Jianghai's 2Q26 results missed Goldman Sachs' expectations, while AI data center business delivery remains insufficient to support the current valuation

Overview

This report assesses Jianghai's 2Q26 results, shipment and customer certification progress for capacitor products related to AI data centers, and whether capacity expansion can translate into verifiable revenue. Goldman Sachs acknowledges growth opportunities arising from recent tight supply, price increases, and AI demand, but believes actual orders and certification progress remain insufficient to support the company's planned capacity utilization and current valuation, and therefore maintains its Sell rating.

Core views

Jianghai's 2Q26 revenue, gross profit, operating profit, and net profit were Rmb1.767bn, Rmb435mn, Rmb251mn, and Rmb221mn, respectively, representing yoy growth of 15%, 12%, 11%, and 7%, but coming in 6%, 8%, 8%, and 14% below Goldman Sachs' expectations, respectively. The market's key focus is shipments of AI data center products and certification by new customers, while the identifiable incremental contribution remains limited at this stage: 2Q26 industrial control and UPS revenue was approximately Rmb579mn, higher than Rmb497mn in 2Q25, but it has yet to demonstrate a significant incremental contribution from AI. By business segment in 1H26, aluminum electrolytic capacitor revenue was Rmb2.498bn, up 12% yoy, with a gross margin of 27.0%, up 0.3 percentage points yoy; film capacitor revenue was Rmb311mn, up 34% yoy, with a gross margin of 15.2%, up 0.8 percentage points yoy; supercapacitor revenue was Rmb249mn, up 54% yoy, with a gross margin of 19.5%, up 2.7 percentage points yoy. Despite revenue growth, the overall gross margin in 2Q26 declined by 0.8 percentage points yoy and 1.3 percentage points qoq, mainly due to annual product price reductions of 3%-5% and rising raw material costs. Price increases ranging from 4.5%-20% are expected to be fully implemented only in 3Q26; management expects gross margin to improve sequentially from August, and Goldman Sachs also expects a modest recovery in 2H26E. Aluminum electrolytic capacitors represent the most scalable part of the AI data center opportunity, but a gap remains between capacity plans and validated demand. Traditional products have an average selling price of approximately Rmb10-12 per unit, monthly capacity of approximately 30mn units, and annual capacity of approximately 360mn units, corresponding to annual output value of approximately Rmb4.3bn, broadly in line with 2025 revenue, implying that further growth in traditional and non-AI products will still require capacity expansion. Among products for AI data centers, approximately 20% are priced at Rmb20-30 per unit, approximately 60% at Rmb40-60 per unit, and the remaining products at higher prices; current monthly capacity is approximately 2mn units and is expected to increase to 4mn units by end-2026E and 10mn-12mn units by end-2027E. Management stated that August shipments of aluminum electrolytic capacitors for AI data centers were approximately 2mn units, mainly supplied to Megmeet, corresponding to revenue of approximately Rmb108mn, and expects to secure additional orders from Delta Electronics in September, but Goldman Sachs is still awaiting validation. Based on the tight supply environment described by management and assuming 80% capacity utilization through year-end, Goldman Sachs incorporates approximately Rmb480mn of AI-related revenue into its 2026E forecast; for 2027E, it assumes Jianghai obtains a 20% market share, broadly consistent with management guidance. However, potential capacity corresponds to output value of approximately Rmb4.5bn, and Goldman Sachs has yet to see sufficiently concrete demand drivers to support full utilization. EDLC has begun shipping for AI applications, with cumulative shipments of approximately 1mn units, but AI-related supercapacitor revenue in 1H26 was only approximately Rmb17mn. Current EDLC monthly capacity is approximately 1mn units, while Goldman Sachs estimates industry-wide monthly demand of approximately 8.3mn units in 2026E, corresponding to an approximately 13% market share for Jianghai; the company plans to expand monthly capacity to more than 3mn units in 2027. Goldman Sachs believes EDLC is more likely to be a transitional solution, with demand potentially shifting gradually toward LIC as technology adoption evolves. Current LIC monthly capacity is approximately 50k units, with a target of more than 100k units by end-2026E and 300k units by end-2027E. Capacity