Chip-to-Chip Optical Interconnects Market | Size, Growth Forecast, Market Share 

 Market Summary and Growth Forecast

The global Chip-to-Chip Optical Interconnects Market is valued at $1,185 million in 2026 and is expected to appreciate to $4,060 million by 2035, at a CAGR of 14.7%. The market covers optical technologies that move data directly between processors, accelerators, switches, memory devices, and other high-performance chips using optical signals rather than relying only on conventional electrical traces. Its business relevance is increasing as AI workloads, high-performance computing, and advanced data centers push interconnect bandwidth beyond what traditional electrical connections can efficiently deliver.

Market indicator	2026	2035	Outlook
Global market value	$1,185 million	$4,060 million	14.7% CAGR
AI/HPC-related demand	~42% of demand	~58% of demand	Fast expansion
Data-center applications	~61% of demand	~68% of demand	Core revenue base
High-bandwidth optical architectures	Emerging	Mainstream in leading systems	Strategic shift

Several forces are shaping the Chip-to-Chip Optical Interconnects Market through 2035. AI accelerators are creating much higher communication requirements between compute units, while chiplet-based architectures are increasing the number of connections inside advanced computing platforms. Optical links can help address bandwidth, reach, signal integrity, and power constraints as electrical interconnects become harder to scale.

Production economics also matter. Optical engines, photonic integrated circuits, laser sources, modulators, coupling structures, and advanced packaging must mature together. Greater integration should gradually reduce component count and improve manufacturing consistency.

Key consumers include hyperscale data-center operators, cloud service providers, AI infrastructure companies, HPC system manufacturers, semiconductor companies, networking equipment suppliers, and advanced computing platform developers. The strategic value of optical interconnects is shifting from simply increasing bandwidth toward improving the overall energy efficiency and scalability of compute systems.

Market Segmentation and Forecast Scope

The Chip-to-Chip Optical Interconnects Market can be assessed across product architecture, application, end user, and geography. These dimensions help separate demand generated by AI computing from conventional networking and identify where optical links are moving from specialized deployments toward broader system adoption.

By Product Type

The market includes Optical Engines, Co-Packaged Optics, Optical I/O Modules, and other integrated optical interconnect solutions. Optical engines currently represent a major revenue pool because they can be integrated into high-bandwidth computing and switching platforms without requiring a complete redesign of the host system.

Co-packaged and tightly integrated optical architectures are more strategic for the longer term. They place optical functionality closer to the processor or switch ASIC, reducing electrical path length and potentially improving bandwidth density and power efficiency.

By Application

Applications include AI and Machine Learning, High-Performance Computing, Data Center Networking, Telecommunications, and other advanced computing systems. AI and machine learning is the fastest-growing application because accelerator clusters increasingly depend on rapid communication between processors and memory or between multiple compute devices.

For system designers, the commercial question is no longer only how much bandwidth an interconnect provides. Power consumed per transmitted bit and the ability to scale the architecture are becoming equally important.

By End User

The end-user base covers Hyperscale Data Centers, Cloud Service Providers, Semiconductor and System Companies, HPC Institutions, and networking equipment manufacturers. Hyperscale and cloud operators remain the most strategically important customers because large AI clusters can justify the higher initial cost of optical infrastructure.

By Region

The geographic scope covers North America, Europe, Asia Pacific, and LAMEA. North America remains a leading market because of its concentration of hyperscale computing, AI infrastructure, semiconductor design, and advanced data-center investments. Asia Pacific has strong manufacturing depth and a growing base of semiconductor and electronics companies.

Segmentation dimension	Leading / strategic segment	2026 indicative share	Growth view
Product type	Optical Engines	~38%	Established base
Application	AI & Machine Learning	~34%	Fastest-growing
End user	Hyperscale Data Centers	~46%	Strong expansion
Region	North America	~43%	Leading revenue market

The most important forecast shift is toward higher-density optical integration. As compute architectures become increasingly modular, demand should move from discrete optical components toward solutions designed as part of the processor, package, or system architecture.

Market Trends and Business Innovations

Innovation in the Chip-to-Chip Optical Interconnects Market is moving toward shorter optical paths, higher bandwidth density, lower energy consumption, and deeper integration with advanced semiconductor packaging. The focus is gradually shifting from standalone optical modules to optical connectivity that is designed alongside the compute architecture.

