As the demand for artificial intelligence (AI) accelerates, the computing power required to train increasingly complex models has surged, leading to a pivotal juncture in data center technology. Companies are facing a critical bottleneck in high-speed interconnection capabilities, which is becoming the primary constraint on AI performance breakthroughs. Traditional electrical transmission methods, constrained by energy consumption and distance limitations, are struggling to keep pace with the massive data traffic required for advanced AI applications.
With transmission rates now exceeding 400Gbps and evolving towards 1.6Tbps, the limitations of copper wires are becoming apparent, as they suffer from significant signal attenuation and rising energy costs. In response, the tech industry is increasingly turning to **silicon photonics technology**, which utilizes photons for data transmission, addressing the challenges of energy consumption and signal delay. Silicon photonic chips, which integrate semiconductor and photonics technologies, are being recognized as essential in overcoming these obstacles, offering high bandwidth, low power consumption, and compatibility with existing CMOS processes.
Industry analysts anticipate that 2026 will mark a significant turning point for the widespread commercialization of silicon photonics technology, with a Nomura Securities report projecting a doubling of shipments for 800G and 1.6T optical modules within that year. The penetration of silicon photonics in this market is expected to reach between 50% and 70%, positioning these chips as a crucial driver of industry growth. As the core component of optical modules, silicon photonic chips account for 30% to 70% of their overall cost, making their production capacity and technological sophistication vital to the progress of downstream industries.
However, a production capacity gap of 5% to 15% is anticipated in 2026, intensifying competition among foundry giants. As demand surges, the design logic of chips is being reshaped, resulting in a renewed contest in the global wafer foundry sector. This burgeoning market indicates that controlling silicon photonics foundry capabilities is becoming synonymous with securing a foothold in next-generation AI and high-performance computing technologies.
In this competitive environment, companies like **Tower Semiconductor** are ramping up efforts. Tower plans to double its silicon photonics manufacturing capacity by mid-2026, currently operating facilities in the United States and Japan. CEO Russell Ellwanger highlighted the company’s leadership in silicon-germanium and silicon photonics technologies, attributing a tripling of Tower’s market value to the heightened demand for silicon photonics. In Q4 2025, Tower reported record revenues of $440 million, marking a 14% increase year-over-year, with net profits of $80 million.
Tower’s aggressive strategy includes a planned increase in total investment in silicon photonics to $920 million, with hopes to quintuple wafer production capacity by late 2026. The company has secured over 70% of its production capacity through customer pre-orders, underscoring the urgency and eagerness of clients for silicon photonic solutions.
Strategic partnerships are also reshaping the landscape. Tower is collaborating with **NVIDIA** to develop 1.6T optical modules, reflecting the rapid expansion of AI clusters and the need to enhance data transmission efficiency among GPUs and network devices. As demand for high-performance optical devices grows, Tower’s ability to deliver reliable, high-quality manufacturing processes will be pivotal for success.
**GlobalFoundries** is also making significant moves, having acquired **Advanced Micro Foundry** in late 2025 to expand its silicon photonics capabilities. This acquisition positions GlobalFoundries as the largest pure-play silicon photonic chip foundry by revenue. The integration of Advanced Micro Foundry’s resources will enable GlobalFoundries to enhance its technology portfolio and R&D capabilities in the critical Singapore market, expanding its offerings in photonic chips that play a crucial role in AI, automotive, and quantum computing sectors.
GlobalFoundries’ Fotonix platform, developed in 2022, marks a key advancement, integrating 300-mm photonics features with RF-CMOS processes to enable faster, more efficient data transmission. This combination of photonic technology and advanced manufacturing is expected to meet the increasing demands for speed and low power consumption in data centers.
Meanwhile, **UMC** is pursuing a collaborative strategy by partnering with **imec**, acquiring the iSiPP300 silicon photonics process to enhance its own technology roadmap. This partnership aims to leverage imec’s proven 12-inch wafer technology, bolstering UMC’s capabilities in producing scalable photonic integrated circuits.
As the AI data load continues to escalate, companies are recognizing that silicon photonics technology will play a pivotal role in addressing the limitations of traditional copper interconnections. With numerous foundries vying for leadership in this space, the next few years will be critical in determining which players will secure their position in the evolving landscape of high-performance computing and AI infrastructure.
See also
Turiyam.ai Secures $4M in Pre-Seed Round to Develop Full-Stack AI Hardware Platform
NTT DOCOMO and NTT Demonstrate Low-Latency AI Video Analytics via 5G and Remote GPUs
AI Tools Transform Civil Engineering: Enhance Efficiency with OpenSite+ and ALICE Technologies
OpenAI Secures $200M Defense AI Contract with Pentagon Amid Anthropic Controversy
Canada Invests $1.16M in Landing Zones Canada to Advance AI for Weather Sampling and Defence
