실리콘 포토닉스 반도체 시장 규모, 점유율 및 성장 예측(2026-2032년)

The Silicon Photonics Semiconductor Market is estimated at USD 3.40 billion in 2026 and is projected to reach USD 12.45 billion by 2032, representing a CAGR of 24.1% during 2026-2032.

Silicon photonics is becoming an important semiconductor technology for moving large volumes of data between processors, switches, accelerators, and network infrastructure. The technology integrates optical functions on silicon-compatible wafers and can combine waveguides, modulators, photodetectors, multiplexers, couplers, and thermal-control elements within a photonic integrated circuit. This approach allows manufacturers to use semiconductor fabrication and packaging infrastructure to increase optical integration and production scale.

Datacom and telecommunications remain the largest commercial applications. Silicon-photonics PICs are used in 400G, 800G, and emerging 1.6T optical systems. However, the market is entering a broader phase of development as optical conversion moves closer to compute and networking silicon. Near-package optics, co-packaged optical engines, and processor and accelerator optical-I/O chiplets are becoming important growth areas. These architectures reduce the length of high-speed electrical connections and address increasing power and bandwidth requirements in AI infrastructure.

The market is also benefiting from expanding foundry capacity. Semiconductor manufacturers are moving silicon photonics from specialist production toward higher-volume manufacturing. This creates opportunities for fabless photonic companies to access qualified process platforms without building their own wafer fabs. At the same time, advanced photonic packaging, laser integration, wafer-level testing, and known-good-die qualification are becoming increasingly important to commercial deployment.

Market Drivers

AI Networking Is Moving Optical Conversion Closer to Compute

• The rapid expansion of AI computing is creating significant demand for higher bandwidth between accelerators, switches, memory, and processing systems. As electrical connections extend across boards and racks, power consumption and signal integrity become more challenging.
• Silicon photonics can convert electrical signals into optical signals closer to processors and network ASICs. This reduces the length of high-speed electrical SerDes connections and can improve bandwidth density and power efficiency.
• Optical-I/O chiplets and co-packaged optical engines are therefore becoming important areas of market expansion. Intel's Optical Compute Interconnect, GlobalFoundries' SCALE platform, and Marvell's 1.6T light engine demonstrate the movement toward processor-adjacent optical connectivity.

Foundry Capacity Is Moving Toward Commercial Production

• The availability of merchant foundry capacity is supporting the transition of silicon photonics into a scalable semiconductor market. Tower Semiconductor has contracted USD 1.3 billion of silicon-photonics revenue for 2027 and has reported higher customer wafer commitments for 2028.
• GlobalFoundries is also expanding its silicon-photonics and advanced-packaging capabilities. These developments allow fabless photonic semiconductor companies to use established wafer-processing infrastructure and reduce the need for captive manufacturing facilities.
• Greater foundry capacity can also improve customer portability and reduce qualification barriers. Process design kits, device libraries, wafer yields, testing, and packaging capabilities are becoming important competitive factors alongside optical device performance.

Higher-Speed Optical Generations Increase Photonic Content

• The transition from 400G and 800G toward 1.6T and future 3.2T optical systems is increasing the number of optical lanes and the required integration density. Silicon photonics becomes more valuable as systems require multiple modulators, detectors, wavelength channels, and routing functions in compact packages.
• Higher bandwidth also increases the importance of wavelength-division multiplexing and integrated passive photonic functions. These developments increase the semiconductor content of optical engines while supporting higher bandwidth without proportional increases in package size.

Applications Are Expanding Beyond Datacom

• Datacom remains the primary revenue source, but silicon photonics is expanding into LiDAR, optical compute interconnect, photonic quantum computing, sensing, and other photonic applications.
• LiDAR systems can use integrated photonic functions for frequency-modulated continuous-wave architectures. Photonic quantum computing creates another emerging application area. Tower Semiconductor's collaboration with Xanadu demonstrates the use of semiconductor manufacturing approaches for photonic quantum processors.
• These applications remain smaller than datacom, but they broaden the potential customer base and create additional development and foundry opportunities.

Market Restraints

Laser Integration and III-V Material Requirements

• Silicon is an effective platform for many passive and electro-optic functions but is not an efficient light emitter. Silicon-photonics systems therefore often require external lasers or the integration of III-V semiconductor materials.
• This increases manufacturing complexity and creates additional supply-chain and packaging requirements. Laser attachment, coupling efficiency, thermal management, and reliability must all be controlled at production scale.
• The need to integrate different semiconductor materials can therefore limit the simplicity of an otherwise silicon-compatible photonic platform.