expansion is relatively cautious because most customers remain in the testing and certification stage, and demand has yet to fully translate into firm orders. The company is optimistic about retrofit demand from installing supercapacitors in legacy servers, believing that this could expand the potential market beyond incremental demand from new AI servers; Goldman Sachs emphasizes that completing customer certification and actually capturing demand are the more critical factors to monitor. MLPC has not yet entered formal mass production for AI servers and is still undergoing technical performance, reliability, and environmental validation, while being jointly developed with domestic strategic customers. The company's current total MLPC monthly capacity is approximately 20mn units, of which approximately 5mn units can be repurposed for AI applications by modifying existing production lines; it will continue upgrading production lines and determine whether to add capacity based on validation progress and customer adoption. Goldman Sachs believes customer certification for LIC and MLPC remains limited and that product quality still trails leading peers, so their market shares cannot yet be assessed as those of mature businesses. On the supply chain side, the key upstream materials for aluminum electrolytic capacitors are etched foil and formed foil. Jianghai internally supplies more than 90% of its high-specific-capacitance foil and is expanding upstream material and capacitor capacity in tandem, making related supply pressure relatively manageable. The company's investment thesis also highlights a formed foil self-sufficiency rate of approximately 75%, higher than approximately 30% for Japanese peers, which the report says can provide an approximately 18% gross margin advantage and enhance control over product performance. In contrast, high-end activated carbon required for EDLC could become a significant bottleneck; management said supply is extremely tight, and Kuraray reported a 30% price increase. LIC is similarly affected by carbon material supply, but its usage per unit is lower than that of EDLC. The company expects to pass upstream price increases on to customers. Electricity costs are also a potential source of margin pressure. Electricity prices in Xinjiang are broadly stable at Rmb0.31/kWh, but the gradual phaseout of preferential electricity tariffs for certain strategic emerging industries in Inner Mongolia could increase electricity costs by Rmb0.01-0.02/kWh. Some formed foil plants in Urumqi and Kuitun were affected by electricity prices and environmental factors in 1H26; the company is studying mitigation measures such as direct procurement of green electricity, improved energy efficiency, and stronger supply chain control, and expects regional electricity prices and environmental requirements to gradually converge over the medium term. Goldman Sachs believes Jianghai's positives include vertical integration in key aluminum electrolytic capacitor processes and an industrial customer base with high barriers to entry and strong brand recognition, capabilities that can be extended to other product lines. The report expects demand from AI data centers for aluminum electrolytic capacitors, LIC, and MLPC to become the main earnings driver beginning in 2026E, with potential support from customer wins, higher capacitor content per rack, price increases, and next-generation power architectures. However, these opportunities have yet to be fully realized, while the recent share-price rerating has moved ahead of medium-term fundamentals. Reverse valuation analysis further supports Goldman Sachs' cautious view. At 2028E P/E multiples of 40x, 30x, and 25x, the current share price implies that Jianghai would capture 48%, 64%, and 77% shares of the AI data center aluminum electrolytic capacitor market, respectively, or 41%, 55%, and 66% shares of the LIC market, all above Goldman Sachs' base-case forecasts. Goldman Sachs therefore believes the market has priced in overly optimistic opportunities and market share gains for aluminum electrolytic capacitors, supercapacitors, and MLPC. Goldman Sachs lowered its 2026-30E EPS forecasts by 0%-1% and set its 12-month target price at Rmb54.4, with the valuation still based on a 30x 2028E P/E multiple and discounted to 2027E at a 10.5% cost of equity. Given results below expectations, insufficient realization of shipments and certifications, and a valuation already reflecting aggressive market share assumptions, Goldman Sachs reiterates its Sell rating.

