Advanced Photonics Coalition

𝐈𝐧𝐧𝐨𝐥𝐮𝐦𝐞 𝐂𝐨𝐨𝐫𝐝𝐢𝐧𝐚𝐭𝐞𝐬 𝐭𝐡𝐞 𝐇𝐢𝐏𝐄𝐑 𝐂𝐏𝐎 𝐏𝐫𝐨𝐣𝐞𝐜𝐭 𝐭𝐨 𝐃𝐞𝐯𝐞𝐥𝐨𝐩 𝐋𝐚𝐬𝐞𝐫 𝐒𝐨𝐮𝐫𝐜𝐞𝐬 𝐟𝐨𝐫 𝐂𝐨-𝐏𝐚𝐜𝐤𝐚𝐠𝐞𝐝 𝐎𝐩𝐭𝐢𝐜𝐬 As the performance of modern data centers continues to increase alongside the rapid growth of artificial intelligence, new demands are being placed on data transmission speed, energy efficiency, and the integration of optical components. One of the most promising approaches is Co-Packaged Optics (CPO), an architecture in which optical modules are integrated in close proximity to high-performance switches and processors. To support the development of this technology, Innolume GmbH is coordinating the HiPER CPO (High PowerLasers for Co-Packaged Optics) research project, funded by the German Federal Ministry for Research, Technology and Space (BMFTR). The project runs from February 1, 2026, to January 31, 2029. 𝐃𝐞𝐯𝐞𝐥𝐨𝐩𝐢𝐧𝐠 𝐭𝐡𝐞 𝐍𝐞𝐱𝐭 𝐆𝐞𝐧𝐞𝐫𝐚𝐭𝐢𝐨𝐧 𝐨𝐟 𝐋𝐚𝐬𝐞𝐫 𝐒𝐨𝐮𝐫𝐜𝐞𝐬 Within the HiPER CPO project, Innolume is developing laser sources designed to meet the critical requirements of modern data centers. The development focuses on achieving: 🔹 high optical output power; 🔹 an optimized optical spectrum; 🔹 high reliability; 🔹 energy-efficient operation; 🔹 cost-effective volume manufacturing. These characteristics are essential for enabling the next generation of high-speed optical interconnects. 𝐐𝐮𝐚𝐧𝐭𝐮𝐦 𝐃𝐨𝐭 𝐓𝐞𝐜𝐡𝐧𝐨𝐥𝐨𝐠𝐲 𝐚𝐭 𝐭𝐡𝐞 𝐂𝐨𝐫𝐞 𝐨𝐟 𝐭𝐡𝐞 𝐏𝐫𝐨𝐣𝐞𝐜𝐭 The project is based on Quantum Dot technology, using quantum dots incorporated into semiconductor structures grown by Molecular Beam Epitaxy (MBE). HiPER CPO leverages the unique properties of Quantum Dots to develop semiconductor lasers that combine: 🔹 high energy efficiency; 🔹 high reliability; 🔹 suitability for scalable volume production. 𝐅𝐫𝐨𝐦 𝐒𝐢𝐦𝐮𝐥𝐚𝐭𝐢𝐨𝐧 𝐭𝐨 𝐄𝐱𝐩𝐞𝐫𝐢𝐦𝐞𝐧𝐭𝐚𝐥 𝐕𝐚𝐥𝐢𝐝𝐚𝐭𝐢𝐨𝐧 A key feature of the project is its comprehensive development approach. All technological process steps are modeled before fabrication, allowing device designs to be optimized prior to experimental validation. The developed structures are then experimentally evaluated to verify their performance. This approach enables more efficient development of advanced semiconductor lasers while reducing the overall development cycle. 𝐀𝐛𝐨𝐮𝐭 𝐭𝐡𝐞 𝐇𝐢𝐏𝐄𝐑 𝐂𝐏𝐎 𝐏𝐫𝐨𝐣𝐞𝐜𝐭 HiPER CPO develops semiconductor laser technologies for next-generation high-speed optical communication systems entirely in-house, leveraging its own industrial and research capabilities. Through its role as the project's coordinating partner and its expertise in Quantum Dot laser technology, Innolume contributes to the advancement of Co-Packaged Optics, one of the key technologies shaping the future of optical communication infrastructure. #CoPackagedOptics #CPO #QuantumDot #SemiconductorLasers #Photonics #OpticalInterconnects #SiliconPhotonics #DataCenters

