AI Server Demand Drives Up MLCC Channel Prices, Some Part Numbers Surge Over 20x
Miles Bennett
Surging AI-server demand has pushed high-capacitance MLCC channel prices up ~5% in a single day at Huaqiangbei, with one Murata part number trading at 25 yuan versus a 1 yuan factory price — a 20× markup driven not by official price hikes but by channel speculation.
What is an MLCC, and why do AI servers need so many?
An MLCC — multilayer ceramic capacitor, a fingernail-sized passive component that stabilizes power delivery to chips — is one of the most heavily used parts in every server.
AI servers draw extreme GPU power and require larger-capacitance MLCCs to keep voltage stable. These "high-cap" part numbers have low manufacturing yield and consume disproportionate production capacity, making them far scarcer than standard grades.
This means → not all capacitors are in shortage. Only part numbers used in AI servers and wireless charging are genuinely tight.
How far have channel prices moved?
Murata 1206-476 (47 μF, a server-standard part) closed at 725–755 yuan per 2,000 units at Huaqiangbei on July 5 — up ~5% from the morning quote.
GPU ultra-high-cap parts had no volume stock; loose units carried a 15%+ premium, stalls rationed supply, and quotes changed by the hour.
The most extreme case: Murata's 107-series part has a factory price of ~1 yuan but trades at ~25 yuan in the channel — a 20× markup.
In plain terms = a part the manufacturer sells for 1 yuan becomes 25 yuan by the time it reaches the channel. That is not normal supply-demand pricing — it is speculative hoarding.
The factories haven't raised prices — so who is?
A manufacturer source told Shanghai Securities News that Japanese OEMs have not publicly announced price increases. Sustained channel speculation is the main driver of abnormal price spikes on certain part numbers.
Server-grade standard MLCCs can currently be sourced from essentially only two OEMs: Murata and Samsung Electro-Mechanics. Supply is highly concentrated.
This means → channel traders are betting that OEM capacity won't free up soon, and they are hoarding within that window. Once capacity catches up, speculative prices could drop sharply.
The bigger picture: passives have been repricing for months
From H2 2025 through February 2026, global passive components went through multiple rounds of broad-based price hikes, with increases typically in the 5%–30% range.
A new round began in April 2026. OEMs prioritized AI and automotive-grade customers; channel prices moved to "a new quote every day."
On July 1, passive-component giant Yageo notified customers of price increases across its full capacitor lineup — tantalum, MLCC, aluminum, solid-state aluminum, film, and supercapacitors — covering roughly 50% of Yageo's revenue and, for the first time, extending to direct customers (EMS/OEMs).
This reflects a shift: price increases have spread from channel middlemen to factory level, and from a single category to the full product line — this is no longer just speculation but an industry-wide repricing.
Who benefits from this cycle?
Shenzhen Huaqiang (深圳华强) disclosed on July 6 that AI infrastructure demand and a capacity siphoning effect are lifting prices across memory chips, power devices, and passive components. The company expects to benefit from both higher volumes and higher per-unit margins in its distribution business.
Sunlord Electronics (商络电子) said its full-category "active + passive" component portfolio positions it to capture both passive-component demand expansion and incremental active-component orders.
In plain terms = distributors earn on both sides of a price-up cycle — they sell more, and each unit is more profitable. How long this logic holds depends on the next question.
Can this rally last?
OEM sources stress this is a structural rally: many MLCC part numbers are not in short supply at all — only specific application segments are tight.
Murata holds over 40% of global MLCC market share and roughly 70% of the global AI-server MLCC market. That concentration means the pace of capacity release is largely one company's decision.
This means → the key variable is whether channel speculation cools before OEM capacity is released. If capacity expansion comes faster than expected, part numbers currently trading at a 20× premium could see a sharp correction.
Content is for reference only, not financial advice.