to "specifically weak," representing a qualitative advancement.

Chapter 5　Downstream Application Structure

1. IC Wafer Fabs: 55–60%, the Largest and Most Demanding Segment

China mainland wafer capacity grew ~14% YoY in 2025 to ~10.1 million wafers/month (8-inch equivalent). Each additional 10,000 wafers/month of 12-inch capacity requires ~50–80 million yuan in annual ESG procurement.

SMIC: Primarily mature nodes (≥28 nm), actively pushing to 14 nm, multi-site expansion. Most comprehensive demand for NF₃, WF₆, SiH₄, ultra-high purity NH₃.

YMTC (Yangtze Memory): Most advanced domestic NAND, >200-layer 3D NAND. Per-wafer WF₆ consumption is far higher than planar NAND — with each tungsten word-line layer requiring a full CVD tungsten deposition cycle. Monthly capacity target: 150,000+ wafers/month. CNOOC Special Gas's WF₆ qualification at YMTC is one of the most landmark localization milestones.

CXMT (Changxin Memory): DRAM specialist, 2025 capacity ~200,000 wafers/month, target 300,000. Has explicitly set domestic material procurement targets; Jinhong Gas ultra-high purity NH₃ qualification is a landmark success.

2. Display Panels: 15–20%, NF₃'s Second-Largest Customer Group

TFT-LCD and OLED fabs share similar CVD-based process logic. China's panel fabs (BOE, CSOT) account for ~55% of global panel capacity, providing the earliest large-scale domestic NF₃ adoption — NF₃ localization in the panel segment now exceeds 80%, making it the most complete localization success story in any ESG category.

3. Photovoltaics: 10–15%, SiH₄ Primary Driver

HJT (heterojunction) solar cells require PECVD amorphous silicon deposition using SiH₄ — consuming 2,000–3,000 MT/GW of capacity, 3–5x more than PERC. As HJT capacity expanded through 2024–2025, SiH₄ demand surged. The challenge: photovoltaic-grade SiH₄ price collapsed alongside panel prices, compressing margins for suppliers targeting this market. The path to profitability lies in qualifying for the more demanding IC market (6N+ grade, 10x+ price premium).

4. LED and Compound Semiconductors: 5–10%, Ultra-High Purity NH₃ Core Market

GaN-based LED and power devices (GaN/SiC) require large volumes of ultra-high purity NH₃ (nitrogen source for MOCVD epitaxy) and H₂ (carrier gas). SiC power devices' rapid adoption in 800V EV platforms is creating additional SiH₄ demand (SiC epitaxy CVD).

Chapter 6　Key Player Analysis

1. International Majors: China Strategy

Air Liquide and Linde maintain deep local operations — Air Liquide with comprehensive coverage of domestic fabs including SMIC and CXMT; Linde expanding NF₃ and WF₆ capacity in Taiwan by 40% each (2024 announcement). Both have chosen "deep localization" — building domestic production capacity for mid-range grades while preserving technology moats in premium categories (excimer gases, ALD precursors, high-purity rare gases).

2. CNOOC Special Gas (688146): National Champion

2025 revenue: 2.26 billion yuan (+15.88%), net profit: 346 million yuan (+12.43%). Q1 2026 revenue: 701 million yuan (+36%), reflecting direct WF₆ market restructuring benefit. NF₃ capacity: 18,500 MT/year (domestic #1, global top 2). WF₆ capacity: 2,000 MT/year (world's largest single-base supply). Customers: TSMC, Micron, SK Hynix, SMIC, YMTC. WF₆ export average price in April 2026: +203.83% MoM; 7N grade: 5 million yuan/ton.

3. Huate Gas (688268): Mixed Gas Specialist

2025 revenue: 1.419 billion yuan (+1.70%), net profit: 135 million yuan (-26.75%). Short-term pressure from depreciation, but photolithography mixed gas (+36% volume growth) is the most competitive growth driver. Achieved 90%+ customer coverage among China's 8-inch and 12-inch IC manufacturers; 50+ product localization substitutions.

4. Jinhong Gas (688106): Ultra-High Purity Ammonia Pioneer

2025 revenue: 2.781 billion yuan (+6.65%), net profit: 132 million yuan (-48.65%). 10,000 MT/year ultra-high purity NH₃ built and operational; 20+ new semiconductor customers in 2025. High-leverage pressure (25 billion yuan in debt at year-end) is the key financial risk.

