被忽視的 2026 年供應風險被動元件為何開始緊缺

☘️ Article

癌大觀點

• good read
• 供需失衡始於領導廠商走向高價值，並退出低毛利業務，若剛好遇到需求大增就出狀況
• 時間不好抓，類似記憶體在 2023 AI 出來之際就不難想到會爆發，但直到 sora 那個時間點附近才催化出來
• 18-19 的 mlcc 缺貨是村田、tdk、太陽誘電退出 legacy 和通用品，並有智慧型手機、電車等用量陡增造成失衡
• 市場都會有一個把大廠產能吸掉的主軸，這次是 AI 伺服器。目前參照 ems 的料件評估，開始有缺的是 tan cap 和 sp cap，mlcc 還在警戒線外有待觀察
• 原文是說 2026 會大缺，但現在看應該是還要煮一陣子
• ---
• " What Can the 2018–2019 Passive Shortage Tell Us About 2026?
• The current passive market is not without precedent. The last major disruption occurred in 2018 and 2019, when multilayer ceramic capacitors (MLCCs) became one of the most constrained component categories in the supply chain.
• That shortage was driven by two forces acting at the same time.
• First, demand surged across multiple fast-growing end markets. Smartphones began requiring well over 1,000 MLCCs per device, while electric vehicles and advanced automotive electronics drove thousands more per platform. At the same time, the expansion of connected devices and early IoT adoption added further pressure to already tightening supply.
• Second, supply was structurally constrained. Between 2016 and 2019, major manufacturers such as Murata, TDK, and Taiyo Yuden reduced or exited production of legacy and general-purpose MLCC lines in favor of higher-margin, high-capacitance products designed for premium automotive and advanced handset applications. Capacity expansion was limited, in part because adding MLCC production lines requires complex layering processes and long qualification cycles, often taking eight to twelve months or longer. Investment flowed toward the most profitable segments rather than commodity components.
• The result was a prolonged imbalance where demand accelerated faster than supply could realistically respond.
• Why Does This Matter Heading Into 2026?
• Today’s passive market is not experiencing the same broad-based shortage conditions seen in 2018 or during the 2021 through 2023 supply cycle. However, the underlying pattern is beginning to re-emerge.
• Similar to 2018, demand for passives is increasing sharply as 2026 server and GPU platforms require a higher concentration of advanced passive components. This growth is being driven by high-density compute platforms, data center infrastructure, and next-generation server architectures that demand greater power regulation, stability, and signal integrity. At the same time, material costs have risen, and capacity has remained relatively flat or has been reduced following the oversupply conditions of 2024, when manufacturers focused on cost controls rather than expansion. The result is a tighter starting point for 2026, with less buffer in the system and fewer opportunities to absorb demand shifts once capacity is spoken for.
• Why Is Demand for Passives Rising So Quickly?
• One of the most important forces shaping passive availability in 2026 is not how many systems are being built, but how many more passive components each system now requires.
• Modern server, networking, and high-density compute platforms consume substantially more capacitors and connectors to support power delivery, voltage regulation, and signal integrity. As architectures become more power-dense, passive content per system continues to rise, even if overall unit volumes remain relatively stable.
• This dynamic can be easy to miss. While attention is often focused on processors and memory, the passive footprint of each system has grown quietly but meaningfully.
• Why Aren’t Manufacturers Expanding Passive Capacity Faster?
• Passive components are built and scaled very differently than advanced processors and memory, and that difference directly affects how capacity is added.
• Leading-edge semiconductors such as CPUs, GPUs, and high-performance memory are produced in extremely high volumes around a relatively small number of standardized designs. That concentration supports large, fast investments in new capacity because demand is deep, predictable, and consolidated.
• Passive components, by contrast, are produced across thousands of variations that differ by size, voltage, material, tolerance, and qualification requirements. Even when overall demand increases, that demand is spread across many individual part numbers rather than concentrated into a small number of high-volume SKUs. Expanding capacity for one specification does not automatically relieve constraints elsewhere.
• As a result, capacity expansion tends to be incremental and selective. Suppliers prioritize committed, long-term programs where volumes are locked in early and production can be planned efficiently, leaving less flexibility for spot demand or late-stage adjustments.
• How Are Material Costs Affecting Passive Pricing?
• Heading into 2026, supply risk is increasingly being shaped by economics as much as availability.
• Throughout 2025, manufacturers have faced sustained increases in copper and silver, two materials that sit at the core of many passive components. Copper is widely used in connector contacts, lead frames, and internal capacitor structures, while silver is commonly used in electrode pastes and conductive layers within ceramic and tantalum capacitors. As those input costs rise, they directly increase the cost to produce finished components.
• At the same time, upstream materials used in printed circuit boards (PCBs) and related assemblies, including copper foil and laminate materials, have also seen cost inflation. Together, these pressures are influencing both pricing strategies and supplier prioritization decisions.
• Which Passive Categories Are Most at Risk in Early 2026?
• While exposure varies by specification and use case, several passive categories are consistently flagged as higher risk heading into Q1 and Q2 2026:
• Tantalum capacitors, where lead times have extended significantly and pricing pressure continues to build
• Polymer capacitors, which are seeing early shortages and announced price increases
• Aluminum electrolytic capacitors, where lead times have stretched beyond historical norms
• High-capacitance MLCCs, particularly those used in server and data center platforms
• Connectors, where rising copper and gold costs are driving price increases
• Risk is not uniform across all parts, but availability is becoming increasingly dependent on specification, qualification, and timing.
• https://www.fusionww.com/insights/why-passive-components-are-emerging-as-a-2026-supply-risk

