On a factory floor in Japan, autonomous mobile robots glide past human workers, stopping abruptly when someone steps into their path. Nothing remarkable there—except these machines aren't relying on the laser scanners that have dominated industrial safety systems for decades. They're listening.
Sonair, a three-year-old Norwegian company barely known outside robotics circles, thinks it's found something the industry has been groping toward: a sensor that sees in three dimensions using sound waves, costs half what competitors charge, and might actually pass the safety certifications that separate serious industrial equipment from expensive science projects.
It's a bold claim in a sector where incumbents like SICK and Keyence have spent years refining their offerings. Whether Sonair can deliver—and whether manufacturers will trust ultrasound where they've long relied on lasers—remains an open question. But the early signals suggest at least some big players are willing to place their bets.
Sound Over Light
ADAR, as Sonair calls its sensor (acoustic detection and ranging, naturally), made its debut at Automate 2025 in Detroit last May. By September, units were shipping to customers across three continents. Three months later, in February, the company announced series production had begun.
The timing matters. Autonomous mobile robots have proliferated across warehouses and manufacturing floors over the past five years, driven by labor shortages and falling robot costs. But safety sensors remain stubbornly expensive—often eating up 30 percent or more of a robot's total parts cost. Any technology that promises to cut that figure draws attention.
What makes ADAR unusual isn't just its use of ultrasound—industrial ultrasonics have been around for ages. It's the marriage of 3D coverage with safety certification, a combination that has eluded other ultrasonic sensor makers. Traditional 2D laser scanners sweep a horizontal plane, which means they can miss forklifts with elevated pallets, overhanging obstacles, or low-lying objects that duck beneath the beam. ADAR's hemispherical field of view—180 degrees in both directions, out to four meters—theoretically catches all of that.
The sensor itself is compact: 200 grams, roughly the size of a deck of cards plus change. Inside, piezoelectric MEMS arrays generate ultrasonic pulses between 70 and 85 kilohertz, well above human hearing range. Software beamforming steers these acoustic beams electronically across the coverage area, eliminating the need for moving parts. Twenty times per second, ADAR spits out fresh data through a standard Ethernet connection.
Sonair claims the system can spot objects as small as 30 millimeters in diameter at distances up to three meters, with range accuracy within two centimeters. That's precise enough to stop a robot before it clips a worker's ankle—assuming the software and certification hold up.
The Certification Gauntlet
Here's where things get interesting, or at least more consequential. Industrial safety isn't a marketing exercise. Equipment operating in European factories must meet EN ISO 13849-1:2023 Performance Level d standards. Globally, manufacturers look for IEC 61508 SIL2 compliance. These certifications involve extensive testing, documentation, and third-party verification. They're expensive and time-consuming to pursue, which is why many sensor startups avoid them.
Sonair hasn't. The company says ADAR's certification process, conducted with safety assessment firm exida, reached its "final stage" by February 2026. Completion is planned for sometime early this year, though anyone familiar with certification timelines knows those dates can slip.
To streamline the process, Sonair built ADAR's firmware using Ferrocene, a safety-certifiable Rust programming toolchain whose core library earned TÜV SÜD certification for IEC 61508 SIL2 in December 2025. The company integrated this certified toolchain by late December—a technical detail that matters mainly because it suggests Sonair is taking the certification seriously rather than treating it as an afterthought.
No other airborne 3D ultrasonic sensor appears to have crossed the functional safety threshold for robot protective stopping functions. Competitors like Germany's Toposens offer 3D ultrasonic sensing for collision avoidance, but without the published PLd or SIL2 certifications that unlock safety-critical applications. That gap, if Sonair bridges it, could prove meaningful.
Early Traction
FUJI Corporation, a Japanese manufacturer known for electronics assembly equipment, has already ordered ADAR units for an upcoming AMR product line. Koji Kawaguchi from the company's Innovation Promotion Department went on record endorsing the technology—the kind of public commitment that carries weight in conservative industrial circles. A Swiss autonomous cleaning robot maker also signed on, though Sonair hasn't disclosed that customer's name.