5. Nanda Optoelectronics (300346): Dual-Track Leader

2025 net profit: ~385 million yuan (+42% YoY). Phosphine mixed gas passed TSMC qualification (landmark for domestic dopant gas in advanced node). NF₃ capacity: 5,400 MT/year at Inner Mongolia + expansion of 2,000 MT planned. ALD precursor (Hf/Zr) rapidly scaling — highest-margin future growth engine.

6. Yarke Technology (002409): Comprehensive Platform

2025 revenue: 8.611 billion yuan (+25.49%), net profit: 1 billion yuan (+14.77%). ESG revenue: 417 million yuan (-11.12%) from price competition. Core value from precursors, electronic materials, and diversified customer coverage (TSMC, Samsung, Intel, SMIC).

7. Sinohua Technology (600378): Fluorine Integration Champion

2025 high-end fluorine materials revenue: 9.939 billion yuan (+30.83%), adjusted net profit +122%. Electronic chemicals: 1.138 billion yuan (+23.89%). WF₆ (600 MT/year) direct beneficiary of Japan exit.

Chapter 7　Localization Progress and Industry Data Insights

1. Three-Tier Localization Reality

Tier 1 — Substantially complete (>70%, ~20–25% of market value): Bulk electronic-grade N₂/O₂/Ar (>80%); panel-grade NF₃ (>80%); ultra-high purity NH₃ for LED/GaN (~60%+).

Tier 2 — Active substitution (25–60%, ~50–55% of market value): WF₆ (jumped from 35–40% to 60%+ after Japan exit); IC-grade NF₃ (35–50%); IC-grade SiH₄ (~~30%); G5 electronic HF (~~45%); high-purity H₂ (~40–50%).

Tier 3 — Difficult substitution (<25%, ~25–30% of market value): High-purity PH₃/AsH₃ dopant gases (15–20%); ArF/KrF excimer photolithography gases (<5%); high-end ALD precursors (<10%).

2. Industry Database Procurement Insights

Tianxia Gongchang, a B2B platform covering 4.8 million active factories, provides first-hand procurement trend data. Three notable observations: "domestic-first" buyer preferences in semiconductor materials procurement up ~40% vs. 2023; geographic dispersal of gas manufacturers from coastal to inland regions accelerating; procurement demand gaps for electronic HF, ArF gases, and electronic-grade ammonia most visible — precisely marking the next frontiers for localization investment.

3. Competitive Ranking (FY2025)

Growth quality: Nanda Optoelectronics (net profit +42%, dual-track) → Sinohua (+122% adj., WF₆ windfall) → CNOOC Special Gas (+12%, stable with WF₆ option) → Yarke (+15%, diversification buffer) → Huate (-27%, short-term pressure, long-term clear) → Jinhong (-49%, ramp pain).

Technology depth: Nanda (excimer/precursor, TSMC) → CNOOC (NF₃/WF₆ 7N engineering) → Huate (mixed gas depth, KrF) → Sinohua (fluorine vertical integration) → Yarke (breadth over depth) → Jinhong (ultra-pure NH₃, narrow focus).

Chapter 8　Pricing and Business Models

1. ESG Price Spectrum

Ultra-premium (>1 million yuan/MT): WF₆ 7N (5 million yuan/MT in 2026), GeH₄, AsH₃, ArF excimer gas
Premium (100,000–1 million yuan/MT): NF₃ 5N (300,000–600,000), PH₃, C₄F₆, IC-grade SiH₄ 6N (150,000–300,000)
Mid-range (10,000–100,000 yuan/MT): Ultra-high purity NH₃ 6N+ (80,000–150,000), high-purity H₂ 6N (30,000–80,000), G5 electronic HF (50,000–100,000)
Lower range (<10,000 yuan/MT): High-purity N₂, O₂, Ar

The pricing structure directly correlates with proximity to core process steps: gases used in photolithography, ion implantation, and ALD are the most expensive and hardest to localize; gases used in peripheral/utility steps are less expensive and more easily substituted.

2. Three Business Models

Cylinder gas supply (most common): Cylinder delivery, billed by weight. Margin depends on product category — premium grades 30–50%+, commodity grades 10–25%.