⚡ The OCP Solid State Transformer (SST) Specification is officially public! In a major milestone for next-gen power infrastructure, Google, Microsoft, and NVIDIA have collaborated to establish a unified specification framework to accelerate our transition to Low Voltage Direct Current (LVDC) native data centers. As AI workloads scale to gigawatt densities, traditional power chains are hitting physical limits. SSTs solve this by stepping medium-voltage utility AC directly down to 800V DC distribution and bypassing bulky legacy conversion stages, improving grid to chip power efficiency, and dramatically shrinking the equipment footprint. We’re excited to drive this open standard with the community to scale AI systems with maximum velocity and efficiency. Access the official OCP SST Design Specification: https://lnkd.in/gb8u5f6m Big shout out to all the authors and contributors that enabled this: Jing Wang, Sangyoun “SY” Kim, P.E., Satarupa Bal, Ashish Upreti, Dragan Ristanovic, Kei Hao, Christina Peabody, Krishnanjan Gubba Ravikumar, Ehsan Nasr, Pritesh Patel, Xianyong Feng, Mike Tu, P.E. and many more who have dedicated countless hours working on this technology, and especially all of the technology partners across the industry that are working to make this technology into a scalable data center product. #OpenComputeProject

Excited to become a member of the Advanced Photonics Coalition and to participate in the CPO working group. We look forward to working with the APC community and contributing to the future of optical networking for AI infrastructure. See you at PECC in Sunnyvale this October. https://resolight.ai #AI #CPO #Photonics #OpticalNetworking #Energy #DataCenters Mari Mineta Clapp

The preliminary program is now available for the PECC Summit – Optica + Advanced Photonics Coalition Photonic-Enabled Cloud Computing Summit. Explore the expanded two-day agenda featuring eight technical sessions on the technologies shaping the future of AI infrastructure. Sessions will cover AI infrastructure architectures, optical scale-up and scale-out networking, AI fabrics, optical switching, advanced photonic devices and components, and the manufacturing innovations required to bring these technologies to volume deployment. The program brings together leaders from across the AI infrastructure and photonics ecosystem, including speakers from Arista Networks, AMD, Google, Meta, Microsoft, Oracle, Coherent Corp., Lumentum, ams OSRAM, Ayar Labs, GlobalFoundries, and many more. View the full program and register today: https://bit.ly/4x5nLKa 20–21 October 2026 Hewlett Packard Enterprise Aspiration Dome, Sunnyvale, California #PECCSummit #AIInfrastructure #Photonics #OpticalNetworking Thank you to our Platinum sponsors: Hewlett Packard Enterprise, Hitachi, Opsydia, Quantifi Photonics, Ciena, and Resolight.ai

𝐒𝐩𝐨𝐧𝐬𝐨𝐫 𝐓𝐚𝐥𝐤: Ravi Mahatme (Marvell Technology) "A Shared Memory Architecture for Token-Efficient AI Infrastructure" (register: https://lnkd.in/dsVsUr8B) We are excited to feature a sponsor talk from Ravi Mahatme, Senior Director, Product Management, Photonic Fabric Business Unit at Marvell Technology. 𝐀𝐛𝐬𝐭𝐫𝐚𝐜𝐭: As AI inference scales, infrastructure performance is increasingly constrained by the ability to move data efficiently between compute and memory rather than by compute alone. Traditional server architectures, built around fixed memory-to-compute ratios, limit memory capacity, infrastructure utilization, and overall AI efficiency. This presentation introduces a shared memory architecture that enables memory to scale independently of compute while maintaining the low latency and high bandwidth required for modern AI workloads. It examines the architectural principles behind a new shared memory tier, the role of optical connectivity in extending memory across servers and racks, and the system-level considerations required to deliver near-local memory performance at pod scale. Attendees will gain insight into how this architectural approach improves infrastructure utilization, reduces unnecessary data movement and recomputation, and enables more token-efficient AI systems. The session also explores how advances in optical interconnect technology are making shared memory practical for next-generation AI infrastructure. 𝐁𝐢𝐨𝐠𝐫𝐚𝐩𝐡𝐲: Ravi Mahatme is a product and systems leader focused on building next-generation AI infrastructure at the intersection of silicon, photonics, systems, and software. He joined Marvell Technology through its acquisition of Celestial AI and is currently Senior Director of Product, where he leads product strategy for rack-scale AI memory and interconnect platforms based on Photonic Fabric™ optical technology. Prior to Celestial AI, Ravi worked on distributed training and profiling capabilities for large-scale GPU workloads at AWS SageMaker. At Arm, he led product management for Ethos NPUs and spent more than a decade across engineering, business development, and product leadership roles, helping bring processor, accelerator, and platform technologies from concept to deployment. Ravi holds an M.S. in Computer Engineering from North Carolina State University and a B.E. in Electronics Engineering from the University of Mumbai. Don’t miss this opportunity to learn how shared memory architectures and optical interconnects are enabling more token-efficient AI infrastructure. Join us at IEEE Hot Interconnects 2026 to explore the technologies shaping the future of AI networking. Follow us on LinkedIn for updates on IEEE Hot Interconnects. IEEE Computer Society #hoti #hoti26

