Solinide raises €4m to turn microcomb photonics into AI data-centre hardware

Swedish deep-technology company Solinide Photonics AB has raised €4 million in seed financing to commercialize silicon nitride photonic integrated circuits and prepare its microcomb technology for scalable manufacturing, targeting the rapidly growing optical interconnect requirements of artificial intelligence data centres. Navigare Ventures and PSV Hafnium co-led the round, with Chalmers Ventures, Turbine Capital, Norrsken Evolve and Almi Greentech Fund also participating. Solinide plans to use the capital to expand engineering and commercial teams, strengthen internal prototyping, improve European manufacturing readiness and deepen work with strategic industry partners. The financing is modest compared with the billion-dollar capital flows surrounding AI compute, but its significance lies in whether a university spin-out can convert a technically demonstrated photonics platform into hardware that data-centre operators and equipment manufacturers will deploy at scale.

Solinide was founded in 2021 as a spin-out from Chalmers University of Technology and was previously known as Iloomina. The company develops optical microcombs, which can generate multiple precise wavelengths of light from one integrated source rather than requiring an individual laser for every wavelength channel. Solinide says this architecture can reduce the electricity consumption, physical footprint and cost associated with moving increasingly large quantities of data between computing systems.

Why has moving data between AI chips become almost as important as the chips themselves?

Artificial intelligence performance depends on more than the speed of individual accelerators. Large training and inference systems connect enormous numbers of processors, memory devices and switches, meaning information must continually move between components at extremely high bandwidth. When communication links cannot keep pace, expensive computing hardware spends more time waiting for data and less time performing useful calculations.

Electrical interconnects become increasingly difficult to scale over longer distances and at higher data rates because signal integrity and energy consumption deteriorate. Optical connections use light to carry information and can move larger volumes of data more efficiently across greater distances, which is why photonics is becoming an increasingly important part of AI data-centre architecture.

The challenge is that conventional optical systems can require many separate laser sources. Adding more wavelength channels can therefore increase the number of components, electricity consumption, thermal load and packaging complexity. Solinide is trying to attack that scaling problem by producing many optical wavelengths from one integrated microcomb source.

How does Solinide’s microcomb replace racks of individual lasers?

A microcomb creates multiple stable optical frequencies, often described as different colours of light, from a single photonic device. Each wavelength can carry a separate data stream through the same fibre using wavelength-division multiplexing, allowing aggregate bandwidth to increase without installing a separate physical fibre for every channel.

Solinide says its silicon nitride chip can produce dozens of these wavelengths from one source, effectively replacing numerous discrete lasers. The company is targeting co-packaged optics, where optical components are brought much closer to computing and networking silicon to reduce the energy required to move data between chips and switches.

The advantage is potentially multiplicative. Reducing the number of separate laser components can lower physical footprint and packaging complexity, while generating multiple channels from one source can make each fibre carry more aggregate data. The commercial challenge is maintaining stability, reliability and manufacturing yield at the standards hyperscale data centres demand.

Solinide reported in December 2025 that its rack-mounted silicon nitride microcomb system had achieved optical conversion efficiency above 60% and supported 28 wavelength channels. Those are company-reported performance figures rather than independent production benchmarks, and Solinide still needs to demonstrate that the technology can preserve its performance and reliability as manufacturing volumes increase.

Why is €4 million enough to matter but nowhere near enough to guarantee commercialization?

Photonics startups often face a difficult funding gap between laboratory demonstrations and commercial hardware. A design can work technically while still requiring extensive engineering around packaging, manufacturing yield, reliability testing, customer qualification and supply-chain integration before a hyperscaler or networking vendor will use it in production.

Solinide says the €4 million round will support technology readiness, commercialization, prototyping, manufacturing preparation and customer engagement. Those are precisely the steps required to move a photonics technology toward deployment, but each can consume substantial time and capital.

The seed round therefore buys Solinide an opportunity to establish the next set of milestones rather than completing the company’s journey. Success would likely create a need for larger later-stage financing to support manufacturing scale, inventory, qualification programmes and commercial expansion.

What makes the current round more interesting is that management says it already has an integrated rack-mounted product and is preparing for commercial launch and real-world deployment. Solinide also stated in December that trial integration with NVIDIA was underway, although the company has not disclosed a commercial supply agreement, deployment volume or revenue associated with that work.

Could Solinide help Europe build a larger position in the AI photonics supply chain?

Much of the advanced AI computing and data-centre networking ecosystem is concentrated around companies and manufacturing supply chains in the United States and Asia. Solinide is positioning its European manufacturing-readiness work as part of an effort to retain more photonics capability within the region. The funding announcement specifically identifies European manufacturing preparation as a use of proceeds.

Silicon nitride is attractive for photonics because it can support low-loss optical waveguides and can be manufactured using processes related to the broader semiconductor ecosystem. Solinide also operates a multi-project-wafer service intended to give photonics developers access to shared fabrication runs, adding a potential platform component alongside its own microcomb products.

A successful commercial product could consequently create value beyond one startup. It could contribute intellectual property, engineering capability and manufacturing demand to Europe’s photonics ecosystem while giving equipment developers another source of optical technology.

Strategic relevance does not substitute for commercial competitiveness, however. Data-centre customers will buy components based on performance, reliability, cost, integration and supply assurance rather than geography alone. European industrial-policy value can support investment, but customers still need a technical reason to choose the product.

What are the most important commercial proof points after Solinide’s €4 million funding round?

The first milestone is a product launch with clear specifications rather than another laboratory demonstration. Solinide’s management has indicated that product news is expected as the company moves toward commercial readiness, making packaging, form factor and target applications important indicators of how close the technology is to deployment.

The second is customer qualification. A trial integration or engineering evaluation becomes much more meaningful when it progresses into a paid design win, supply agreement or production deployment. Data-centre hardware has long qualification cycles because failure rates acceptable in a prototype are unacceptable when components are installed across thousands of servers or network switches.

Manufacturing yield and reliability will be equally important. A photonic chip can show excellent efficiency in a controlled demonstration but still become commercially uneconomic if too few devices meet specification at production scale. Solinide will therefore need to prove that its silicon nitride process and packaging architecture can be reproduced consistently.

The fourth measure is whether customers value the system-level savings enough to change existing optical architectures. If one microcomb source genuinely replaces many lasers while maintaining performance and reliability, its impact could extend beyond component cost to power consumption, cooling requirements and physical density across AI networks.

Solinide’s €4 million raise is small beside the enormous investments flowing into GPUs and data-centre campuses, but that imbalance is precisely why the story is worth watching. The next constraints in artificial intelligence infrastructure may not be solved simply by buying more accelerators. Moving information between those accelerators with less power and fewer components is becoming its own engineering race, and Solinide is now funded to find out whether its microcomb architecture can move from an impressive photonics result into that commercial supply chain.