R&D Evolution

Research is increasingly centered on optical I/O, silicon photonics, advanced laser integration, optical coupling, and packaging methods that can support very high data rates. Developers are also working to reduce the electrical distance between compute chips and optical conversion points.

This matters because conventional electrical links can face increasing losses and power requirements as signaling speeds and physical distances rise. Optical conversion placed closer to the chip can reduce some of these limitations.

Technology Evolution

Several technologies are progressing together:

• Silicon photonics is becoming an important platform for integrating optical functions with semiconductor manufacturing processes.
• Co-packaged optics is being developed for systems where conventional pluggable optics may not provide sufficient bandwidth density or power efficiency.
• Optical I/O is gaining attention for direct connectivity between processors, accelerators, memory systems, and networking components.
• Advanced packaging is becoming a critical enabler because optical and electronic dies must operate as a tightly coordinated system.
• Higher-speed signaling is increasing interest in optical links for next-generation accelerator and switch architectures.

AI is relevant mainly as a demand catalyst rather than as an optical technology itself. Large AI clusters require fast movement of model parameters, intermediate data, and memory traffic. This creates a strong incentive to improve interconnect bandwidth while controlling energy consumption.

Partnerships and Industry Activity

The ecosystem is becoming more collaborative. Semiconductor designers, photonics specialists, optical component suppliers, packaging companies, and hyperscale infrastructure developers are increasingly working together. The commercial emphasis is shifting toward complete interconnect architectures rather than individual optical components.

Major industry development programs have also focused on optical I/O, co-packaged optics, silicon photonics, and chiplet-based architectures. These initiatives are important because successful deployment requires compatibility across chips, packages, optical engines, manufacturing processes, and system software.

Innovation area	Primary objective	Expected business impact
Optical I/O	Bring optical connectivity closer to compute	Higher bandwidth density
Co-packaged optics	Integrate optics with switching/compute packages	Lower interconnect power
Silicon photonics	Increase optical integration	Better scalability
Advanced packaging	Combine optical and electronic components efficiently	Smaller system footprint
Chiplet architectures	Enable modular compute designs	More interconnect demand

In our view, the next phase of the market will be shaped less by optical bandwidth alone and more by how effectively optical connectivity can be integrated into the complete semiconductor package. This may create new opportunities for suppliers that can combine photonics, packaging, testing, and high-volume manufacturing capabilities.

As these technologies mature, the Chip-to-Chip Optical Interconnects Market should increasingly become part of the core architecture of AI and high-performance computing systems rather than remaining a specialized connectivity layer.

Competitive Intelligence and Benchmarking

The Chip-to-Chip Optical Interconnects Market remains technology-intensive, with competition centered on bandwidth density, energy per bit, packaging integration, manufacturing readiness, and compatibility with emerging AI accelerator architectures. The competitive field includes established semiconductor companies as well as specialized photonics developers.

Intel Corporation

Intel Corporation has one of the broadest technology positions in the segment, combining silicon photonics, optical engines, semiconductor manufacturing, and compute architecture. Its portfolio spans high-speed photonic components and optical compute interconnect technology designed for direct integration with CPUs, GPUs, and other compute devices. The company has demonstrated an optical I/O chiplet integrated with a processor, positioning it strongly in the transition from rack-level optical connectivity toward package-level optical links. Its advantage is the ability to connect photonics R&D with large-scale semiconductor manufacturing.

Broadcom Inc.

Broadcom Inc. has a strong position across switching silicon, optical connectivity, SerDes, DSPs, and co-packaged optics. Its strategy is particularly focused on integrating optical engines closely with high-bandwidth switching and accelerator silicon. The company is also targeting AI networking, where optical connectivity can reduce electrical path losses. Broadcom’s broad semiconductor and networking customer base gives it an important commercial advantage as CPO moves toward production deployment.

Marvell Technology

Marvell Technology is positioning itself around optical connectivity for AI scale-up and scale-out systems. Its portfolio combines silicon photonics, optical engines, DSP technologies, and custom accelerator connectivity. The company has demonstrated high-bandwidth optical engine technology and is developing CPO architectures for custom AI accelerators. Its market position is strengthened by its relationships across cloud infrastructure and custom silicon ecosystems.

