Dutch photonics startup Photon Bridge raises €10M to tackle AI's looming laser bottleneck
Dutch photonics company Photon Bridge has raised €10 million to tackle an emerging bottleneck in AI infrastructure: producing the millions of lasers that will be needed to connect increasingly large ...
Dutch photonics startup Photon Bridge has secured €10 million in funding to address an emerging laser supply bottleneck in AI infrastructure. The investment round was led by Amadeus APEX Technology Fund and 360 Capital, with existing investors Innovation Industries, Faber, and Brabant Development Agency (BOM) also participating.
Omdia predicts that over 20 million AI data-center accelerators will be shipped annually by 2028. As AI clusters scale from thousands to hundreds of thousands of processors, inter-chip data transfer becomes a performance constraint. Optical connections, particularly co-packaged optics, offer faster and more efficient data transmission closer to processors as they integrate with chips.
However, this shift creates a challenge: the current high-power laser supply chain was designed for lower-volume telecom and subsea networks. AI could transform lasers into semiconductor-scale components requiring manufacturing at that scale. Photon Bridge's technology combines and multiplexes light from multiple lower-power semiconductor lasers, reducing manufacturing defects and enabling efficient scaling with increased volumes.
CEO Paul Marchal states that AI is transforming computing infrastructure, moving data transfer closer to processors. This requires optics to be integrated with compute. Photon Bridge's approach focuses on manufacturing yield, crucial as laser production moves from traditional telecom volumes to AI infrastructure scales. Investment partners Ion Hauer and Cesare Maifredi highlight the company's technology as addressing the manufacturing challenge by combining deep photonics expertise with scalability potential for millions of laser sources annually.
Written by urgent.news from Tech.eu's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.