How the chip works
The transmitter is built around a 5-by-5 grid of vertical-cavity surface-emitting lasers, or VCSELs, the same class of semiconductor laser already mass-produced for smartphone face-recognition sensors and data-center fiber links. Each VCSEL is paired with a micro-optic element that shapes its beam to cover a specific zone inside a room, much like a spotlight grid on a theater ceiling. During testing, 21 of the 25 lasers were active. Each one carried between 13 and 19 Gbps of data, and their combined output reached the 362.7 Gbps aggregate figure. Crucially, the researchers also showed the system splitting its capacity across several receivers at once, serving multiple users in parallel rather than funneling everything into a single link. The design philosophy is deliberate simplicity. Instead of electronically steering a single beam to track moving devices, the chip fires many fixed beams shaped to blanket distinct zones. That trades some flexibility for a compact, lower-power architecture that could eventually fit into a ceiling-mounted access point. “The VCSEL-array approach lets us scale bandwidth by adding emitters on the same chip rather than increasing the complexity of each channel,” the study’s authors wrote in the paper, noting that the architecture is designed to keep per-bit energy low even as aggregate throughput rises.Where it fits in the research landscape
This is not the only group pushing indoor optical wireless into triple-digit gigabit territory. A separate team recently demonstrated a 320 Gbps link using thin-film lithium niobate and electronic beam steering, published in Nature Communications. That system uses a phased-array approach to actively direct light toward targets, a fundamentally different architecture that offers more agility but greater complexity. Earlier work established that optical wireless could cover practical room-scale areas, not just narrow benchtop paths. A 2019 study in Nature Scientific Reports demonstrated a 35 Gbps visible-light link using ceiling lighting infrastructure for wide-area indoor coverage. That result was a notable benchmark at the time; more recent visible-light communication experiments have pushed LED-based links higher, but the 2019 study remains one of the most widely cited demonstrations of room-scale optical coverage. The new VCSEL result represents a roughly tenfold speed increase over that benchmark while using infrared lasers instead of visible LEDs. Harald Haas, a professor widely regarded as a pioneer of LiFi technology, has noted in public remarks that the field is advancing faster than many in the wireless industry expected. “We are seeing a convergence of photonics and communications that could reshape how buildings deliver bandwidth,” Haas observed at an IEEE photonics conference in early 2026. Standards bodies including the IEEE 802.11bb working group, which finalized the first LiFi standard in 2024, are watching these laboratory results closely as they consider next-generation specifications. The diversity of approaches is itself telling. Some groups are optimizing for compatibility with room lighting, others for agile beam steering, and the VCSEL-array team for dense parallel emitters and chip-level integration. Future commercial products may borrow elements from several of these designs rather than adopting any single blueprint.What remains uncertain
The 362.7 Gbps figure was recorded under controlled laboratory conditions. No field-trial data from real offices, classrooms, or homes has been published. That distinction matters because real rooms introduce problems a lab can minimize: people walking through beam paths, furniture blocking line of sight, sunlight flooding through windows, and the constant need for a reliable uplink channel from devices back to the transmitter. Interference from ambient light is a related concern. Fluorescent fixtures, LED panels, display screens, and outdoor glare can all inject noise into optical receivers. The published work characterizes signal quality under defined conditions but does not yet map performance across the wide variety of lighting environments found in actual buildings. Eye safety is another area where the public evidence is incomplete. The journal article discusses laser safety classification and accessible emission limits under the IEC 60825 framework, but no consumer safety certification for this specific VCSEL array has appeared in the available literature. Closing the gap between a lab demonstration that appears to meet safety limits on paper and a certified product that passes full regulatory review can take years, particularly when multiple beams and complex room geometries are involved. Cost and manufacturability are largely unaddressed. VCSELs themselves are inexpensive at volume thanks to their use in consumer electronics and data centers, but the integrated beam-shaping micro-optics and precise alignment requirements could complicate assembly. Without bill-of-materials estimates or yield data, it is hard to judge whether these transmitters would be affordable for mass-market access points or limited to specialized enterprise deployments. Scaling to the terabit-per-second rates that 6G roadmaps envision for indoor hotspots is also an open challenge. The new chip gets roughly a third of the way there from a single transmitter. Whether tiling multiple chips across a ceiling or combining them with phased-array steering can reach that threshold without exceeding energy and safety budgets is a design problem no published prototype has yet solved.What to watch for next
For consumers and IT planners, the practical takeaway is cautious optimism. The combination of speed and energy efficiency demonstrated here is real, peer-reviewed, and addresses two genuine pain points: rising bandwidth demand and tightening power budgets. Companies like pureLiFi and Signify have already begun shipping early LiFi products based on older, slower optical wireless technology, which suggests commercial interest in the broader category is not purely academic. But the path from a 25-laser chip on a lab bench to a ceiling-mounted unit in a conference room runs through safety certification, manufacturing scale-up, protocol standardization, and integration with the radio-frequency systems that will still handle mobility, uplink traffic, and wide-area coverage. Wi-Fi is not going away; optical wireless is more likely to complement it, handling ultra-high-bandwidth downlinks in fixed zones while radio covers everything else. The most useful signals to track from here: follow-up publications testing similar systems under realistic multiuser conditions, independent replication of comparable performance by other research groups, and early moves by standards bodies to define how optical wireless fits into future network specifications. Until those milestones arrive, 362.7 Gbps remains a compelling proof of concept, not a product roadmap. More from Morning Overview*This article was researched with the help of AI, with human editors creating the final content.