Lightmatter

Lightmatter is a specialist photonic computing and interconnect company focused heavily on AI infrastructure. Its technology portfolio centers on 3D photonic interconnects, optical engines, and co-packaged architectures designed to overcome the limited I/O perimeter of conventional chips. The company is differentiating itself through area-based optical I/O and high-density photonic packaging. Its strategic position is strongest in next-generation AI systems where conventional electrical connectivity becomes a scaling constraint.

Ayar Labs

Ayar Labs focuses on optical I/O designed to move data between processors, accelerators, and memory systems. Its approach combines photonic chiplets, optical interfaces, and tightly integrated packaging. The company has attracted interest from semiconductor and computing ecosystem participants because its architecture is designed specifically around the growing need for energy-efficient chip-to-chip data movement. Its position is particularly relevant to AI and high-performance computing.

Cisco Systems

Cisco Systems participates through optical networking, silicon, switching infrastructure, and data-center connectivity. Its relevance to the market is broader than a pure chip-to-chip optical supplier, but its networking ecosystem gives it exposure to the transition toward higher-bandwidth optical architectures. The company is well positioned where optical interconnects converge with AI networking, switching, and large-scale data-center infrastructure.

Celestial AI

Celestial AI is focused on photonic fabric technology for AI and memory-intensive computing. Its approach emphasizes optical connectivity between compute and memory resources, addressing bandwidth and power limitations in increasingly distributed accelerator architectures. The company occupies a specialized position in the emerging photonic fabric segment and could benefit as AI systems place greater emphasis on memory bandwidth and resource disaggregation.

Company	Primary strength	Competitive position
Intel Corporation	Silicon photonics and optical I/O	Established semiconductor + photonics
Broadcom Inc.	CPO, switching silicon, optical engines	Strong commercial scale
Marvell Technology	Silicon photonics and AI interconnect	Strong infrastructure position
Lightmatter	3D photonic interconnects	High-growth specialist
Ayar Labs	Optical I/O chiplets	Specialized technology player
Cisco Systems	Optical networking and switching	Large infrastructure ecosystem
Celestial AI	Photonic fabric and memory connectivity	Emerging specialist

The competitive advantage is moving toward system-level integration. Companies that can control the optical engine, electrical interface, packaging and manufacturing path together should have a stronger route to high-volume AI deployments.

Regional Landscape and Adoption Outlook

Regional adoption of the Chip-to-Chip Optical Interconnects Market is closely linked to AI infrastructure investment, semiconductor manufacturing capability, hyperscale data-center capacity, advanced packaging, and government support for strategic computing technologies.

United States

The United States remains the leading commercial center. Its advantage comes from hyperscale cloud companies, AI accelerator developers, advanced semiconductor designers, and a strong photonics startup ecosystem. Large investments in AI data centers are creating an early customer base for optical I/O and CPO technologies.

Government support through the CHIPS and Science Act also strengthens domestic semiconductor production and advanced packaging capabilities. This creates a favorable environment for optical interconnect suppliers that need close integration with leading-edge silicon.

Europe

Europe has a strong research base in photonics, semiconductor equipment, telecom infrastructure, and advanced manufacturing. Adoption is more measured than in the United States, but the region benefits from semiconductor research programs and the European Chips Act.

Germany, France, the Netherlands, Belgium, and the United Kingdom are important technology centers. Europe is particularly relevant for photonics research, optical components, semiconductor equipment, and specialized manufacturing.

China

China is developing a large domestic semiconductor and data-center ecosystem. Government-backed investment in advanced computing and semiconductor capabilities supports demand for higher-bandwidth interconnects. Domestic cloud platforms and AI companies are also increasing their requirements for accelerator connectivity.

Export controls on advanced semiconductor technologies create a different market structure. These restrictions can encourage domestic development of optical and photonic alternatives, although access to certain advanced manufacturing technologies remains a constraint.

India

India represents an emerging opportunity rather than a current volume leader. Growth is supported by data-center expansion, cloud adoption, AI infrastructure development, and government efforts to build a domestic semiconductor ecosystem. The India Semiconductor Mission and associated incentive programs are improving the foundation for semiconductor and packaging investments.

The country’s opportunity is strongest in data-center infrastructure and semiconductor ecosystem development. Large-scale domestic production of advanced optical interconnect components remains less mature than in the United States, Japan, South Korea, China, and parts of Europe.