Fiber Attachment and Testing Complexity

• Photonic devices require precise fiber attachment and optical alignment. These processes can be more complex than conventional electrical semiconductor packaging. Wafer-level and die-level optical testing also require specialized equipment and procedures.
• As the market moves toward higher volumes, manufacturers need to improve automated optical testing, fiber coupling, and known-good-die qualification. Yield losses at these stages can significantly affect the economics of photonic semiconductor production.

Serviceability of Deeply Integrated Optics

• Co-packaged optics place photonic components directly beside expensive switching or compute silicon. This can reduce electrical interconnect length and improve system performance, but it also creates serviceability challenges.
• If a photonic component fails within a deeply integrated package, replacing the optical device may be more difficult than replacing a conventional pluggable optical module. This creates a trade-off between integration density and maintainability.

Competition from InP and VCSEL Platforms

• Indium-phosphide electro-absorption-modulated lasers and VCSEL technologies remain competitive in applications where discrete optical performance, established manufacturing processes, or cost considerations are more important than maximum integration density.
• Silicon photonics must therefore demonstrate clear advantages in bandwidth density, power per bit, integration, manufacturing scale, and total system economics to expand beyond established optical technologies.

Technology and Segment Insights

Semiconductor Product

• Datacom and telecom silicon-photonics PICs represent the largest semiconductor revenue pool in 2026. These PICs are already manufactured at scale for optical transceivers and coherent optical systems. The segment captures the semiconductor-layer value of integrated modulators, detectors, and routing elements.
• Optical-I/O chiplets and co-packaged photonic engines are the fastest-growing product category. Their current base is smaller, but they add photonic semiconductor content directly beside networking and compute silicon. Their growth is closely connected to AI scale-up and scale-out infrastructure.
• Standalone modulators and photodetectors remain important for specialized optical systems. WDM and passive photonic ICs support dense optical lane aggregation and routing. Foundry wafer processing is another important value pool because every increase in fabless photonic production creates additional wafer-processing demand.
• Semiconductor-level photonic packaging is also gaining importance. PIC and electronic integrated circuit integration, known-good-die testing, fiber interfaces, and advanced packaging are becoming essential as photonics moves closer to compute.

Integration Architecture

• Pluggable-module PICs remain the established architecture because they fit existing optical-transceiver designs. However, near-package and on-board photonics are gaining relevance as electrical interconnect distances increase.
• Co-packaged optical engines provide tighter integration between optical and electronic components. Processor and accelerator optical-I/O chiplets represent the most advanced architecture in the market. These solutions directly connect processors, accelerators, and memory-related systems through optical interfaces.

Application

• Data center and AI networking represent the dominant application. Increasing accelerator bandwidth and network traffic are driving demand for higher-speed optical connectivity.
• Telecommunications remains an established market, particularly for high-capacity optical networks. Optical compute interconnect is emerging as a significant growth opportunity as photonic connectivity moves into processor and accelerator architectures.
• LiDAR and physical AI, photonic quantum computing, sensing, and other photonic applications provide additional opportunities. These applications are smaller today but can diversify market demand beyond conventional communications.

Manufacturing Model

• Merchant silicon-photonics foundries are becoming increasingly important because they allow fabless companies to scale without building captive wafer fabs. Vertically integrated semiconductor manufacturers maintain control over process technology, photonic design, packaging, and system integration.
• Fabless photonic semiconductor companies focus on device and architecture development while using external manufacturing. Foundry and advanced-packaging ecosystems combine wafer fabrication with specialized photonic assembly and testing.

Customer Type

• Networking semiconductor vendors represent an important customer category because photonic devices are increasingly integrated with switching and networking silicon. Optical component and module suppliers remain major users of silicon-photonics PICs.
• Hyperscale and AI platform companies are becoming increasingly influential because their data-center architectures require rapid increases in bandwidth and power efficiency. Quantum, LiDAR, and photonic-system developers create additional demand, while research and defense programs support specialized development.