Long-term framework agreements (LTAs): Multi-year contracts with annual price review and raw material cost pass-through clauses. CNOOC Special Gas successfully inserted cost pass-through provisions into 2024–2025 LTAs, enabling effective price increases when WF₆ costs surged.

On-site generation (BOT/Take-or-Pay): Only for bulk electronic-grade gases (N₂/O₂/H₂/Ar). Gas company builds dedicated ASU equipment at fab site, supplies via pipeline, 15–20 year contracts. Guanggang Gas's Pearl River Delta on-site network exemplifies this model's stable cash flow quality.

Custom precision mixing (highest margin): 50–80% gross margins; relies on precise blending equipment and deep formula databases accumulated through years of fab-specific collaboration.

Chapter 9　Client Case Studies and Project Timelines

1. Nanda Optoelectronics + TSMC: The Four-Stage Qualification Journey

Stage 1 (6–12 months): Lab evaluation. Parallel testing vs. incumbent Japanese supplier across ICP-MS metal trace (sub-ppb), CRDS moisture (<0.5 ppm), wafer-level MOSFET uniformity, and 6-month batch consistency. Free sample provision of hundreds of kg to multi-ton volumes.

Stage 2 (6–12 months): Production line validation. Hundreds of production lots with SPC-based statistical verification that critical device parameters (Vth, SS, DIBL) show no significant deviation. Any excursion requires root-cause analysis within 24–48 hours.

Stage 3 (3–6 months): Supplier quality audit. TSMC QA team on-site inspection of hazardous chemical management, ISO 9001 execution, emergency supply plans (24-hour guaranteed delivery), and full lot traceability.

Stage 4: Production ramp. Entering at 10–20% share, gradually increasing. Total timeline: ~2 years. Nanda entered as TSMC's second qualified phosphine supplier (alongside incumbent Japanese supplier) — the most significant domestic dopant-gas qualification milestone to date.

2. CNOOC Special Gas + YMTC: The Domestic Partnership Template

WF₆ is the only gas source for CVD tungsten deposition, and 3D NAND with 200+ layers consumes WF₆ proportional to layer count. YMTC's WF₆ demand is among the highest of any single fab globally. CNOOC Special Gas had to invest in ICP-MS instruments (7-figure yuan) and Class 100 clean analytical labs (tens of millions of yuan) to provide reliable quality certificates for sub-ppb purity claims. The three compelling advantages for YMTC: 24-hour local replenishment (vs. weeks for imports); cost savings from elimination of import premium; and supply chain security (no export control risk).

3. Samsung's Supply Chain Restructuring and China's Elevated Position

Samsung's "3-source" strategy originally relied on SK Materials (internal), Kanto Denka (Japanese), and Central Glass (Japanese). With two Japanese suppliers exiting in July 2026, CNOOC Special Gas became the essential replacement, elevating its position from "backup source" to "primary source" in Samsung's WF₆ supply chain — a structural status upgrade that reflects the new geopolitical reality of semiconductor material supply chains.

Chapter 10　Investment and M&A Landscape

1. A-Share ESG Public Company Overview

15+ A-share companies touch the ESG category as of June 2026: pure-play ESG targets (CNOOC Special Gas, Huate Gas, Nanda Optoelectronics), integrated electronic chemicals platform (Yarke Technology), fluorine-to-ESG (Sinohua Technology), bulk+specialty dual-track (Jinhong Gas, Heyuan Gas), electronic bulk on-site (Guanggang Gas).

2. National IC Fund Phase 3: The Coordination Logic

Phase 3 (3,440 billion yuan, 2024, 15-year term) differentiates from Phases 1–2 by targeting "strength from existing foundation" rather than "building from zero." 70% of capital allocated to equipment and materials. Key intervention modes: indirect equity stakes through intermediate platforms (providing political trust signals that accelerate fab qualification timelines); preferential-rate debt for high-risk R&D projects that pure market capital would over-price; top-down coordination (nudging fabs to fast-track specific supplier qualifications).

3. Two Consolidation Models

Horizontal M&A (category expansion): Yarke Technology acquired Chengdu Cometech (fluorine ESGs) and Jiangsu Xianke (precursors) to build a "comprehensive electronic materials platform." The logic: same fab customer base serves multiple product categories; one relationship development effort monetizes across many SKUs.

