AnswersPublished23 sources
Guangdong’s MLCC Trio Rides the AI Compute Wave
AI servers require roughly 15,000–25,000 MLCCs, far above conventional server demand, helping lift volumes and pricing for Guangdong’s Three Circle and Fenghua in H1 2026; Viiyong’s opportunity is visible in its high... Three Circle reported H1 revenue of RMB 6.423 billion and net profit of RMB 1.932 billion, while...
Published byEdited with GPT-5.6 TerraImages generated with GPT Image 2
Research answer
Key takeaways
- AI servers require roughly 15,000–25,000 MLCCs, far above conventional server demand, helping lift volumes and pricing for Guangdong’s Three Circle and Fenghua in H1 2026; Viiyong’s opportunity is visible in its high...
- Three Circle reported H1 revenue of RMB 6.423 billion and net profit of RMB 1.932 billion, while Fenghua reported RMB 3.50 billion in revenue and RMB 290 million in net profit.
- The long term test is qualification in premium server, power and optical applications: Chinese suppliers are gaining ground, yet Japanese and Korean companies still lead the highest end AI server MLCC market.
AI Prompt
openai.comCreate a landscape editorial hero image for this Studio Global article: How have Chaozhou Three-circle (Group), Fenghua and Viiyong—the “three musketeers” of Guangdong’s MLCC industry—benefited from the AI-comput. Article summary: AI infrastructure has turned MLCCs from a low-profile commodity into a power-delivery bottleneck: a mainstream AI server reportedly uses roughly 30,000 capacitors, far more than conventional servers. That is lifting volu. Topic tags: general, general web, user generated. Style: premium digital editorial illustration, source-backed research mood, clean composition, high detail, modern web publication hero. Use reference image context only for broad subject, composition, and topical grounding; do not copy the exact image. Avoid: logos, brand marks, copyrighted characters, real person likenesses, fake screenshots, UI text, readable text, watermarks, charts with fa
MLCCs—multilayer ceramic capacitors—are tiny components with an outsized role in AI infrastructure. They help stabilize and decouple power around processors, memory and power-delivery systems. As AI systems require denser, faster-changing power delivery, MLCC content rises sharply: industry reporting puts an AI server at roughly 15,000–25,000 MLCCs. 53
That shift has put Guangdong’s three prominent MLCC makers—Chaozhou Three-Circle (Group), Fenghua Advanced Technology and Viiyong—in focus. Their H1 2026 results and expansion plans show a real improvement in demand and product mix, although not every gain can be attributed to AI alone.
Why AI servers matter to MLCC makers
AI hardware increases both the number of capacitors required and the value of the specifications that matter. High-capacitance, high-voltage, high-frequency, low-impedance and high-reliability MLCCs are needed in server boards, power systems and related optical-communication equipment. AI-server demand is therefore more valuable than broad consumer-electronics demand for suppliers that can pass demanding qualification requirements. 17
53
The 2026 upcycle has also included improving industry conditions and price recovery in some MLCC specifications. That distinction matters: earnings are being supported by a mix of AI-related demand, automotive electronics, industrial controls, optical communications and a broader recovery in component pricing. 1
2
Three-Circle: scale in MLCCs plus optical-communications exposure
Chaozhou Three-Circle traces its origins to a local radio-component factory established in 1970. It has since built a broader advanced-electronic-ceramics business spanning electronic and ceramic materials, electronic components, communication devices and equipment components. 20
18
For the AI buildout, its advantage is not solely MLCC volume. The company also has exposure to optical-communications ceramic components, linking it to two infrastructure needs at once: power delivery in servers and high-speed optical interconnects between them. 17
24
The financial result was substantial. In the first half of 2026, Three-Circle reported revenue of RMB 6.423 billion, up 54.82% year on year, and net profit attributable to shareholders of RMB 1.932 billion, up 56.18%. Its adjusted net profit rose 69.66%. The company cited higher demand in electronic components and optical communications; for MLCCs, customer recognition, stronger industry conditions, price recovery for some specifications and higher sales volume all contributed. 1
2
Three-Circle also completed its Hong Kong listing on July 9, 2026, becoming an A+H-listed company. The offering generated net proceeds of about HK$7.046 billion, according to reporting on the listing. 20 That additional capital broadens its options for technology investment, automation and overseas expansion, but the value of the listing will ultimately depend on execution in high-end products rather than financing alone.
Fenghua: an early domestic pioneer moves up the specification ladder
Zhaoqing-based Fenghua Advanced Technology was an early Chinese MLCC producer, introducing a full production line from the United States in 1985. 17 Its relevance to the current cycle lies in its move beyond standard consumer-grade capacitors toward higher-capacitance, higher-voltage, automotive, industrial and high-reliability products. Fenghua has said it is advancing MLCCs, resistors and inductors toward higher capacitance, voltage, reliability and precision for markets including AI compute and automotive electronics.