Geography

• North America is the largest design and early-demand region. The United States has a strong concentration of AI networking vendors, cloud operators, optical-semiconductor companies, and merchant foundries. Intel, GlobalFoundries, Marvell, Broadcom, Cisco, Coherent, Lumentum, Lightmatter, and Ayar Labs contribute to the regional ecosystem.
• Asia Pacific remains critical for manufacturing and packaging. Taiwan, Japan, China, Singapore, Malaysia, and South Korea provide semiconductor manufacturing, advanced packaging, and photonic-device capabilities. Tower Semiconductor is expanding 300 mm silicon-photonics capacity in Japan, while regional advanced-packaging companies support the broader ecosystem.
• Europe contributes through imec, STMicroelectronics, Soitec, and research-oriented photonics programs. The region is also developing greater emphasis on sovereign photonic semiconductor manufacturing.

Competitive and Strategic Outlook

• The competitive landscape includes merchant silicon-photonics foundries, vertically integrated semiconductor companies, networking semiconductor suppliers, optical component manufacturers, and fabless photonic specialists.
• Tower Semiconductor and GlobalFoundries compete strongly at the foundry level. Their positioning depends on process design kits, qualified device libraries, wafer capacity, process performance, and customer portability. Intel combines a mature silicon-photonics manufacturing platform with integrated laser capabilities and optical-I/O development.
• TSMC participates through photonic and heterogeneous-integration programs. Marvell, Broadcom, and Cisco connect silicon photonics with networking semiconductor roadmaps. Coherent and Lumentum combine photonic semiconductor capabilities with laser and optical-module expertise.
• Lightmatter and Ayar Labs are focused on optical-I/O architectures that move photonic connectivity toward processors, memory, and AI accelerator systems. imec supports the market through process development, prototyping, and industrial collaboration.
• Competitive differentiation is increasingly based on modulator bandwidth, photodiode performance, power consumption per bit, wavelength density, laser strategy, wafer yield, packaging repeatability, and access to high-volume foundry capacity. As photonic devices become more deeply integrated, known-good-die testing and electronic-photonic co-design will become increasingly important.
• Recent market developments demonstrate the accelerating commercialization of the technology. Tower Semiconductor announced a major 300 mm silicon-photonics and SiGe capacity expansion in Japan in July 2026. In May 2026, the company disclosed USD 1.3 billion in contracted silicon-photonics revenue for 2027 and higher customer capacity commitments for 2028.
• GlobalFoundries launched its SCALE silicon-photonics co-packaged optical platform in May 2026 for AI scale-up infrastructure. Marvell highlighted its 1.6T silicon-photonics light engine for high-bandwidth AI applications in March 2026. Tower Semiconductor and Xanadu expanded their collaboration on production-oriented silicon-photonics manufacturing for quantum hardware in February 2026.
• These developments indicate that the market is moving from conventional optical-module integration toward semiconductor-level optical connectivity. The resulting opportunity extends beyond PIC sales to foundry processing, photonic packaging, optical-I/O chiplets, and integrated electronic-photonic platforms.

Conclusion

• The Silicon Photonics Semiconductor Market is projected to grow from USD 3.40 billion in 2026 to USD 12.45 billion by 2032 at a CAGR of 24.1%. AI networking, higher optical bandwidth requirements, expanding foundry capacity, and the movement of optical conversion closer to compute are the main growth forces. Datacom and telecom PICs will remain the largest revenue pool, while optical-I/O chiplets and co-packaged photonic engines will expand at the fastest pace. North America will remain a major design and early-demand center, while Asia Pacific will remain critical for manufacturing and packaging. Laser integration, optical testing, packaging complexity, and competition from InP and VCSEL technologies will remain important constraints. Long-term market development will depend on improving manufacturing scale, photonic packaging, optical testing, and integration between photonic and electronic semiconductor technologies.

Key Benefits of this Report

• Insightful Analysis: Detailed market insights across regions, customer segments, policies, socio-economic factors, consumer preferences, and industry verticals.
• Competitive Landscape: Understand strategic moves by key players to identify optimal market entry approaches.
• Market Drivers and Future Trends: Assess major growth forces and emerging developments shaping the market.
• Actionable Recommendations: Support strategic decisions to unlock new revenue streams.
• Caters to a Wide Audience: Suitable for startups, research institutions, consultants, SMEs, and large enterprises.

What Businesses Use Our Reports For

• Industry and market insights, opportunity assessment, product demand forecasting, market entry strategy, geographical expansion, capital investment decisions, regulatory analysis, new product development, and competitive intelligence.

