![]() |
市場調查報告書
商品編碼
2069655
共封裝光元件市場:按組件、資料速率、整合類型和最終用戶分類-市場規模、產業動態、機會分析和預測(2026-2035 年)Co-Packaged Optics Market: By Component, Data Rate, Integration Type, End User - Market Size, Industry Dynamics, Opportunity Analysis and Forecast For 2026-2035 |
||||||
全球共封裝光元件 (CPO) 市場正經歷快速且變革性的成長,反映出對現代資料中心和人工智慧 (AI) 基礎設施日益成長的需求。預計到 2025 年,該市場規模將達到約 1.7087 億美元,而到 2035 年,預計將大幅成長至約 7.8087 億美元。這一強勁的成長趨勢體現在 2026 年至 2035 年預測期內高達 35.9% 的複合年成長率 (CAGR) 上,凸顯了該技術在下一代運算生態系統中日益重要的戰略意義。
這一顯著成長主要源自於處理頻寬密集型人工智慧工作負載日益成長的需求,這些工作負載正在重塑全球數位基礎架構的架構。生成式人工智慧、大規模語言模型和高效能運算叢集的快速發展,對伺服器、加速器和儲存系統之間的資料傳輸提出了前所未有的要求。
共封裝光元件 (CPO) 市場目前由少數幾家極具影響力的技術領導企業主導,這些企業正推動著資料中心和人工智慧基礎設施領域的創新、商業化和大規模部署。這些公司正透過先進的矽光電平台、高頻寬交換架構以及光技術與電子處理系統的深度整合,為下一代光連接模組技術制定發展藍圖圖。
憑藉在人工智慧 (AI) 硬體領域的強大實力,NVIDIA 正透過推動共封裝光元件的整合,成為該領域最傑出的領導者之一。博通 (Broadcom) 在共封裝光元件市場也佔主導地位,尤其是在高效能開關晶片領域。 Ayar Labs 是共封裝光元件生態系統中的另一個關鍵創新者,尤其在用於 AI 擴展架構的光輸入/輸出 (I/O) 領域,其影響力不容小覷。
英特爾在共封裝光元件技術的早期發展中發揮了先鋒作用,尤其早在2020年就展示了CPO解決方案。 Marvell Technologies也是共封裝光元件領域的主要參與者,提供專為高速互連應用而設計的先進矽光電引擎。這五家公司共同塑造了共封裝光元件市場的競爭格局和技術演進。
主要成長要素
極高的電力消耗和對提高能源效率的迫切需求是推動全球資訊通訊技術 (ICT) 產業採用共封裝光元件 (CPO) 的核心驅動力。隨著數位基礎設施的快速擴展以支援人工智慧、雲端運算和高效能資料處理,現代運算生態系統的能源需求達到了前所未有的水平。 ICT 產業的年電力消耗量目前已達到約 1,000 太瓦時 (TWh),凸顯了數據驅動技術帶來的全球能源負擔之重。
新機會的趨勢
從2.5D整合到3D整合正成為共封裝光學元件(CPO)市場成長的關鍵趨勢,標誌著先進半導體和光子系統設計進入下一個發展階段。儘管2.5D架構目前憑藉其成熟性、可靠的性能以及與現有製造生態系統的兼容性而佔市場主導地位,但業界正日益轉向3D整合,以期提高性能密度、提升能效並進一步小型化下一代計算基礎設施。
最佳化障礙
製造和整合製程的複雜性是限制共封裝光學元件 (CPO) 市場成長的主要因素。這是因為製造過程需要極高的精度,並且光子元件和電子元件之間需要高度協調。與傳統的半導體組件不同,CPO 需要將光引擎和高性能電子開關及處理單元緊密整合,而且通常需要極其緊湊的封裝尺寸。這種技術融合帶來了巨大的技術挑戰,必須加以解決才能實現大規模商業部署。
The global Co-Packaged Optics (CPO) market is experiencing a phase of rapid and transformative expansion, reflecting the accelerating demands placed on modern data center and artificial intelligence infrastructure. In 2025, the market is valued at approximately USD 170.87 million, but it is projected to surge dramatically to around USD 780.87 million by 2035. This strong upward trajectory corresponds to a robust compound annual growth rate (CAGR) of 35.9% over the forecast period from 2026 to 2035, underscoring the technology's increasing strategic importance in next-generation computing ecosystems.
This remarkable growth is primarily being driven by the escalating need to support bandwidth-intensive AI workloads, which are reshaping the architecture of global digital infrastructure. The rapid expansion of generative AI, large language models, and high-performance computing clusters has created unprecedented data movement requirements between servers, accelerators, and storage systems.