From a prototype to 10,000 units or more: what changes? In integrated #photonics, getting a prototype to work is one thing. But getting that same device manufactured thousands of times, with consistent performance, is a completely different challenge. 𝐏𝐫𝐨𝐭𝐨𝐭𝐲𝐩𝐞 𝐩𝐡𝐚𝐬𝐞: • Prove the concept • Maximize performance • Iterate quickly • Manual processes may be favored 𝐒𝐜𝐚𝐥𝐞-𝐮𝐩 𝐩𝐡𝐚𝐬𝐞: • Optimize manufacturing yield, throughput, and repeatability • Freeze the design • Reduce process variations • Automate wherever possible In light of these shifting priorities, a packaging approach that is perfectly acceptable for 10 devices may become a bottleneck at 1,000. That's why #manufacturability should be part of the conversation from the very beginning, such that all design parameters are optimized in parallel with the total costs. With expertise spanning across engineering and manufacturing, PHIX can help you define a production roadmap that ensures success in each product development stage. To get started, have a look at our free PIC Design Guidelines document. It will help you make some very smart PIC design choices early on! ➤ Read here: https://lnkd.in/eVYv23cK #chipindustry #productdevelopment #siliconphotonics

As the Chair of the D1 technical subcommittee of OFC2027, I want to encourage you to share with the world the great advancements of optical technology you have obtained in your labs. There's no better place and no better time to do that than at the OFC!

Luceda webinar series, 𝘽𝙚𝙮𝙤𝙣𝙙 𝙎𝙞𝙡𝙞𝙘𝙤𝙣: 𝘼𝙙𝙫𝙖𝙣𝙘𝙚𝙙 𝙈𝙖𝙩𝙚𝙧𝙞𝙖𝙡𝙨 𝙛𝙤𝙧 𝙈𝙤𝙙𝙪𝙡𝙖𝙩𝙞𝙤𝙣, continues! 🚀Join us for the second session, where we're excited to welcome OpenLight, a pioneer in heterogeneous silicon photonics. 💡 𝗜𝗻 𝘁𝗵𝗶𝘀 𝘀𝗲𝘀𝘀𝗶𝗼𝗻, OpenLight will introduce its heterogeneous III-V-on-silicon platform, which integrates active components, including lasers, optical amplifiers, and high-speed modulators, directly into silicon photonic integrated circuits. Discover how heterogeneous integration is enabling more scalable, energy-efficient, and higher-performance photonic solutions for next-generation AI and data center optical interconnects. 📆 𝗔𝘂𝗴𝘂𝘀𝘁 𝟭𝟮 | 𝟭𝟯:𝟬𝟬 𝗘𝗧 🎤 𝗦𝗽𝗲𝗮𝗸𝗲𝗿: Dr. Steven Alleston - Senior Director Business Development at OpenLight Photonics 𝗠𝗼𝗱𝗲𝗿𝗮𝘁𝗼𝗿𝘀:  Dr. Matthew Delaney - Business Development Manager – North America at Luceda Photonics Dr. Kevin McComber, CEO & Co-founder at Spark Photonics Design 👉 𝗥𝗲𝗴𝗶𝘀𝘁𝗲𝗿 𝗵𝗲𝗿𝗲: https://lnkd.in/eWBMccgA Register now and learn more about how heterogeneous integration is shaping the future of silicon photonics. #LucedaPhotonics #LucedaWebinars #OpenLight #SparkPhotonics #SiliconPhotonics #HeterogeneousIntegration #AI #DataCenter

Co-Packaged Optics scaling starts with the connections. As optical engines move closer to the ASIC, the conversation around Co-Packaged Optics is expanding beyond bandwidth and power. The next challenge is building systems that can be manufactured, integrated, and maintained at scale. Optical connectivity and fibre management are becoming critical design considerations—impacting density, alignment, routing, reliability, and serviceability. Our latest technical blog explores the role of the optical layer in moving Co-Packaged Optics from innovation toward deployment. Read more: https://lnkd.in/gggasERB SENKO We Value Every Connection #CoPackagedOptics #CPO #AIInfrastructure #DataCenters #FiberOptics #OpticalConnectivity #SiliconPhotonics