AI Drives Prosperity Across the Entire Electronic Hardware Supply Chain; Focus on Tight Supply-Demand Segments
AI summary card
AI Drives Prosperity Across the Entire Electronic Hardware Supply Chain; Focus on Tight Supply-Demand Segments
J.P. Morgan is bullish on the explosive demand for MLCCs, high-end CCL/PCBs, IC substrates, and testing equipment driven by AI servers, noting that high-end capacity is in short supply, and recommends positioning in relevant industry leaders.
- AI server MLCC demand is growing at a CAGR of over 150%, driving a surge in high-end MLCC capacity consumption.
- High-end CCL (M8/M9 grade) is in short supply, with leading manufacturers like EMC continuously gaining market share.
- IC substrates are entering an unprecedented phase of supply tightness, with the trend toward larger sizes exacerbating capacity bottlenecks.
- Chroma dominates GPU SLT testing and data center power supply testing, benefiting from AI computing power expansion.
- YAGEO also benefits from AI infrastructure upgrades through its diversified portfolio in resistors, inductors, and other non-MLCC businesses.
Report interpretation
Overview
This report provides an in-depth analysis of the industrial transformation across four key sub-sectors in Asia—passive components, PCB/CCL, IC substrates, and testing equipment—amid the AI wave. The core conclusion is that the rapid adoption of AI servers is reshaping supply-demand dynamics in each sub-sector, leading to structural shortages in high-end products (e.g., high-capacitance MLCCs, M8/M9 grade CCL, large-format IC substrates, and SLT testing equipment). The firm maintains a positive view on relevant industry leaders, believing that companies with high-end capacity and technological moats will enjoy volume and price gains over the next 2-3 years.
Core views
In passive components, AI servers have become the core engine for MLCC growth. The report notes that the Total Addressable Market (TAM) for AI server MLCCs is expanding at a CAGR of over 150%. Next-generation chips like Blackwell and Rubin significantly increase unit consumption and specification requirements for high-end MLCCs, occupying substantial high-end capacity resources, thereby squeezing mid-to-low-end capacity and triggering a tight supply-demand balance across all categories. Although YAGEO has a relatively smaller share in MLCCs, its leadership in resistors, inductors, and tantalum capacitors, along with its penetration into server/HPC businesses, enables it to fully benefit from AI infrastructure construction, demonstrating a distinct growth logic. In the PCB/CCL sector, AI-driven High-Performance Computing (HPC) demand is fueling rapid growth in the high-end CCL TAM. From 2025 to 2027, CCL demand related to major AI server projects is expected to grow at a CAGR of 79%, with the proportion of M8 and M9 grade materials rising rapidly. Due to high technical barriers and slow yield ramp-ups for top-tier materials like M9, supply cannot quickly respond to the demand explosion, widening the supply-demand gap. As a leader in high-end CCL, Elite Material (EMC) holds a dominant share in key customer projects such as Nvidia and Google, and is poised to continue benefiting from both volume and price increases. The IC substrate industry is experiencing unprecedented supply tightness. As AI chip sizes continue to increase (e.g., Nvidia Rubin Ultra, Google Pumafish), substrate area and layer counts have risen significantly, causing effective capacity supply growth to lag far behind demand growth. The report forecasts a 67% CAGR in substrate demand for AI servers and switches from 2026 to 2028. However, industry capacity expansion is constrained, particularly by shortages of key upstream raw materials like T-glass, further aggravating supply bottlenecks. Leading manufacturers such as Unimicron and Ibiden will secure most high-end orders through Long-Term Agreements (LTAs) and technological advantages. In testing equipment, the increasing complexity of AI chips has significantly extended FT (Final Test) and SLT (System-Level Test) durations, raising the value per unit of equipment. Chroma ATE holds nearly 100% market share in Nvidia GPU SLT testing equipment and leads in data center power supply testing. With surging power consumption in AI data centers, demand for power testing equipment is growing exponentially. Chroma is expected to achieve rapid revenue and profit expansion in the coming years.
Analysis framework
The report employs a combined top-down and bottom-up analytical framework. First, by deconstructing AI server architectures (e.g., Nvidia Blackwell/Rubin, Google TPU), it quantifies specific usage changes (volume) across different chip generations for passive components, CCL, substrates, and testing, as well as ASP increases (price) driven by specification upgrades. Second, combining capacity expansion plans, yield ramp-up status, and upstream raw material supply constraints (e.g., T-glass) for each sub-sector, it builds supply-demand balance models to identify segments with structural gaps. Finally, based on technological moats in high-end product lines, customer certification progress, and market share data, it assesses earnings elasticity and provides valuation judgments.
Methodology notes
Supply-Demand Framework
Judges industry prosperity and price trends by comparing demand growth (driven by AI chip iterations) against supply growth (constrained by capacity expansion cycles and yield limits). This report focuses on analyzing supply-demand gaps in high-end CCL and IC substrates.
Upstream-Midstream-Downstream Value Chain Transmission
Analyzes how AI chip design (upstream) drives technical requirements and volume demand for packaging substrates, PCBs, passive components (midstream), and testing equipment (downstream), revealing value chain transmission paths.
