It was surprising how far-removed Washington and political discussions appeared to be from this meeting of photonics experts from all around the world. Indeed, those who lost their travel grants on very short notice were missed (and colleagues who made it to San Francisco were alarmed). But even with the intense discussions on social media, politics ultimately became a secondary topic at Photonics West. Of course, there were great scientific presentations and lots (nearly 1400) company booths. But the most important thing was socializing. Catching up with old friends and making new ones was and still is the reason people invest considerable resources in travel and, for many companies, exhibitor booths. If there is one lesson that we learned during the pandemic, it is that a virtual meeting cannot replace a handshake. Meeting in person is essential to building trust, and trust is the foundation of any successful future relationship, whether in business or research. In times of rising political tensions it is trust between experts that allows them to continue working together even if they follow different political movements. “Semiconductorization” of photonics If we look at the most powerful trends in high tech, artificial intelligence and big data are top of the list. Within both, their energy consumption is probably the most prominent challenge. Integrated photonic solutions have been widely discussed to solve the problem. But for them to work in this context, they need to be integrated into semiconductor manufacturing processes. The European Photonics Industry Consortium (EPIC) has always demonstrated a deep understanding of the trends and challenges in the photonics industry. So, it is no surprise that its "Summit on Optimizing and Scaling Photonics Manufacturing" was perfectly timed to discuss supply chain issues for integrated photonics. The trend is clear, and Scott Jordan, head of photonics at Physik Instrumente (PI), summed it up with the programmatic term, "semiconductorization of photonics." He showed how the evolution of photonics is following a path that semiconductors took years ago. Minna Gallery played host to an evening jam session for Photonics West attendees. The event was one of several off-location events held during the week-long show. Courtesy of Andreas Thoss. What does this mean in 2025? First, it is about upscaling and process automatization in the production of photonics components. Here, alignment is a major cost driver — and a place for innovation in the near future. Parallelization (for example, in array alignment) is a major way to cost reduction, as Jordan showed. Advancing the ecosystem will be next, promoting automated testing and “first time right” manufacturing procedures. To my surprise, Jordan sees standards as a killer to innovation. I would see it as an enabler for substantial growth across an industry. There were additional exciting presentations in this summit, though the limited space here allows only one more to mention: Leo Y. Lin, business unit director at EXFO, introduced us to advanced testing. He referred to an estimate I have seen a lot lately: In electronics, 90% of the value is in the chip and 10% in testing, assembly, packaging (TAP). In photonics, it is 80% TAP and 20% chip manufacturing. Lin went through many details of testing for PIC production. His conclusions (with a bit of shortening) are: 1)The industry needs efficient testing to characterize and validate large batches of wafers. 2)Wafer-level edge-coupling and high-performance testing are critical to efficiently identify known good dies (KGD) and to ensure quality starting from the wafer level. 3)A single software interface provides an automated flow from CAD files to analysis 4)AI and machine learning can be used to predict and skip bad circuits Lin’s message aligns with recent developments from within industry. In the week before Photonics West, semiconductor giant GlobalFoundries announced that it would establish a $700 million silicon photonics facility in Upstate New York. Of the dollar amount, $575 million will be dedicated to its Advanced Packaging and Photonics Center, which will perform advanced assembly, and testing functions in addition to packaging. Though not one of the dozens of new products released by exhibiting companies around Photonics West, Waymo self-driving cars captured the attention of show attendees. Courtesy of Andreas Thoss. Optical compute is coming As the supply chain for integrated photonics is evolving, people are eagerly searching for new photonic solutions at the chip level. In recent years, some startups have focused on photonics processors. Just recently, activity has turned to photonic interconnects, aiming to widen the data transfer bottleneck between processing units. Against that trend, German startup Q.ANT promoted an entirely photonic processing unit. Q.ANT is funded by the TRUMPF group. The company went through a learning curve with quantum sensors, in which they nurtured idea toward product. The company also worked on quantum