Goldman Sachs: AI-Driven MLCC Demand, Market Size to Grow 7.4x from FY25 to FY28
0xBroomberg
Goldman Sachs projects the AI-server MLCC market will expand 7.4x between FY25 and FY28, driven by two converging forces — high-end price resets and tightening commodity-grade supply — that point to a multi-year tight cycle across the entire MLCC industry.
What is MLCC, and why does AI suddenly need so much of it?
MLCCs — multi-layer ceramic capacitors, fingernail-sized components found on virtually every circuit board — are seeing explosive demand from AI servers.
Goldman's data: AI-server MLCCs already account for over 10% of the market in 2025, and are on track to become the single largest end-use segment within a few years.
This means → MLCC has shifted from "a supporting part in phones and cars" to "a must-have for AI infrastructure" — a qualitative change in demand drivers.
What is driving the price increases? How do the two forces stack up?
Force one: high-end "price reset." AI accelerators demand high-density, miniaturized capacitors, pushing high-end MLCC unit prices far above traditional IT levels.
Force two: commodity-grade supply squeeze. Makers are steering capacity toward the higher-margin AI segment, leaving less room for general-purpose MLCCs — and spot prices are rising.
In plain terms = high-end prices are climbing because of AI demand, and low-end prices are climbing because AI is absorbing the capacity. Both ends tighten at once, lifting the whole industry.
Goldman flags a caveat: headlines about "MLCC price hikes" can mislead — you must look at high-end, mid-range, and commodity grades separately.
Why can't capacity catch up quickly?
Most MLCC makers are already running near 90% utilization; Japanese firms typically expand at just 10%–15% per year.
Key equipment and materials are largely built in-house. This means → scaling up is not simply a matter of spending more money; there is a physical bottleneck.
If AI data-center demand doubles annually as expected, most new capacity will be absorbed by AI, keeping overall supply-demand tight for years.
How does this relate to the semiconductor cycle?
Goldman notes the MLCC cycle lags semiconductors by roughly three years and is still in its AI-driven early stage.
Yet surging per-server power consumption gives MLCC demand a sustained driver independent of the traditional electronics cycle.
This reflects something important: MLCC is not just riding the semiconductor tailwind — it now has its own growth engine.
Is now the time to buy? How does Goldman view valuation?
Goldman uses an estimated FY2028 P/E of ~30x as the target-price benchmark for some companies; recent pullbacks have brought valuations down to the 27–30x range.
Goldman's view: if fundamentals hold over the next several years, current prices may represent a good entry point.
The near-term risk is that the market may correct if valuations run up too fast. Put simply = the fundamentals are intact, but when the stock price gets ahead of itself, the market hits the brakes on its own.
The key variable to watch: whether high-end price-hike expectations are already fully priced in.
Content is for reference only, not financial advice.