Report Coverage

• Historical data from 2021 to 2024, Base year 2025, and Forecast years from 2026 to 2032
• Growth opportunities, challenges, supply chain outlook, regulatory framework, and trend analysis
• Competitive positioning, strategies, and market share evaluation, and trade analysis
• Revenue growth and forecast assessment across segments and regions
• Company profiling including strategies, products, financials, and key developments

TABLE OF CONTENTS

1. Executive Summary

• 1.1. Market Opportunity and Key Findings
• 1.2. Semiconductor Commercialization Outlook
• 1.3. Principal Revenue Pools

2. Market Overview

• 2.1. Silicon Photonics Semiconductor Architecture
• 2.2. PIC, EIC and Laser Integration
• 2.3. Foundry Manufacturing and Wafer Processing
• 2.4. Optical-I/O and Co-Packaged Semiconductor Integration
• 2.5. Silicon Photonics versus InP and VCSEL Platforms

3. Market Size and Forecast, 2026-2032

• 3.1. Global Market Revenue
• 3.2. Annual Growth Analysis
• 3.3. PIC, Foundry and Photonic-Packaging Revenue

4. Market by Semiconductor Product

• 4.1. Datacom and Telecom Silicon-Photonics PICs
• 4.2. Optical-I/O Chiplets and CPO Photonic Engines
• 4.3. Standalone Modulators and Photodetectors
• 4.4. WDM and Passive Photonic ICs
• 4.5. Foundry Wafer Processing
• 4.6. Semiconductor-Level Photonic Packaging

5. Market by Integration Architecture

• 5.1. Pluggable-Module PICs
• 5.2. Near-Package and On-Board Photonics
• 5.3. Co-Packaged Optical Engines
• 5.4. Processor and Accelerator Optical-I/O Chiplets

6. Market by Application

• 6.1. Data Center and AI Networking
• 6.2. Telecommunications
• 6.3. Optical Compute Interconnect
• 6.4. LiDAR and Physical AI
• 6.5. Photonic Quantum Computing
• 6.6. Sensing and Other Photonic Applications

7. Market by Manufacturing Model

• 7.1. Merchant Silicon-Photonics Foundries
• 7.2. Vertically Integrated Semiconductor Manufacturers
• 7.3. Fabless Photonic Semiconductor Companies
• 7.4. Foundry plus Advanced-Packaging Ecosystems

8. Market by Customer Type

• 8.1. Networking Semiconductor Vendors
• 8.2. Optical Component and Module Suppliers
• 8.3. Hyperscale and AI Platform Companies
• 8.4. Quantum, LiDAR and Photonic-System Developers
• 8.5. Research and Defense Programs

9. Regional Market

• 9.1. North America
  • 9.1.1. United States
  • 9.1.2. Canada
• 9.2. Asia Pacific
  • 9.2.1. Taiwan
  • 9.2.2. Japan
  • 9.2.3. China
  • 9.2.4. South Korea and Southeast Asia
• 9.3. Europe
• 9.4. Rest of World

10. Market Dynamics

• 10.1. Drivers
  • 10.1.1. AI Optical I/O and CPO Adoption
  • 10.1.2. Expansion of Foundry Silicon-Photonics Capacity
  • 10.1.3. 1.6T and 3.2T Optical Integration Density
  • 10.1.4. Growth beyond Datacom into LiDAR and Quantum
• 10.2. Restraints
  • 10.2.1. Laser Integration and III-V Material Requirements
  • 10.2.2. Fiber Attach and Photonic Test Complexity
  • 10.2.3. Serviceability of Deeply Integrated Optics
  • 10.2.4. Competition from InP EML and VCSEL Platforms

11. Competitive Landscape

• 11.1. Market Structure and Competitive Intensity
• 11.2. Foundry Capacity and Process-Platform Positioning
• 11.3. PIC and Optical-I/O Semiconductor Strategies
• 11.4. Electronic-Photonic Integration and Packaging Ecosystems
• 11.5. Hyperscale, Networking and Photonic-System Partnerships

12. Company Profiles

• 12.1. Tower Semiconductor
• 12.2. GlobalFoundries
• 12.3. Intel
• 12.4. Marvell Technology
• 12.5. Broadcom
• 12.6. Cisco
• 12.7. Coherent
• 12.8. Lumentum
• 12.9. Lightmatter
• 12.10. Ayar Labs
• 12.11. TSMC
• 12.12. imec
• 12.13. STMicroelectronics
• 12.14. Soitec
• 12.15. Fabrinet

13. Recent Developments

14. Appendix