Samsung Electro-Mechanics Wins $293M MLCC Contract
Samsung Electro-Mechanics signed a 454 billion-won (US$293 million) one-year contract on June 30, 2026 to supply multilayer ceramic capacitors (MLCCs) for AI servers to an undisclosed U.S. cloud service provider, running from January 2027 through December 2027, according to Yonhap and Korea Times. The deal matters because AI servers pack more than 10 times as many MLCCs as conventional servers, an AI GPU alone carries over 20,000 units, and a full rack can hold up to 600,000, so component supply has become a real bottleneck in datacenter buildouts. According to Korea Times, the contract equals roughly 9 percent of the component division's 2025 revenue and is unusually long-term for an industry that typically runs on quarterly or spot orders, reflecting tight AI-driven MLCC supply. Samsung Electro-Mechanics holds over 40 percent of the AI-server MLCC market, trailing only Japan's Murata Manufacturing overall; shares reportedly rose on the news.
For AI infrastructure and datacenter procurement teams, this deal is a concrete data point on how tight the passive-component supply chain has become as AI server buildouts scale: long-term, fixed-volume MLCC contracts are replacing the industry's traditional spot and quarterly ordering, signaling suppliers and hyperscalers are both moving to lock in capacity years ahead of deployment.
What happened
According to Yonhap and Korea Times, Samsung Electro-Mechanics disclosed in a regulatory filing that it signed a 454 billion-won (US$293-294 million) contract to supply multilayer ceramic capacitors (MLCCs) for AI servers to a U.S.-based Big Tech customer. The Korea Times reports the deal runs for one year, from January 1, 2027 through December 2027. The customer was not named due to a confidentiality agreement; Yonhap reports industry observers believe the client is likely one of the major U.S. cloud service providers, such as AWS, Google, or Meta. Samsung Electro-Mechanics CEO Chang Duck-hyun said, according to Korea Times, "The contract demonstrates that Samsung Electro-Mechanics' MLCC technology has been recognized as a key component in the AI era."
Technical context
MLCCs stabilize power delivery for semiconductors and other electronic components; power fluctuations in servers can directly degrade performance. Korea Times reports that AI servers require MLCCs able to withstand higher heat and greater warpage stress than consumer-grade parts, and contain more than 10 times as many MLCCs as conventional servers: a single AI GPU can carry over 20,000 MLCCs, and a full AI server rack can hold up to 600,000. This scale of component demand is a genuine and verifiable driver tying MLCC supply directly to AI datacenter hardware buildout, not just general electronics demand.
Industry context
Korea Times reports the deal equals roughly 9 percent of the component division's 2025 revenue of 5.2 trillion won, and that such multi-hundred-million-dollar, single-year contracts are uncommon in an MLCC market that has traditionally run on quarterly or spot orders. Samsung Electro-Mechanics holds close to 25 percent of the overall global MLCC market behind Japan's Murata Manufacturing, but leads the AI-server MLCC segment specifically with more than 40 percent share, according to Korea Times. TrendForce separately reports over 18 percent annual growth in the adjacent silicon-capacitor market, consistent with sustained AI-driven demand for high-end capacitors.
For practitioners
Infrastructure and procurement teams sourcing or specifying AI server hardware should treat MLCC availability as a real scheduling and cost variable, not a rounding error: with per-rack demand in the hundreds of thousands of units, supplier capacity commitments and lead times can influence datacenter rollout timing. The shift toward long-term supply agreements suggests hyperscalers are actively hedging against component shortages rather than relying on spot-market availability.
What to watch
Watch for further long-term MLCC or passive-component supply agreements from major cloud providers, any capacity-expansion announcements from Samsung Electro-Mechanics or Murata, and whether pricing for server-grade MLCCs continues to rise as AI datacenter investment accelerates. Korea Times reports Samsung Electro-Mechanics is already in discussions with other Big Tech customers about supply expansion beyond 2027.
Key Points
- 1Samsung Electro-Mechanics signed a $293-294M, one-year MLCC supply deal for AI servers with an undisclosed U.S. cloud provider starting January 2027.
- 2AI servers require over 10 times more MLCCs than conventional servers, up to 600,000 per rack, making passive-component supply a real datacenter bottleneck.
- 3The long-term contract structure, unusual for an industry that runs on spot orders, signals hyperscalers locking in scarce AI-server component capacity early.
Scoring Rationale
A $293M single-supplier MLCC contract for AI servers is a genuine, well-evidenced AI-infrastructure component supply signal, not generic electronics news: Korea Times documents that AI server racks use up to 600,000 MLCCs (10x conventional servers), confirming the AI-hardware linkage is real rather than tenuous. However, this remains a single component supplier's contract-win story with modest broader significance for AI practitioners, warranting a below-mid score rather than the prior industry-shaking framing; two independent outlets (Yonhap, Korea Times) corroborate the core facts.
Sources
Primary source and supporting public references used for this report.
Practice interview problems based on real data
1,625 SQL & Python problems across 15 industry datasets — the exact type of data you work with.
Try 250 free problems