Vertical integration (raw material control): Sinohua Technology integrated from fluorite mining through HF production through fluorine ESGs — the most complete fluorine chemical value chain in China, providing inherent cost advantages and raw material security.

Chapter 11　Policy and Standards

1. IC Fund Phase 3 Strategic Intent

Phase 3's electronics materials focus signals that the next upgrade bottleneck in China's semiconductor chain is not chip design (HiSilicon, Haiguang exist) nor fabrication capability (SMIC has 14 nm) — but rather the critical materials and equipment supporting advanced nodes. ESGs sit precisely at this strategic position.

2. Tungsten Export Controls: The First Mineral Weapon in Electronic Materials

China's early 2026 tungsten export controls are historically significant: the first time China directly used rare mineral export control tools to reshape a specific electronic materials supply chain. The mechanism was precise: China dominates global tungsten (60% reserves, 85% production); Japan's two WF₆ producers were 100% dependent on Chinese tungsten powder; controls cut off raw materials; Japan exits permanently; China's manufacturers absorb the 2,200 MT/year gap with structural pricing power.

Following 2023's gallium and germanium export controls, this action establishes a precedent: China will use its mineral dominance as a policy instrument in semiconductor materials — a reality that every supply chain strategist must now factor into long-term planning.

3. Safety Regulation and ESG Compliance

ESG production involves multiple classes of extremely hazardous chemicals (acutely toxic: AsH₃, PH₃; pyrophoric: SiH₄; corrosive: HF). The 2025 specialty gas safety rectification campaign raised industry entry barriers, eliminating weakly-managed small operators and improving competitive discipline. Multiple resource qualification requirements (highly toxic chemical operating license, hazardous chemical safety production license) form meaningful regulatory moats for established players.

Chapter 12　3–5 Year Outlook and Research Judgments

1. Three Demand Drivers

Driver 1: Continued wafer capacity expansion (highest certainty). SMIC multi-site expansion, YMTC Phase 3, CXMT post-IPO scale-up — each corresponds to quantifiable gas demand increments. Even if growth decelerates to 8–10% annually after 2026, the absolute increment remains substantial.

Driver 2: Process migration multiplier. Advanced nodes require more process steps and higher-quality (higher-priced) gases per wafer. Production capacity growth of 10% + process-driven consumption intensity growth of 10% ≈ ~20% actual gas demand growth.

Driver 3: Third-generation semiconductor new demand. SiC power device penetration in 800V EV platforms (projected to rise from 15% to 40%+ in 2025–2030) creates substantial SiH₄ demand for SiC epitaxy CVD — becoming a fourth demand pillar alongside IC, panel, and photovoltaics.

2. Three Supply-Side Evolutions

Evolution 1: Low-end commodity category consolidation — high-purity N₂/O₂, basic SiH₄, standard NF₃ face ongoing price pressure and capacity exit, leading to higher industry concentration over 2–3 years.

Evolution 2: Premium category breakthrough — ArF/KrF excimer gas localization to reach 5–10% by 2027; G5 electronic HF localization to exceed 65% by 2028; each breakthrough creates "winner-takes-most" dynamics due to high certification barriers for latecomers.

Evolution 3: WF₆ super-cycle and normalization — CNOOC Special Gas is planning +1,000 MT/year WF₆ capacity (2027 production). Price premium persists 3–5 years before normalization as global non-Chinese capacity gradually rebuilds.

3. Research Institute Comprehensive Assessment

Tianxia Gongchang, a B2B data platform covering 4.8 million active factories, provides three-year trajectory data on domestic ESG penetration into downstream procurement. Integrating industry dynamics, competitive analysis, and policy ecosystem, key research judgments are:

Short-term (1–2 years, 2026–2027): WF₆ and IC-grade NF₃ are the highest-certainty premium categories; Huate Gas's potential ArF excimer breakthrough is the most significant optionality; NH₃ inflection point at Jinhong/Heyuan expected Q4 2026–Q1 2027. Maintain "premium outperforms, commodity under pressure" differentiation framework.

