Zurich, Switzerland - December 18, 2025 - Enlightra, a Swiss-American deep-tech startup focused on optical technologies for artificial intelligence infrastructure, has emerged from stealth with USD 15 million in funding to commercialise energy-efficient laser systems designed for next-generation AI data centres, as operators grapple with the growing power and bandwidth limits of traditional copper interconnects.
Founded by Maxim Karpov and John Jost, Enlightra is developing chip-scale multiwavelength “comb” lasers that enable optical data transmission at significantly lower energy consumption compared with electrical links.
The company’s technology is aimed at addressing one of the most pressing bottlenecks in AI infrastructure, the rising power and cooling costs associated with GPU-to-GPU and rack-to-rack communication in dense compute clusters.
The USD 15 million funding round was backed by a group of investors including Y Combinator, Runa Capital, Pegasus Tech Ventures, Protocol Labs, Halo Labs, Asymmetry Ventures, and TRAC VC. The capital will be used to advance product development, expand engineering capabilities, and prepare the technology for integration into AI-focused data-centre environments.
Unlike conventional approaches that rely on multiple discrete laser sources, Enlightra’s platform consolidates multiple optical wavelengths into a single integrated laser device, with each wavelength acting as an independent high-bandwidth data channel.
This design allows a single compact source to replace several lasers, reducing power draw, heat generation, and system complexity while enabling higher aggregate bandwidth over optical fibre.
According to the company, early prototypes support eight- and sixteen-channel configurations aligned with emerging AI interconnect requirements, and have demonstrated error-free data transmission at target speeds and power levels. The lasers are manufactured using industry-standard silicon photonics processes, positioning the technology for scalable production as demand for optical interconnects accelerates across hyperscale and enterprise AI deployments.
Commenting on the announcement, co-founder Maxim Karpov said the AI sector is “hitting the physical and economic limits of copper-based connectivity,” while co-CEO John Jost highlighted that leveraging multiple wavelengths allows optical networks to operate closer to their theoretical capacity without a proportional rise in energy use.
While Enlightra’s near-term focus is on in-data-centre AI interconnects, the company sees broader applications for its technology across data-centre optical links, long-haul networks, and future high-performance computing systems. Pilot production is expected later this decade as the company works with ecosystem partners to validate deployment pathways.
The announcement comes amid intensifying industry efforts to curb the energy footprint of AI infrastructure, with optical technologies increasingly viewed as critical to sustaining performance growth without unsustainable increases in power consumption and cooling demand.