The co-packaged optics (CPO) market is currently shaped by a small group of highly influential technology leaders who are driving innovation, commercialization, and large-scale deployment across data centers and AI infrastructure. These companies are defining the technological roadmap of next-generation optical interconnects through advanced silicon photonics platforms, high-bandwidth switching architectures, and deep integration of optics with electronic processing systems.
NVIDIA has emerged as one of the most prominent leaders in this space, leveraging its strong position in artificial intelligence hardware to advance co-packaged optics integration. Broadcom also plays a dominant role in the co-packaged optics market, particularly through its leadership in high-performance switching silicon. Ayar Labs represents another critical innovator in the CPO ecosystem, particularly in the domain of optical input/output (I/O) for AI scale-up architectures.
Intel has played a pioneering role in the early development of co-packaged optics technologies, particularly through its demonstration of CPO solutions as early as 2020. Marvell Technology is also a key player in the co-packaged optics landscape, offering advanced silicon photonics engines designed for high-speed interconnect applications. Collectively, these five companies are shaping the competitive dynamics and technological evolution of the co-packaged optics market.
Core Growth Drivers
Extreme power consumption and the urgent need for improved energy efficiency are becoming central forces driving the adoption of co-packaged optics (CPO) across the global information and communications technology (ICT) sector. As digital infrastructure expands rapidly to support artificial intelligence, cloud computing, and high-performance data processing, the underlying energy requirements of modern computing ecosystems have escalated to unprecedented levels. The ICT industry now consumes approximately 1,000 TWh of electricity annually, highlighting the scale of the global energy burden associated with data-driven technologies.
Emerging Opportunity Trends
The shift from 2.5D to 3D integration is emerging as a major opportunity trend for growth in the co-packaged optics (CPO) market, signaling the next phase of evolution in advanced semiconductor and photonic system design. While 2.5D architectures currently dominate the market due to their maturity, proven performance, and compatibility with existing manufacturing ecosystems, the industry is increasingly exploring 3D integration as a pathway toward higher performance density, improved energy efficiency, and further miniaturization of next-generation computing infrastructure.
Barriers to Optimization
Manufacturing and integration complexity represents a significant constraint on the growth of the co-packaged optics (CPO) market, as the underlying fabrication processes require an exceptionally high degree of precision and coordination between photonic and electronic components. Unlike traditional semiconductor assembly, CPO involves the close integration of optical engines with high-performance electronic switching and processing units, often within extremely compact form factors. This convergence of technologies introduces substantial engineering challenges that must be addressed before large-scale commercial deployment can be fully realized.
By data rate, the "Up to 800G" segment continues to hold a dominant position in the co-packaged optics (CPO) market in 2026, accounting for approximately 58% of total market share. This strong foothold reflects the industry's pragmatic approach to scaling network infrastructure in response to rapidly increasing bandwidth demands from modern data centers. As artificial intelligence workloads, cloud computing services, and high-performance computing applications continue to expand, operators are prioritizing data rate solutions that deliver substantial performance improvements while maintaining production stability and economic viability.
By integration type, 2.5D integration continues to dominate the co-packaged optics (CPO) market, accounting for approximately 52% of total market share in 2026. This sustained leadership reflects its position as the most commercially viable and technologically balanced approach for integrating optical and electronic components in advanced semiconductor systems. As demand for higher bandwidth, lower latency, and improved energy efficiency continues to accelerate, 2.5D integration has emerged as the preferred architecture for scaling next-generation optical interconnect solutions across data centers, artificial intelligence infrastructure, and high-performance computing environments.
By application, AI and machine learning networking has rapidly become the dominant force shaping the co-packaged optics (CPO) landscape, accounting for an overwhelming 65% market share in 2026. This dominance reflects the structural transformation occurring within modern computing infrastructure, where artificial intelligence workloads have shifted from experimental deployments to large-scale, mission-critical systems. The rise of generative AI, foundation models, and advanced machine learning applications has fundamentally redefined networking requirements, pushing traditional interconnect technologies beyond their practical performance limits.
By end user, hyperscale and cloud operators form the undisputed foundation of the co-packaged optics (CPO) market, accounting for a dominant 72% share of the global ecosystem. This overwhelming concentration reflects the fact that CPO technology is primarily designed to solve scaling challenges in extremely large, high-performance computing environments. These operators run some of the world's most complex and bandwidth-intensive infrastructure, where even marginal improvements in latency, power efficiency, and data throughput translate into substantial operational and financial benefits. As a result, hyperscale cloud providers have become the earliest and most aggressive adopters of CPO solutions.
By Component
By Data Rate
By Integration Type
By Application
By End User
By Region
Geography Breakdown