Japan

Japan has strong capabilities in optical components, precision manufacturing, semiconductor materials, packaging, and electronics. Government support for semiconductor revitalization is encouraging new investment across the broader chip ecosystem.

Japanese companies are also important suppliers of optical components, materials, equipment, and manufacturing technologies. This gives Japan an influence that extends beyond domestic demand.

South Korea

South Korea is strategically important because of its leadership in memory, semiconductor manufacturing, advanced electronics, and AI infrastructure. The country’s large semiconductor manufacturers create a natural ecosystem for high-bandwidth optical connectivity.

Government-backed semiconductor investment and the country’s focus on AI and advanced packaging should support adoption. Optical interconnects are particularly relevant as memory and accelerator architectures become more bandwidth intensive.

Middle East

The Middle East is an emerging demand center, particularly Saudi Arabia and the United Arab Emirates. Large AI and cloud infrastructure programs are creating opportunities for high-bandwidth networking and optical connectivity. However, the region remains primarily an infrastructure adopter rather than a major manufacturing base.

Region / country	Adoption position	Main advantage	Key constraint
United States	Leading	AI infrastructure, hyperscalers, chip design	High deployment cost
Europe	Established research base	Photonics and semiconductor R&D	Slower commercial scaling
China	High-growth	Domestic AI and semiconductor ecosystem	Technology access restrictions
India	Emerging	Data centers and semiconductor incentives	Limited local photonics manufacturing
Japan	Strong technology base	Optics, materials, precision manufacturing	Smaller domestic AI market
South Korea	Strategic growth market	Memory and semiconductor leadership	Concentrated industry structure
Middle East	Emerging adopter	New AI/data-center investment	Limited local manufacturing

North America should remain the earliest large-scale commercialization market, while Asia Pacific is likely to become increasingly important as semiconductor production, AI infrastructure, and advanced packaging capacity expand.

Recent Developments + Opportunities & Restraints

Recent Developments

March 2025 — United States: Lightmatter introduced a new 3D co-packaged optical architecture designed for AI processors and switches. The company stated that its architecture supports 32 Tbps and 64 Tbps configurations and more than 200 Tbps of total I/O per package. The development highlights the move toward placing optical connectivity across the chip surface rather than restricting I/O to the package edge.

March 2025 — United States: Broadcom expanded its optical interconnect portfolio at OFC 2025, highlighting CPO, 200G/lane SerDes and DSP technology, 400G optical connectivity, and PCIe Gen6 over optics. The company also outlined a roadmap toward 200T optical interconnect systems for AI infrastructure.

March 2025 — United States: Marvell demonstrated a 1.6T silicon photonics light engine using 200 Gbps per lane connectivity for rack-scale AI infrastructure. The development is aimed at reducing power and simplifying integration of high-bandwidth optical connectivity.

August 2025 — United States: Lightmatter demonstrated a 16-wavelength bidirectional optical link over a single standard single-mode fiber, reaching 800 Gbps bidirectional bandwidth. The development targets the increasing fiber-density requirements created by large AI clusters.

March 2026 — United States: Lightmatter and Qualcomm Technologies announced successful sampling of a CPO chiplet demonstrating 1.6 Tbps per fiber using a 16-wavelength architecture. The companies linked the development to the need for higher-bandwidth AI accelerator connectivity.

Opportunities

AI accelerator connectivity: Growing AI clusters create a clear opening for optical I/O between processors, accelerators, switches, and memory resources. Higher bandwidth density can help reduce the physical and power burden of conventional electrical links.

Advanced packaging: As chiplet architectures become more common, optical connectivity can become part of the package itself. This creates opportunities for suppliers that can combine photonics, packaging, testing, and semiconductor interfaces.

Emerging data-center markets: India, the Middle East, and other fast-developing digital infrastructure markets can create new demand as AI-ready data centers expand.

Business Restraints

The main barriers remain high integration costs, optical packaging complexity, thermal management, manufacturing yield, testing requirements, and the need for ecosystem-level compatibility. Customers also need strong reliability before replacing established electrical or pluggable architectures at scale.

The commercial inflection point will come when optical I/O delivers a clear total-system advantage rather than simply a higher headline bandwidth figure.