Analysis framework

The report first compares 2Q26 revenue and profit with their yoy performance and Goldman Sachs' expectations, and then breaks down 1H26 revenue and gross margins by aluminum electrolytic capacitors, film capacitors, and supercapacitors. It subsequently examines, item by item, the selling prices, capacity, shipments, customer certification, and potential orders for AI data center-related aluminum electrolytic capacitors, EDLC, LIC, and MLPC, translating capacity utilization and market share assumptions into revenue opportunities. Finally, the report assesses the margin trajectory by considering the supply chain, price increases, and electricity costs, and evaluates whether expectations are excessive through P/E valuation and the market shares implied by the current share price.

Methodology notes

• Valuation MethodP/E and PEG Valuation

  Forward P/E valuation and cost-of-equity discounting

  Goldman Sachs calculates value by applying a 30x P/E multiple to 2028E earnings and then discounts it to 2027E at a 10.5% cost of equity, deriving a 12-month target price of Rmb54.4.

• Valuation Method

  Reverse-engineering the market share implied by the current share price

  Under different 2028E P/E assumptions, the report reverse-engineers the share of the AI data center aluminum electrolytic capacitor or LIC market that the company would need to capture to justify its current share price and compares it with Goldman Sachs' base-case forecast to determine whether market expectations are overly aggressive.

• Industry/Sector Analysis FrameworkSupply-demand framework

  Alignment of capacity, customer demand, and raw material supply

  The report compares the company's capacity expansion plans with validated orders, customer certification, and industry demand, while analyzing supply constraints for formed foil and high-end activated carbon to determine whether capacity can be fully utilized.

• Industry/Sector Analysis FrameworkVolume-price decomposition

  Shipment volume, utilization, selling prices, and revenue estimates

  Based on the price ranges, shipment volumes, and an 80% utilization assumption for AI capacitor products, the report estimates 2026E AI-related revenue and analyzes subsequent margin changes in conjunction with the magnitude of price increases.

• Competitive and Strategic FrameworkValue chain analysis

  Vertical integration of upstream materials

  Using the internal supply ratios of formed foil and high-specific-capacitance foil, the report assesses Jianghai's competitive advantages in raw material security, costs, gross margins, and control over product performance.

Asset mapping & comparison

Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).

• Nantong Jianghai Capacitor Co., Ltd. (002484.SZ)

  AI data centers offer potential growth opportunities for aluminum electrolytic capacitors, EDLC, LIC, and MLPC, but actual shipments, certification, and market share remain insufficient to support the current valuation.

  Strengths

  Vertical integration in key aluminum electrolytic capacitor processes, an approximately 75% formed foil self-sufficiency rate, a solid industrial customer base, and the ability to extend its brand and customer relationships to other capacitor products.

  Weaknesses

  2Q26 results missed Goldman Sachs' expectations, LIC and MLPC remain in customer testing or certification, MLPC has not yet entered formal mass production, and some products still trail leading peers in quality.

  Comparison

  The formed foil self-sufficiency rate is approximately 75%, higher than approximately 30% for Japanese peers; however, the AI-related market shares implied by the current share price exceed Goldman Sachs' base-case forecasts.

  Risks

  If LIC and MLPC certification, market shares, or AI data center demand and pricing progress significantly faster than Goldman Sachs expects, the cautious assessment underlying the Sell rating could face upside risk.