✍️ Abstract

被忽視的 2026 年供應風險：被動元件為何開始緊缺？

• 擴產困境：被動元件規格繁雜，涵蓋數千種 尺寸、電壓、材質 組合，擴張速度較 處理器、記憶體 緩慢。
• 供應僵化：廠商傾向優先支應長約客戶，導致現貨市場面對突發需求時缺乏調整彈性。
• 需求激增：AI 伺服器、GPU 平台 嚴格要求 電量調節、電壓穩定、訊號完整性，推升單機元件用量。
• 風險重現：2024 年擴產停滯，隨著 2026 年新一代高密度架構推出，供應壓力將再次浮現。
• 歷史借鏡：2018~2019 年缺貨源於需求激增，如：智慧型手機、電動車 銷量增加，且大廠退出低毛利業務。
• 當前徵兆：AI 伺服器、資料中心、新世代架構 對 電源調節、信號完整性 要求攀升，推升元件密度。
• 成本推升：銅、銀、金 等關鍵原物料價格持續上漲，直接增加 連接器、電極漿料、電路板 的生產成本。
• 高風險類別：鉭質電容、聚合物電容、鋁電解電容、高容量 MLCC 為 2026 年初供應風險最高的元件。

癌大觀點分析

• 分析價值：原文深入解析被動元件供需失衡的深層背景。
• 失衡開端：領導廠商轉向高毛利產品、退出低階業務，若遇突發需求大增即產生缺口。
• 缺貨發酵期：時間點難以精準預估，如：2023 年 AI 記憶體缺口，直到 Sora 推出才被真正催化。
• 歷史主軸：2018~2019 年 MLCC 缺貨主因是 村田、TDK、太陽誘電 等大廠退出 舊有 legacy、通用品市場，並轉向高需求領域。
• 產能吸盤：市場總會有吸走大廠產能的主軸，當前關鍵即為 AI 伺服器。
• 缺貨跡象：依 EMS 廠評估，Tan Cap (鉭電容)、SP-Cap (聚合物電容) 漸顯缺口，MLCC 仍在警戒線外。
• 趨勢發展：2026 年雖估有大缺貨，預期目前仍需一段時間 醞釀、發展。

專有名詞

• 積層陶瓷電容器 (MLCC)：由交替堆疊的陶瓷介電層和金屬電極層組成的電容器，是電子設備中使用量最廣泛的被動元件，用於濾波、去耦和儲能。
• 鉭質電容器 (Tantalum Capacitor)：以金屬鉭作為陽極的電解電容，具備體積小、電容量大且在高溫下極其穩定的特性，常用於醫療、軍工及高階伺服器。
• 聚合物電容器 (Polymer Capacitor)：使用導電聚合物作為電解質的電容，具有低等效串聯電阻 (ESR) 和長壽命優點，適合處理高速數位電路中的高頻雜訊。
• SP Cap (鋁聚合物電容器)：特定類型的固態鋁電解電容，因其超低 ESR 特性，常被用於顯示卡與伺服器處理器的電壓調節模組。
• EMS (電子製造服務)：專門為電子品牌商提供設計、製造、測試及物流服務的外包廠商。
• 訊號完整性 (Signal Integrity)：信號完整性，是指電信號在通過電路傳輸後，仍能保持其原始特性 (如：電壓、時序) 的能力，這對高速運算系統至關重要。
• Legacy (遺留產品/舊款產品)：指技術較為 成熟、毛利較低、競爭激烈、逐漸被廠商減少產能投入 的舊規格元件。
• SKU (庫存單位)：用於追蹤庫存的特定產品編碼，每一種規格差異都會被視為一個獨立的 SKU。