computers — but put this aside to do something that is much more needed. With an optical processing unit, Q.ANT wants to reduce the energy consumption in AI systems by a factor of 30 on chip level. Michael Förtsch, founder and CEO, goes all in. “We have a roadmap where we target to outperform GPUs for AI-specific tasks within two and a half years,” he said to me. The Q.ANT native processing unit, an analog computer, is an AI accelerator and will reduce energy consumption by up to a factor of 30. Q.ANT uses several different technologies, but in the core, it is thin-film lithium niobate. This could become the killer application for which most PICs researchers were looking. Photonics West 2025 welcomed more than 24,000 registered guests. The organization's AR/VR/MR show was again co-located in 2025. And SPIE additionally announced a VisionTech event to debut in 2026. This exhibit and technical program will showcase the photonics technologies enabling the imaging and vision markets. It too will be located at Moscone Center. Courtesy of Photonics Media. Fusion? Fusion! Much has been said about laser fusion at the last two Photonics West meetings. At this year's LASE Conference, Constantin Haefner presented his thoughts on the benchmarks that a roadmap to fusion should include. According to Haefner, bottlenecks (scientific, technical, and economic) need to be identified, supply chains built, and regulatory issues resolved. He said, “Where there are challenges, there are immense opportunities for innovation and success. The fusion power plant market will demand far more than the current global production capacity of optics and laser components can provide.” Following Haefner’s talk, I spoke with Ed Moses, a legend in laser fusion research and founder of the startup Longview Fusion Energy Systems. Moses drew up plans for a first fusion power plant more than a decade ago, and today he wants to update those plans in light of the recent success of the National Ignition Facility (NIF). He envisions 15 shots per second. In the target, lead would replace gold, and a plastic shell would replace the precious diamond sphere. The changes could bring the cost down to 25 cents per target. "We are the only company that has proof of the physics," is Moses’ central statement for his company, regarding the experiments at NIF. Based on this fact, he sees laser fusion within a decade. I am a great admirer of NIF, but I am a bit less optimistic. Nobody has yet shown a wall plug efficiency >1. Neither have we seen a diode-pumped laser system of the necessary size. There are more questions to be answered before the goal is met. The good news is that several projects have started to build necessary supply chains. In Germany, for example, a set of such projects — for example, to develop substantially more efficient and cheaper diode lasers — has earned governmental funding. No matter if fusion comes or goes, there will be good use for cheaper and more efficient diodes in numerous fields. Adjacent activities I could go on with many more topics. Literally thousands of people where presenting and exchanging new ideas in San Francisco. Many prizes have been awarded to upcoming or established talents. I am sure you may read about it elsewhere. Something I haven’t seen outside social media is the growing number of social activities outside the convention center. EPIC is again a frontrunner here. This year, EPIC debuted a bicycle tour on the Sunday before the main exhibition. I took part, and, believe it or not, some photonics luminaries started to generate new ideas in an Uber packed with six people on the way back to the conference. It truly was networking at its best. On Wednesday night, partying became a challenge. First, I hit the Optica cocktail hour, which served as a good warm-up for the Prism Award ceremony that followed. This "see and be seen" event remains the most glamorous photonics celebration. And from there, it was easy to meet up with some from the German community on their way to the German Evening, held this year on the premises of Germany's most valuable public company, SAP. The celebration is a good place to meet people I don't get to see at home. My party discovery of the year, however, is the jam session of about a dozen specially gifted photonics people performing at the Minna Gallery. I counted 10 on the stage, all performing songs from the good old days — live (and very professionally). Everybody, including one well-known German CEO, took part in the dancing. The laid-back event is certainly much needed during a busy week-long show. I absolutely loved it! Why am I mentioning these activities? Because they are the ultimate means of creating and strengthening trust. In my humble opinion, this is precious and necessary, especially in these times. With free global trade and cooperation under political pressure, maintaining trust across borders seems like a valuable idea. You may not share the same political opinion, but that fades into the background when people rock together. th@thoss-media.com