Senior / Principal Silicon Photonic Packaging & Process Engineer

Voyant targets AI data centres with SiPh. Tuesday 13th October 2026 The company is developing a chip-scale optical circuit switching engine based on technology industrialised for FMCW LiDAR. Voyant Photonics is developing a silicon photonics optical circuit switching (OCS) engine for AI data-centre networks, adapting its production-proven switching architecture originally developed for frequency-modulated continuous-wave (FMCW) LiDAR. The company plans to supply the compact switching engine to optical networking companies for integration into their own platforms, rather than developing complete networking systems. The architecture uses solid-state optical paths without moving switching elements and is designed to deliver switching times below five microseconds. Voyant said its technology has been engineered for trillions of switching cycles, with a focus on reliability, low optical loss and semiconductor manufacturability. The company is adapting the platform to meet data-centre requirements, including optical loss and crosstalk, fibre attachment, port density, packaging, testability and compatibility with network control systems. Optical circuit switching can enable dynamic reconfiguration of network connections between GPUs and other computing resources, potentially reducing the need for repeated optical-electrical-optical conversions in AI infrastructure. Voyant also sees potential for its silicon photonics architecture to support closer integration with co-packaged optics (CPO) as data-centre networking requirements evolve.

Voyant Photonics has silicon that will 'make LiDAR as common as cameras' At CES, all about circuits interviewed Voyant Photonics' CEO to learn about the company's vision for LiDAR's camera-like future with solid-state 4D FMCW sensing. Weeks before CES 2026, Voyant Photonics announced a new lineup of solid-state 4D FMCW LiDAR sensors. Voyant claims its silicon photonics architecture will do for LiDAR what CMOS image sensors did for cameras - marking a major transition point for the industry. LiDAR has spent more than a decade constrained by moving parts, optical alignment challenges, and cost structures that limited deployment to niche or premium systems. Voyant argues that the next phase will depend on chip-scale integration and software-defined operation. At CES, we spoke to Voyant Photonics' CEO, Clément Nouvel, to learn about the company's new LiDAR architecture, the state of solid-state LiDAR, and where the technology is heading. All about circuits contributor jake hertz (left) with Voyant Photonics CEO Clément Nouvel (right) at CES 2026. From time-of-flight to FMCW on silicon. Nouvel framed Voyant's work as part of a generational shift in LiDAR architecture. "The first generation was time-of-flight at 905 nm with spinning polygons," he said, describing systems that rely on pulsed lasers and mechanical scanning. "LiDAR 2.0 is FMCW. It's the same principle as radar. You send a continuous wave, and you get Doppler velocity directly." Voyant's approach implements FMCW LiDAR entirely on a silicon photonics platform. The chip integrates the laser source, amplifiers, transmit and receive paths, and coherent detection on a single photonic integrated circuit. Operating at 1550 nm allows higher eye-safe output power and improved range performance compared with 905-nm systems. More importantly, FMCW inherently measures radial velocity at every point. "Every point in the cloud already tells you if it's moving toward or away from the sensor," Nouvel said. "That changes perception completely." Voyant's Carbon sensors. Voyant's Carbon family is the company's first step toward high-volume production. Carbon sensors use single-axis, on-chip beam steering combined with a low-speed oscillating mirror for the second axis. That hybrid approach reduces mechanical complexity while allowing Voyant to ship complete sensors. The company's Carbon 30 option targets mid-range perception with up to 150 m of range, a 30° by 120° field of view, and sub-centimeter range precision in specific modes. The Carbon 60 module prioritizes wide field-of-view sensing, offering 60° by 90° coverage, a maximum range of 75 m, and point rates approaching 977,000 pps. Both variants support direct per-point velocity measurement up to 63 m/s. Voyant's Carbon module on display at CES 2026. "We froze the designs and moved from a technology company into a production company," Nouvel said. "We are launching this in higher-series production by the end of the year. That's a big step for Voyant." Helium and the move to true solid-state. The more consequential announcement is Helium, Voyant's fully solid-state LiDAR platform. Helium eliminates all moving parts, including MEMS mirrors and voice coils. Instead of a mirror, the chip uses a dense two-dimensional array of surface-emitting optical antennas with fully integrated 2D beam steering. A fixed lens and on-chip electronics complete the system. "This one has no mirror whatsoever," Nouvel said. "You have the die and the lens. It looks very much like a camera." Voyant designed Helium for medium-range sensing with a 75-m maximum range, roughly 0.3-cm range precision, and 0.7-cm/s velocity precision. The sensor supports up to 819,200 samples per second in a package measuring roughly 3 cm x 4 cm x 4 cm. Nouvel compared the architecture directly to the evolution of radar. "Radar used to cost thousands of dollars with spinning antennas," he said. "Once everything moved into silicon, radar became a ten-dollar device. We are doing the same thing with LiDAR." Why 4D changes perception. Voyant made it clear that 4D LiDAR delivers more than depth maps. FMCW enables direct Doppler measurements, adding semantic information to raw point clouds. "With time-of-flight, you need heavy computer vision to classify objects and then infer motion frame to frame," Nouvel explained. "Here, the velocity is baked into the data." A rendering of the Voyant Helium module. Image used courtesy of Voyant Photonics. He demonstrated micro-Doppler effects, in which different parts of a moving object exhibit different velocities. "If a pedestrian turns, part of the body moves toward the sensor and part moves away," he said. "You can capture intent earlier. That reduces compute and improves safety." "Humanoid robots at home will have to meet safety constraints similar to cars," he said. "Industrial vehicles in warehouses need the same precision. These environments will stay mixed with humans and robots for many years." Modular optics and an ecosystem model. Voyant also plans for Helium to support modular optics, giving OEMs the option to tailor the field of view and range through lens selection rather than redesigning the sensor. While resolution will stay defined by the silicon, optics will adapt the system to the application. "It's like having a Sony image sensor with different lenses," Nouvel said. Voyant plans to sell both complete sensors and reference modules. "There are too many verticals for us to address alone," he said. "Some customers want the module and will build their own sensor around it. Others want a finished product." Scaling toward ubiquity. Nouvel described the next year as an execution phase focused on manufacturing scale, environmental validation, and early access programs. Evaluation hardware is planned for early 2026. "We are step by step driving LiDAR toward what cameras are today," Nouvel said. "The value is in the chip. Everything around it becomes simpler, smaller, and cheaper." For Voyant, that direction is the main driver of its product roadmap and its view of when solid-state LiDAR will finally become pervasive.