Shandong Sinocera - A Strong valuation re-rating; AI-related core businesses; Potential upside of multiple growth pillars; Stay OW

Research evidence excerpt

Shandong Sinocera - A Strong valuation re-rating; AI-related core businesses; Potential upside of multiple growth pillars; Stay OW

Lei Mu Asia Pacific Equity Research

(86-21) 6106 6319 27 May 2026

lei.mu@jpmorgan.com

Positive driver #1: Strong demand outlook for high-end

MLCC powder

Downstream demand contributor #1 – AI servers

MLCC demand has historically been closely tied to consumer electronics. As smart

hardware evolved continuously, manufacturers inserted more passive components to

support multi-functional features such as 5G, wireless charging, full-screen displays,

and multi-camera systems. Additional MLCC demand is now being shifted to AI

servers, which are emerging as the single most powerful growth engine for high-end

MLCC powder. The exponential rise in AI computing power has created

unprecedented demand for MLCC, dwarfing even the growth seen in smartphones.

Figure 1: MLCC unit demand per device (unit: thousand)

Source: J.P. Morgan estimates

The scale of the demand surge lies in the sheer volume of MLCC required by AI

servers compared to traditional consumer electronic devices. The latest smartphones

require around 1,500 MLCC units per device, while traditional servers need 1,500-

2,000 MLCC units per device.

For AI servers, due to the unique working conditions, there are three key

requirements: 1) Extreme power consumption: high-performance GPUs and

accelerators demand highly stable power delivery, therefore, the device requires far

more MLCCs for filtering and voltage regulation; 2) Complex power distribution

architecture: AI servers use advanced vertical power delivery systems, multiplying

the number of MLCC needed across the board; and 3) Reliability requirements: 24/7

operation data centers demand the highest quality, high-temperature, high-