By October 2024, more than 20 companies across industrial automation, automotive technology, and healthcare robotics had joined Sonair's early adopter program. First shipments went out in July 2025, a couple months behind the May announcement but ahead of full production ramp.
The company lined up Cornes Technologies as its distribution partner in Japan, a strategic choice given that market's appetite for robotics and automation. NVIDIA added Sonair to its Inception Program last June—a stamp of approval that carries marketing value if not necessarily technical validation. ADAR also collected awards: "Best of Sensors 2024" in the Automotive/Autonomous category, a finalist spot in Automate's Startup Challenge, Germany's inVISION "Top Innovations 2026" designation.
Awards are cheap. Repeat orders matter more, and those numbers aren't public yet.
The Economics of Safety
Sonair's pitch hinges on cost as much as capability. The company claims ADAR can cut sensor package costs by up to 50 percent compared to conventional approaches. That figure presumably reflects ADAR replacing multiple 2D scanners or supplementing existing sensor arrays with fewer total units. Whether that math pencils out in real-world deployments remains to be seen—Sonair hasn't published list pricing, and cost comparisons depend heavily on specific robot configurations and safety requirements.
But if the economics work even roughly as advertised, manufacturers will pay attention. Robotics companies operate on tight margins. Shaving thousands of dollars off a mobile robot's bill of materials while maintaining safety compliance would be significant.
The technical advantages are easier to verify. Ultrasound works in conditions that bedevil optical sensors: complete darkness, heavy dust, high-glare environments. ADAR detects transparent materials like glass panels and highly reflective surfaces—both notorious blind spots for cameras and some laser scanners. The fixed-array design means no spinning components to wear out or maintain.
Multiple ADAR units can operate in the same space without interfering with each other, handled through built-in crosstalk algorithms. The technology draws on roughly two decades of ultrasonic research at SINTEF MiNaLab in Norway, the research institute from which Sonair spun out.
Following the Money
Investors seem convinced, at least at early-stage valuations. Sonair raised $6.8 million through October 2024, led by Skyfall Ventures and RunwayFBU. A $6 million round followed in September 2025, bringing in Scale Capital, Investinor, SINTEF, ProVenture, and TYR.vc. That's nearly $13 million total—modest by Silicon Valley standards but substantial for a Norwegian hardware startup still proving its technology.
The funding timeline suggests investor confidence grew as Sonair moved from prototype to production. Whether that confidence is justified depends on what happens next: certification completion, production scaling, customer adoption beyond early adopters. And whether ADAR actually delivers the cost savings and safety performance Sonair promises.
The Bigger Picture
Step back, and Sonair's story fits a broader pattern. As robots colonize more shared human spaces—factories, warehouses, hospitals, eventually public areas—the pressure to prove safety credentials intensifies. Regulators in Europe and elsewhere are tightening requirements. Insurance companies are paying attention. Manufacturers can't wave away safety concerns with vague assurances.
That creates an opening for technologies that genuinely improve safety while cutting costs. It also creates opportunities for clever marketing to run ahead of proven capability. Sonair seems to be making the right moves: pursuing certification, shipping actual products, landing name-brand customers. But the gap between promising technology and market-dominant product is littered with well-funded startups that couldn't close it.
The established players aren't standing still. SICK, Keyence, and Omron didn't build their market positions by ignoring new approaches. If 3D ultrasonic sensing proves valuable, they have the resources and distribution networks to catch up—or simply acquire promising startups.
For now, Sonair has a head start. Whether that translates to lasting advantage depends on execution, certification, and whether manufacturers ultimately trust sound waves to keep their workers safe around machines that still don't think quite like we do.
At LogiMAT in Stuttgart later this year, Sonair plans to announce a strategic AMR partnership. That reveal will offer another data point on whether this bet on ultrasound is gaining momentum—or whether the market remains skeptical of acoustics in a world dominated by light.
