STMicroelectronics Launches 300mm Wafer Production of PIC100 Silicon Photonics for AI Data Centers

July 26, 2026

EE Times International

Overview

STMicroelectronics has begun mass production of its PIC100 silicon photonics platform on 300mm wafers to meet the escalating demands of AI data centers. This platform integrates BiCMOS for laser drivers with advanced packaging options, delivering more compact, high-bandwidth optical engines. As AI clusters expand, optical interconnects are becoming as crucial as computational power, resolving data transfer bottlenecks and improving power efficiency, thereby accelerating the scaling of AI workloads.

In Depth

Key Findings

STMicroelectronics has transitioned its PIC100 silicon photonics platform to mass production on 300mm wafers to address the burgeoning demands of AI data centers. This strategic decision reflects the increasing importance of optical interconnects within AI clusters, now considered as critical as computational power itself. This move facilitates the realization of more compact, higher-bandwidth optical engines, effectively breaking through the limitations of traditional electrical interconnects.

Technical / Clinical Details

STMicroelectronics’ PIC100 platform utilizes a BiCMOS process that integrates silicon photonics with CMOS technology, specifically optimizing the performance and efficiency of laser driving circuits. Coupled with advanced packaging options, this achieves miniaturization and higher density for the entire optical module. Manufacturing on 300mm wafers is essential for significantly increasing production capacity and reducing per-unit costs. As AI workloads require massive amounts of data to be exchanged at high speeds between GPU clusters, optical interconnects are gaining traction as a solution that dramatically enhances data transfer rates and substantially reduces power consumption compared to conventional copper wiring.

Background & Context

With the evolution of AI, particularly for large language models and complex inference tasks, AI clusters comprising tens of thousands of GPU chips operating in concert are being built. In such environments, data transfer between chips and servers becomes a bottleneck, significantly impacting overall performance. As Moore’s Law approaches its limits, the semiconductor industry is focusing not only on improving data processing capabilities but also on enhancing data transfer efficiency and power efficiency. Silicon photonics is a promising solution to this challenge, with major players like NVIDIA and Intel investing heavily in this domain. STMicroelectronics’ transition to 300mm wafers is seen within the industry as accelerating the full-scale industrialization of optical communication technology in the AI era.

Strategic Significance & Outlook

The initiation of 300mm wafer silicon photonics mass production by STMicroelectronics holds the potential to significantly improve the cost and energy efficiency in the construction of AI data centers. The widespread adoption of this technology will enhance the scalability of AI computing, enabling the deployment of larger and more complex AI models. In the future, a transition towards more tightly integrated optical interconnect solutions, such as co-packaged optics (CPO) and on-chip optical I/O, is expected to accelerate. This will lead to fundamental changes in data center design, strengthening the technological foundation for AI to permeate deeply into all aspects of society.

Source: https://www.eetimes.com/ai-data-centers-push-silicon-photonics-toward-300-mm-scale/

Get our weekly technology intelligence — free

Receive an infographic that lets you judge at a glance whether each field’s analysis report is worth reading.

Subscribe Free — Weekly Tech Intelligence

By subscribing, you’ll receive Troy-Technical’s weekly technology intelligence newsletter.