Rising AI server demand is squeezing China’s semiconductor supply chain, with MLCC makers and MCU suppliers facing capacity constraints.

By CII (China Industry Intel) – Contributing Analyst | June 20, 2026

AI Server Demand Squeezes China’s Chip Supply Chain From MLCCs to MCUs

China’s semiconductor supply chain is showing fresh signs of pricing strain as surging demand for AI servers collides with capacity constraints in passive components and microcontrollers. The squeeze, reported by DIGITIMES on June 18, spans the entire value chain from high-end multilayer ceramic capacitors (MLCCs) used in AI accelerators to basic 8-bit microcontrollers in industrial and consumer applications.

The pressure is not new — component shortages have plagued the industry since the pandemic — but the AI boom has added a new dimension. Every Nvidia H100 or H200 server rack requires thousands of high-capacity MLCCs, and the rapid scaling of AI data centers in China and globally is absorbing capacity that would otherwise serve automotive, industrial, and consumer electronics customers.

The MLCC crunch: AI’s hidden bottleneck

MLCCs — tiny ceramic capacitors that store and regulate electrical charge — are among the most ubiquitous components in electronics. A single AI server can contain 15,000-20,000 MLCCs, roughly 3-5 times more than a traditional server. With global AI server shipments projected to exceed 2 million units in 2026, the demand for high-end MLCCs has effectively doubled capacity utilization at major suppliers.

Taiyo Yuden, the Japanese MLCC maker, has been raising prices since early 2026 and recently announced plans to expand AI server MLCC capacity. Murata Manufacturing, the world’s largest MLCC producer, is reportedly considering its own price increases. Samsung Electro-Mechanics is expected to follow. The pricing power has shifted decisively to suppliers.

Ripple effects through China’s component ecosystem

For Chinese component makers, the MLCC crunch creates both opportunity and pain. Domestic MLCC suppliers like Fenghua Advanced Technology and Three-Circle Group have seen orders increase as customers seek alternatives to constrained Japanese and Korean suppliers. But these Chinese manufacturers primarily serve the mid-range and low-end market — the high-capacity, high-reliability MLCCs needed for AI servers remain dominated by Murata, Samsung, and TDK.

The squeeze extends to microcontrollers. As AI server production absorbs more wafer capacity at foundries like TSMC and SMIC, the allocation for commodity chips — 8-bit and 32-bit MCUs used in everything from washing machines to factory automation — tightens. Chinese MCU makers including GigaDevice, GD32, and Puya Semiconductor are facing rising wafer costs and longer lead times.

DIGITIMES reported that Chinese MCU makers and passive component suppliers are facing “rising costs, tighter capacity, and surging demand from AI servers” — a triple squeeze that compresses margins for everyone except the top-tier suppliers who can pass costs through.

The Japan-Korea-China triangle

The MLCC market is dominated by four Japanese and Korean companies: Murata (Japan, ~40% global share), Samsung Electro-Mechanics (South Korea, ~20%), TDK (Japan, ~15%), and Taiyo Yuden (Japan, ~8%). Together they control over 80% of the global MLCC market by value.

Chinese MLCC manufacturers hold roughly 5-7% of the global market, concentrated in low-capacitance, consumer-grade components. The AI server boom is widening this gap because the high-capacity, high-temperature MLCCs required for GPU power delivery are precisely the segments where Chinese suppliers are weakest.

This dynamic mirrors the broader semiconductor equipment story: China’s semiconductor equipment self-reliance has surged past 35% for mature nodes, but advanced components remain dependent on foreign suppliers. The same pattern is playing out in passive components — Chinese companies are gaining share in commodity segments but struggling to break into the high-performance tier that AI infrastructure demands.

Capacity reallocation: who loses?

When Murata or Samsung shifts production capacity toward high-end MLCCs for AI servers, somebody loses. The most vulnerable segments are:

• Automotive: Cars use thousands of MLCCs each, and automotive-grade components require expensive qualification. Shifting capacity away from auto-grade MLCCs creates shortages that take 6-12 months to resolve.
• Industrial automation: PLCs, motor drives, and sensor modules all depend on mid-range MLCCs that compete for the same production lines.
• Consumer electronics: Smartphones, laptops, and IoT devices use commodity MLCCs, but these customers have less pricing power than AI server makers.

The result is a two-tier market: AI server OEMs with deep pockets and long-term supply agreements get priority, while everyone else faces longer lead times and higher prices. Chinese manufacturers in the automotive and industrial sectors are particularly exposed because they lack the purchasing leverage of Nvidia’s server partners.

What to watch

The MLCC super-cycle that industry observers have been predicting since late 2025 appears to be materializing. Key indicators:

• Murata’s pricing decisions: If the world’s largest MLCC maker raises prices in Q3 2026, others will follow within weeks
• SMIC capacity allocation: China’s largest foundry is running at high utilization; any shift toward AI-related wafer starts reduces MCU and analog chip output
• Chinese MLCC M&A: Domestic suppliers may pursue acquisitions of Japanese or Korean MLCC technology to close the gap
• AI server shipment data: Nvidia’s Q2 earnings (expected August 2026) will reveal whether the AI server buildout is accelerating or plateauing

CII Analysis

The AI server boom is not just a semiconductor story — it is reshaping the entire electronics component supply chain. China, as both the world’s largest electronics manufacturer and a rapidly scaling AI infrastructure market, sits at the center of this disruption.

Chinese component makers face a strategic choice: compete in the commodity segments where margins are thin and capacity is being reallocated toward AI, or invest aggressively in the high-performance components where Japanese and Korean suppliers still dominate. The MLCC market is a microcosm of the broader technology self-sufficiency challenge — China can make most of what it needs, but the most critical, highest-value components still come from abroad.

For investors, the pricing cycle in passive components creates opportunities in Japanese MLCC makers (Murata, TDK, Taiyo Yuden) and risks for Chinese downstream manufacturers who face rising input costs without the ability to pass them through. The AI infrastructure buildout is a rising tide that lifts all boats in the component supply chain — but it lifts some boats much higher than others.

Follow CII on LinkedIn for daily analysis of China’s semiconductor supply chain.

Sources

1. DIGITIMES — China chip supply chain feels squeeze from AI server MLCCs to 8-bit MCUs (June 19, 2026)
2. DIGITIMES — Taiyo Yuden to boost AI server MLCC capacity (June 17, 2026)
3. DIGITIMES — AI servers squeeze high-end MLCC supply (June 11, 2026)
4. DIGITIMES — AI sparks tight supply warning for high-end passive components (June 4, 2026)
5. DIGITIMES — Taiyo Yuden raises MLCC prices, Murata takes the lead (April 17, 2026)