Medium-term (3–5 years, 2028–2030): ALD precursors (Nanda's strongest first-mover advantage); G5 HF substitution substantially complete; SiC-driven SiH₄ uplift; industry consolidation accelerates — top-5 players to control 70%+ of industry; China ESG market to exceed 40 billion yuan before 2028.

Long-term structural challenge (post-2030): ArF/EUV photolithography gases remain the core structural gap. Their localization depends on domestic lithography ecosystem maturation. The breakthrough, when it comes, will be more consequential than NF₃ or WF₆ — it will mark the completion of China's semiconductor materials self-sufficiency at the process layer.

Chapter 13　Risk Assessment

1. Supply Glut and Price Compression (Most Immediate Risk)

Multiple companies simultaneously expanded NF₃, SiH₄, and ultra-high purity NH₃ capacity in 2025. Huate Gas net profit -26.75%, Jinhong Gas -48.65% in 2025 — not projected risks but realized profit pressure. The key differentiation for investors: distinguish "industry-wide short-term pressure" (most companies' margins under pressure) from "structural value impairment for leading players" (which is not happening).

2. Wafer Expansion Schedule Uncertainty

ESG demand depends fundamentally on fab capacity expansion, which is subject to: export controls on EUV and advanced equipment limiting leading-edge expansion; memory price cycles (NAND/DRAM price crashes can slash capex); macro demand shocks reducing terminal electronics demand. China's semiconductor expansion plans have delayed before; investors should use actual production milestones rather than announced timelines for demand modeling.

3. Geopolitical Risk: Bilateral Exposure

China's ESG companies face bilateral geopolitical exposure: cross-strait/PRC-Korea tensions create revenue risk from TSMC and Samsung contracts; China's mineral export controls (tungsten) that benefited WF₆ suppliers also create retaliatory risk — other countries may respond with counter-restrictions on equipment or chemical inputs. This bilateral nature makes geopolitical risk impossible to evaluate as uniformly positive or negative, requiring continuous monitoring.

4. High-Leverage Liquidity Risk

Jinhong Gas: ~29.88 billion yuan in interest-bearing debt (year-end 2025) against net profit of only ~1.32 billion yuan. If capacity utilization fails to ramp as expected, debt service stress rises meaningfully. A liquidity event at any mid-sized operator could trigger broad-market anxiety about the sector even if leading-company fundamentals are intact.

5. Long Certification Cycles and Reversal Risk

Certification cycles of 3+ years for premium categories, with uncertain outcomes, represent significant sunk-cost risk for companies investing heavily in qualification. Even existing qualifications may require re-qualification when fabs upgrade to new process nodes — "certification never ends." No ESG market position is permanently secure; competitive advantages require continuous re-validation.

Data Sources

Data for this report from the following sources as of June 18, 2026. All figures have been cross-validated across multiple sources; reasonable variation from actual figures is expected, and readers should consult primary sources for independent verification:

Tianxia Gongchang B2B Industrial Database (procurement and supply-demand trend observations across 4.8 million active factories)
Company reports and annual disclosures (FY2025): CNOOC Special Gas (688146), Huate Gas (688268), Jinhong Gas (688106), Nanda Optoelectronics (300346), Yarke Technology (002409), Sinohua Technology (600378), Heyuan Gas (002956), Guanggang Gas (688239)
Persistence Market Research, Electronic Specialty Gases Market Forecast to 2032 (2025)
Custom Market Insights, Global Semiconductor Specialty Gas Market Size 2025–2034 (2025)
Valuates Reports, Special Gas (NF₃) Market 2025–2031 (2025)
SEMI World Fab Forecast — China wafer capacity data (March 2025 edition)
Everbright Securities, Industrial Gas Sector: ESG Localization Wave Has Arrived (March 2025)
Orient Securities, CNOOC Special Gas Deep Dive (March 2024)
Zhiyan Consulting, 2025 ESG Market Size and Major Player Market Share Analysis (2025)
China Securities Pengyu, Semiconductor Materials: Industry Recovery, High-End Localization Still Needs Improvement (November 2024)
Financial media tracking of listed company announcements (April 2026)
JW Insights, 2025 China ESG Listed Company Research Report (2025)
Shenqitou Industrial Research Institute, 2025 ESGs: Wide Localization Potential, Capacity Glut Risk Emerging (December 2025)