Key data

• 2Q26 RevenueRmb1,767mnUp 15% yoy, 6% below Goldman Sachs' expectation
• 2Q26 Gross ProfitRmb435mnUp 12% yoy, 8% below Goldman Sachs' expectation
• 2Q26 Operating ProfitRmb251mnUp 11% yoy, 8% below Goldman Sachs' expectation
• 2Q26 Net ProfitRmb221mnUp 7% yoy, 14% below Goldman Sachs' expectation
• 1H26 Aluminum Electrolytic Capacitor Revenue and Gross MarginRmb2,498mn; 27.0%Revenue up 12% yoy, gross margin up 0.3 percentage points yoy
• 1H26 Film Capacitor Revenue and Gross MarginRmb311mn; 15.2%Revenue up 34% yoy, gross margin up 0.8 percentage points yoy
• 1H26 Supercapacitor Revenue and Gross MarginRmb249mn; 19.5%Revenue up 54% yoy, gross margin up 2.7 percentage points yoy
• 2Q26 Industrial Control and UPS RevenueApproximately Rmb579mnApproximately Rmb497mn in 2Q25, with the incremental AI contribution still limited
• August Shipments of Aluminum Electrolytic Capacitors for AI Data CentersApproximately 2mn units; approximately Rmb108mn revenueMainly supplied to Megmeet, with subsequent orders still awaiting validation
• 2026E AI-Related Revenue ForecastApproximately Rmb480mnAssumes 80% capacity utilization by year-end
• Monthly Capacity of Aluminum Electrolytic Capacitors for AI Data CentersCurrently approximately 2mn units; 4mn units by end-2026E; 10-12mn units by end-2027EPotential capacity corresponds to output value of approximately Rmb4.5bn, but demand for full utilization has yet to be validated
• EDLC ScaleCumulative shipments of approximately 1mn units; 1H26 AI-related revenue of approximately Rmb17mnCurrent monthly capacity is approximately 1mn units, equivalent to approximately 13% of 2026E industry demand
• LIC Monthly Capacity PlanCurrently approximately 50k units; more than 100k units by end-2026E; 300k units by end-2027EMost customers remain in the testing and certification stage
• MLPC Monthly CapacityTotal capacity of approximately 20mn units, of which approximately 5mn units can be repurposed for AI applicationsHas not yet entered formal mass production for AI servers
• Product Price Increases4.5%-20%Expected to be fully implemented in 3Q26 to offset annual price reductions of 3%-5% and rising raw material costs
• EPS Forecast Revision2026-30E lowered by 0%-1%Slightly reduced following this earnings update
• 12-Month Target PriceRmb54.4Based on a 30x 2028E P/E multiple and discounted to 2027E at a 10.5% cost of equity

Impact & implications

The report believes AI data center demand could drive Jianghai's aluminum electrolytic capacitor, LIC, and MLPC revenue, but capacity expansion does not necessarily equate to realized demand: customer certification, firm orders, and product quality will still determine actual utilization. Near-term price increases could help gross margin improve from August 2026, but raw material and electricity costs are offsetting factors. Because the current valuation already implies market shares above Goldman Sachs' base-case forecasts, even slightly slower business delivery would make existing expectations difficult to support.

Risks

• LIC and MLPC customer certification may progress faster than Goldman Sachs expects.
• The company's market shares in LIC, MLPC, and aluminum electrolytic capacitors may exceed Goldman Sachs' expectations.
• AI data center demand, growth in capacitor content per rack, and product pricing may be stronger than Goldman Sachs expects.
• Tight supply of high-end activated carbon required for EDLC could become a capacity expansion bottleneck, and upstream suppliers have already reported a 30% price increase.
• The gradual phaseout of preferential electricity tariffs in Inner Mongolia could increase electricity costs by Rmb0.01-0.02/kWh.

What to watch

• Verify whether the company's stated August shipments of approximately 2mn aluminum electrolytic capacitors for AI data centers can be sustained.
• Monitor whether the Delta Electronics orders expected by management in September materialize.
• Track the progress of LIC and MLPC customer testing, certification, and conversion into formal orders.
• Observe whether gross margin can improve sequentially from August after the 4.5%-20% price increases are fully implemented in 3Q26.
• Test the assumptions of 80% capacity utilization by end-2026E and approximately Rmb480mn of AI-related revenue.
• Monitor the pace of the transition from EDLC to LIC and whether retrofit demand from installing supercapacitors in legacy servers can create an incremental market.