2
In H1 2026, Fenghua reported RMB 3.50 billion in revenue, up 26.28% year on year, and RMB 290 million in attributable net profit, up 74.01%. Its second-quarter net profit was RMB 202 million, approximately 127% higher than in the first quarter. 1
9
The company’s disclosures also temper the pure-AI narrative. High-end MLCC products accounted for about 40% of MLCC revenue in the first half, while MLCC revenue itself accounted for about 40% of total company revenue. 10 In other words, Fenghua is gaining from AI-related demand and product upgrades, but its results still reflect a diversified passive-components business and growth in automotive, industrial and other end markets.
Viiyong: the younger high-end specialist
Viiyong, formally Guangdong Viiyong Electronic Technology, was founded in 2017 and is headquartered in Luoding, Yunfu, Guangdong. It specializes in high-end MLCCs. 32
Its strategic position differs from the two listed incumbents. Viiyong’s IPO materials emphasize miniature, high-capacitance and high-reliability products across consumer, industrial and automotive markets. Based on 2024 sales volume, its prospectus cites Frost & Sullivan data ranking it first in China and second globally in ultra-miniature MLCCs, behind Murata. 33
For AI-server applications, the relevant capability is high-capacitance production using high-precision stacking and lamination. Reporting on its IPO effort says Viiyong developed a process exceeding 1,200 layers and had begun batch supply of relevant models to leading server manufacturers. 21 The Shenzhen Stock Exchange accepted its ChiNext IPO application on April 28, 2026; the proposed RMB 1.675 billion raise was weighted heavily toward high-end MLCC capacity expansion.
21
Unlike Three-Circle and Fenghua, Viiyong did not have a comparable publicly reported H1 2026 earnings release in the provided materials. Its AI exposure is therefore better assessed through product capability, customer qualification and planned capacity than through half-year profit figures.
A Guangdong cluster with three distinct strengths
The three companies are based in different parts of Guangdong: Three-Circle in Chaozhou, Fenghua in Zhaoqing and Viiyong in Luoding, Yunfu. Industry reporting describes them collectively as Guangdong’s MLCC “three musketeers” and part of China’s largest, most complete electronic-components cluster. 17
Their differences are commercially useful:
- Three-Circle combines electronic-ceramics depth, MLCC scale and optical-communications components.
- Fenghua brings a long-established, broad passive-component platform and a push into higher-end MLCC specifications.
- Viiyong is a newer specialist concentrated on miniaturization and high-end capacity expansion.
Together, they illustrate why AI infrastructure can pull through demand far beyond GPUs and servers: materials expertise, precision manufacturing, component qualification and optical interconnects all become relevant upstream capabilities.
The key caveat: demand is not the same as leadership
The 2026 numbers show that higher MLCC volumes and better pricing are reaching company earnings. But a durable AI advantage requires more than participation in a cyclical shortage. Suppliers must reliably meet the stringent performance and qualification requirements of server, power and optical applications.
TrendForce reported that Chinese MLCC suppliers are gaining approvals and making progress in lower- and mid-range markets, while Japanese and Korean suppliers remain dominant in the highest-end AI-server segment. 34 That makes the investment and industrial question straightforward: the winners will be the companies that turn present demand into repeatable high-end design wins, not simply those benefiting from an industry-wide price recovery.
For now, Guangdong’s MLCC trio offers three versions of the same AI-infrastructure story: Three-Circle has delivered the strongest reported earnings and added Hong Kong capital-market access; Fenghua has translated product upgrades into faster profit growth; and Viiyong is pursuing the high-end specialist route through technology and capacity expansion. The AI compute wave has raised the stakes for all three, but their ability to compete at the top end remains the decisive next chapter.
Studio Global AI
Continue your research
This page includes a source-backed answer you can continue inside Studio Global.
People also ask
What is the short answer to "Guangdong’s MLCC Trio Rides the AI Compute Wave"?
AI servers require roughly 15,000–25,000 MLCCs, far above conventional server demand, helping lift volumes and pricing for Guangdong’s Three Circle and Fenghua in H1 2026; Viiyong’s opportunity is visible in its high...
What are the key points to validate first?
AI servers require roughly 15,000–25,000 MLCCs, far above conventional server demand, helping lift volumes and pricing for Guangdong’s Three Circle and Fenghua in H1 2026; Viiyong’s opportunity is visible in its high... Three Circle reported H1 revenue of RMB 6.423 billion and net profit of RMB 1.932 billion, while Fenghua reported RMB 3.50 billion in revenue and RMB 290 million in net profit.
What should I do next in practice?
The long term test is qualification in premium server, power and optical applications: Chinese suppliers are gaining ground, yet Japanese and Korean companies still lead the highest end AI server MLCC market.