PE/PEG Valuation
Evaluates the reasonableness of Price-to-Earnings (PE) ratios in conjunction with expected net profit growth rates (G), particularly applicable to high-growth AI hardware supply chain companies like Chroma and Elite Material.
Earnings Quality Analysis
Focuses on trends in Gross Margin (GM) and Operating Profit Margin (OPM) to assess the actual improvement in profitability resulting from product mix upgrades (e.g., transition to high-end MLCCs and M9 CCL).
Asset mapping & comparison
Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).
- YAGEO (2327.TW)Beneficiary
- Strengths
- Global leader in resistors; leading positions in inductors and tantalum capacitors; increasing share of server/HPC business; diversified product line reduces single-dependency risk.
- Weaknesses
- Relatively lower market share in MLCCs compared to Japanese peers like Murata; slower expansion pace in high-end MLCC capacity.
- Comparison
- Compared to pure-play MLCC manufacturers, YAGEO offers more stable AI exposure through its full-category passive component portfolio.
- Risks
- Recovery in traditional consumer electronics demand falls short of expectations; raw material price volatility.
- Elite Material (2383.TW)Beneficiary
- Strengths
- Global leader in high-end CCL (M7/M8/M9); primary supplier to core AI customers like Nvidia/Google; extremely high technological barriers.
- Weaknesses
- Product structure highly concentrated in the high-end market; significant volatility possible if AI capex slows.
- Comparison
- Market share and profitability in AI server CCL far exceed peers (e.g., Panasonic, Isola).
- Risks
- M9 material yield ramp-up below expectations; competitor technological breakthroughs.
- Unimicron (3037.TW)Beneficiary
- Strengths
- Global IC substrate leader; deeply integrated with major clients like Intel/Nvidia; leading FCBGA capacity scale; orders secured via long-term agreements.
- Weaknesses
- Massive upfront capex leading to high depreciation pressure; near-term margins sensitive to capacity utilization.
- Comparison
- Advantages in large-scale mass production capability and cost control compared to Ibiden and SEMCO.
- Risks
- Delays in downstream chip shipments; lag in new capacity commissioning.
- Chroma ATE (2360.TW)Beneficiary
- Strengths
- Exclusive/primary supplier of Nvidia GPU SLT testing equipment; leader in data center power testing solutions; strong technological monopoly.
- Weaknesses
- Extremely high customer concentration; heavy reliance on a few AI chip giants.
- Comparison
- Virtually no direct competitors in AI testing equipment; greater exclusivity in specific niches compared to Advantest/Teradyne.
- Risks
- Changes in testing technology roadmaps; customers developing in-house testing solutions.
Key data
- AI Server MLCC TAM CAGR (2025-2028)>150%Explosive growth in high-end MLCC demand driven by AI servers
- HPC CCL TAM CAGR (2025-2027)60%Major AI server projects drive expansion of the high-end CCL market
- M9 Grade CCL Share (2027e)8%Share of high-end materials rises rapidly from 0% in 2025 to 8% in 2027
- IC Substrate Demand CAGR (2025-2028)36%Sustained high growth in substrate demand for AI servers and switches
- Chroma SLT Equipment ASP (Rubin)~US$1.4mnValue per SLT testing unit rises significantly with increasing chip complexity
- Elite Material Share in Nvidia Projects (2026e)~38%EMC holds a dominant position in high-end CCL supply for key AI customers
Impact & implications
The report believes that accelerating AI infrastructure construction will profoundly impact the upstream hardware supply chain. For passive component manufacturers, the scarcity of high-end capacity will support firm pricing, benefiting companies capable of mass-producing high-capacitance MLCCs. For CCL/PCB manufacturers, faster technological iteration will widen the gap between leaders and tier-2/3 players, concentrating market share among leaders like Elite Material. For IC substrate manufacturers, prolonged supply tightness will enhance bargaining power, benefiting firms like Unimicron and Ibiden that have secured LTAs with major clients. For testing equipment vendors, the dual increase in AI chip testing duration and equipment value will deliver certain high earnings growth, with Chroma ATE as a primary beneficiary. Investors should focus on industry leaders possessing high-end technological moats and capacity advantages across sub-sectors.
Risks
- Slowdown or shortfall in AI server capex growth
- Yield ramp-up hurdles for high-end materials (e.g., M9 CCL, T-glass)
- Supply chain disruptions due to geopolitical risks
- Macroeconomic downturn impacting traditional consumer electronics demand
- Intensified industry competition leading to price wars
What to watch
- Shipment progress of Nvidia Blackwell/Rubin series chips and changes in substrate/CCL specifications
- Yield rates and capacity expansion of M9 grade materials by major CCL manufacturers
- Severity of supply tightness for upstream raw materials like T-glass for IC substrates
- Order confirmation pace for Chroma's next-generation GPU SLT testing equipment
- AI infrastructure investment guidance from major cloud providers (AWS, Google, Microsoft)