Light people: Radha Nagarajan, Marvell CTO and NAE member

Wan and Guo Light: Science & Applications (2026)15:250 https://doi.org/10.1038/s41377-026-02305-6 LIGHT PEOPLE www.nature.com/lsa Open Access Light people: Radha Nagarajan, Marvell CTO and NAE member 1✉ and Chenzi Guo2 ✉ 1234567890():,; 1234567890():,; 1234567890():,; 1234567890():,; Yating Wan Short Bio of Radha Nagarajan. Dr. Nagarajan is currently the Senior Vice President and Chief Technology Officer of Marvell’s Optical Engineering Group. At Marvell, he manages the development of the company’s optical platform products and technology. Concurrently, he is a Visiting Professor at the Department of Electrical and Computer Engineering at the National University of Singapore. He received his B.Eng. from the National University of Singapore, M.Eng. from the University of Correspondence: Yating Wan () or Chenzi Guo () 1 Electrical and Computer Engineering, the Computer, Electrical and Mathematical Sciences and Engineering Division, King Abdullah University of Science and Technology, Thuwal, Saudi Arabia 2 Changchun Institute of Optics, Fine Mechanics and Physics, Chinese Academy of Sciences, Changchun, China Tokyo, and Ph.D. from the University of California, Santa Barbara, all in Electrical Engineering. Dr. Nagarajan has been elected to the National Academy of Engineering (US). His other recognitions include the IEEE/LEOS Aron Kressel Award, the IPRM Award and the OPTICA David Richardson Medal for breakthrough work in the development and manufacturing of photonic integrated circuits. He has been awarded more than 250 US patents. He is a Fellow of the IEEE, OPTICA, and IET. Q: You have worked across several companies, and your remarkable achievements have led to your election to US National Academy of Engineering. Through your career path, what dominates in your thinking about ‘what matters’ as integration scaled up? A: After my PhD with Prof. John Bowers, for the past 30 years, I’ve been working on photonic integration—first at Infinera, then Inphi, and now Marvell. If you look at these roles, they are a snapshot of how the industry has evolved. When I started at Infinera in 2001, our goal was to build photonic integrated circuits based on indium phosphide, primarily targeting long-haul applications. At that time, people were not talking about datacenter interconnects. Servers were still connected using conventional electrical cables, such as Cat5 or Cat6. In 2013, I transitioned to Inphi and moved from indium phosphide to silicon photonics. While indium phosphide was already a mature platform, silicon photonics was emerging quickly. At that stage, our focus was still largely on applications outside the datacenters. Later, when I moved to Marvell, the explosion of AI and HPC dramatically expanded the market. Silicon photonics is now being deployed at large scale—not only for outside the datacenter but also within the datacenter. My work during this period has been more centered on heterogeneous integration. © The Author(s) 2026 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. Wan and Guo Light: Science & Applications (2026)15:250 A common theme that has guided me in my career all along is turning advanced technology into a real product. This requires achieving large-scale manufacturing, maintaining yield at a profitable level, and ensuring reliable deployment in the field. These challenges are where we Fig. 1 Dr. Radha Nagarajan elected to US National Academy of Engineering Fig. 2 Dr. Radha Nagarajan was Prof. John Bowers’ first Ph.D. student to graduate, The photo was taken at John’s retirement party at UCSB, with two others who graduated subsequently Page 2 of 7 invested most of our effort, always with a strong focus on market needs. People often underestimate how much time, effort, and investment it takes to turn a promising technology into a commercial product, but this process has consistently been at the core of my work. (Figs. 1, 2) Q: At what point do you think silicon photonics will take over discrete components/modules? A: I would say silicon photonics has already taken over to a large extent, especially at 100 G and higher data rates. Yield is not something unique to just silicon photonics—it is a critical requirement for any practical technology. However, what silicon photonics uniquely enables is particularly important for the industry. First, it allows companies without their own foundry, like Marvell, to design and manufacture highly complex photonic components by leveraging the existing semiconductor infrastructure. Second, it provides true scalability. Marvell, for example, has been one of the well-known companies implementing silicon photonics for applications between datacenters. Intel pioneered silicon photonics for applications inside the datacenters. Today, many of the large transceiver companies have mostly transitioned to silicon photonics. Given this level of adoption, it is fair to say that silicon photonics has already ‘taken over’ in many key segments of the optical interconnect market. Q: Many silicon photonics systems still rely on external laser packaging. Do you see integrated light sources becoming mainstream, or will off-chip lasers dominate for practical reasons? A: If you look at the integrated laser work published over the years, including our early demonstrations, Intel’s results, and developments at Infinera (now Nokia), it is clear that integrating light sources heterogeneously on silicon or monolithically in InP is commercially viable. The real obstacle is not technical capability, but perception—many people still assume that integrated lasers will suffer from reliability issues. One concern stems from the behavior of indium phosphide lasers at elevated temperatures. It is true that performance can degrade at high temperatures, but when operated around a base temperature of roughly 70–75 °C, these lasers can already operate perfectly well. In fact, Intel has demonstrated strong reliability in this regime. The key is careful design. Temperature variations can be significant, and the resulting wavelength shifts must be properly managed. In practice, engineers have already developed effective strategies to handle these challenges. So, f (...truncated)