Photonic integrated circuit market to reach $54B by 2035

Photonic integrated circuits (PICs) are revolutionizing optical communications and computing by offering higher data transfer speeds and lower energy consumption compared to traditional electronic circuits. With growing demand from artificial intelligence (AI) data centers, telecommunications, quantum computing, and sensing applications, the PIC market is projected to surpass $54B by 2035, according to IDTechEx’s latest report, Silicon Photonics and Photonic Integrated Circuits 2025-2035: Technologies, Market, Forecasts (see Fig. 1).

Role of PICs within AI data centers

The rapid expansion of AI use, as well as expansion of foundation models requiring huge numbers of parameters, is leading to an exponential increase in data processing requirements within data centers. This surge necessitates high-bandwidth, low-latency communication channels to efficiently handle the massive data flows between AI accelerators and storage systems. PICs have emerged as a pivotal technology to address these demands because they offer enhanced data transfer rates and energy efficiency over traditional electronic interconnects.

Imperative for high-bandwidth optical interconnects

Training and executing large modern AI models requires substantial data movement within data centers due to the large amounts of matrix math that must take place, which is frequently split across multiple graphic processing unit (GPU) nodes. Traditional electronic interconnects face challenges meeting the required bandwidth and latency specifications, which often leads to bottlenecks that impede overall system performance.

PICs mitigate these issues by using light for data transmission to achieve higher bandwidths and reduced latencies. Energy losses via light adsorption are typically lower than equivalent losses via electrical resistance, which improves energy efficiency and increases range.

Recent developments underscore the critical role of PICs within AI data centers. For example, STMicroelectronics is introducing a new silicon-germanium PIC production process, in collaboration with Amazon Web Services (AWS). Production is slated to begin during the latter half of this year and pluggable transceivers manufactured via the process will be integrated into AWS’ infrastructure—highlighting the industry’s shift toward photonic solutions to meet AI’s data transfer demands.

Co-packaged optics: Integrating photonics and electronics

To further enhance performance and energy efficiency, the integration of photonics directly with electronic components is gaining traction. Co-packaged optics (CPO) involves the close coupling of photonic and electronic integrated circuits, which reduces signal loss and power consumption associated with traditional interconnects. This approach is particularly beneficial within AI data centers, where the minimization of electrical path lengths not only reduces latency while increasing throughput, but also reduces power consumption via minimization of resistive losses.

Silicon photonics: Advantages and limitations

Silicon photonics is the leading technology within the PIC market due to its compatibility with mature semiconductor manufacturing processes. It’s normally based on a silicon-on-insulator (SOI) platform of silicon on silica, although in some cases silicon nitride is used instead of silicon on these wafers.

TSMC announced its entry into the silicon photonics market mid-2024, and focused first on pluggable transceivers and later on CPO, which marks another milestone in confidence for this technology.

Key advantages of silicon photonics include:

Integration with complementary metal-oxide semiconductor (CMOS) manufacturing. Silicon photonics leverages existing foundry infrastructure, which reduces production costs and enables mass deployment.

Scalability. Silicon photonics operates within standard semiconductor processes, so it can be scaled alongside conventional electronic circuits.

Energy efficiency. Compared to electrical interconnects, silicon photonics enables lower power consumption for high-speed data transfer.

Despite these benefits, silicon photonics has notable limitations that alternative material platforms could address. These include the difficulty of laser integration, since silicon is an indirect bandgap semiconductor and therefore a poor laser material. Compared to some alternative materials, modulation performance and losses in optical waveguides are also inferior.

Material innovations for PICs

While silicon dominates the PIC market, emerging materials are addressing its limitations (see Fig. 2):

Indium phosphide (InP). A direct-bandgap semiconductor that enables efficient light generation and detection, InP is widely used in telecommunications and is expected to gain market share as demand for high-speed optical interconnects rises.

Thin-film lithium niobate (TFLN). With strong optoelectric properties and low loss, TFLN is emerging as a candidate for high-performance modulation applications, including quantum computing and ultrahigh-speed optical communication. But its technology readiness level remains low due to challenges integrating it with standard CMOS processes.

As demand for high-performance transceivers increases, opportunities for materials with extremely high modulation performance are expected to grow. Alongside evolution in data center transceiver requirements, widening applications for silicon photonics are expected to increase the opportunity for alternative material platforms during the next 10 years.

Our report also categorizes the opportunities for a broader range of emerging PIC materials, including barium-tin oxide and polymer-on-insulator.

PICs beyond communications

While data communications remains the largest market for PICs, other applications are emerging:

LiDAR. Frequency-modulated continuous-wave (FMCW) LiDAR, enabled by PICs, enhances the performance of autonomous vehicles and industrial automation.

Biosensors and gas sensors. PICs fabricated with silicon nitride are enabling compact, highly sensitive sensing devices for healthcare and environmental monitoring.

Quantum computing. Photonic quantum computing is a rapidly evolving field, and PICs are playing a crucial role in scaling quantum architectures. Companies developing trapped-ion and photon-based quantum processors are investing in integrated photonic solutions to improve stability and scalability.

Market growth and industry landscape

The silicon photonics market alone is expected to maintain a compound annual growth rate (CAGR) of 20% during the next decade. While this is largely driven by the ongoing AI boom, emerging use cases for PICs including LiDAR, biosensing, and quantum computing are expected to see significant markets within the latter half of this period.

Our report provides in-depth analysis of key players, technology trends, and market forecasts, including:

• A breakdown of co-packaged optics and its potential impact on the industry.
• Benchmarking of photonic materials, including emerging options such as TFLN and BTO.
• Market forecasts for PIC transceivers in AI, data centers, 5G, and quantum applications.
• Manufacturing techniques and supply chain dynamics.

As demand for high-speed, energy-efficient optical solutions grows, PICs are poised to play a fundamental role shaping the future of computing, communications, and sensing technologies.

For more details from IDTechEx’s latest report, visit www.IDTechEx.com/SemiPIC.