Metalenz buys Israel's PxE Imaging to add color, depth to AI vision
What's the deal? Metalenz has acquired PxE Imaging, an Israeli startup whose technology lets a standard image sensor capture color and depth simultaneously. Terms were not disclosed.
What each does: Boston-based Metalenz, spun out of Harvard University in 2016, builds metasurface optics patterned directly onto CMOS image sensors and says it has shipped more than 300 million of them. PxE, founded by chief executive officer Yoav Berlatzky, developed Holocoder technology that is likewise patterned onto the sensor.
Why it matters: Most cameras use a Bayer filter to produce color, but it absorbs about 60% of incoming light. Holocoder replaces that filter and sorts light without absorbing it, delivering 2.5 times more light to the sensor and improving performance in dark and bright scenes alike.
What's the endgame? The deal extends Metalenz beyond polarization and near-infrared sensing into visible imaging. By decoding color and depth in hardware rather than reconstructing them in software, the company aims to cut computational load so AI can run more efficiently on devices — from consumer electronics and automotive to drones and robotics.
What could go wrong? PxE's technology is not yet on the market. Metalenz points to consumer applications no earlier than late 2027 or early 2028, so the claimed performance still has to be turned into shipping products.
Metalenz brings a commercialization edge: its Polar ID face authentication system entered mass production at UMCDealroom has a profile for this one. Try Dealroom → last year, and it operates at semiconductor-fab scale. "With PxE's technology and team, we're expanding our capabilities into visible sensing, addressing a longstanding problem of low light performance," said Rob Devlin, Metalenz's co-founder and chief executive officer.
Berlatzky said joining Metalenz "gives us the manufacturing platform and commercial reach to bring this technology to everyday devices."
The signal: The acquisition reflects a broader push to give edge-AI systems richer input while reducing reliance on multiple cameras and heavy software reconstruction. Whether that translates into scale will depend on execution over the next two years.
Read more: globenewswire.com
Image credit: Zach Dischner
