Scintil Photonics raises $58 million to scale AI chip tech

Scintil Photonics has secured $58 million (€50M) in a Series B funding round led by Yotta Capital Partners and NGP Capital.

It included new investment from NVIDIA and BNP Paribas Développement, as well as ongoing support from existing investors including Supernova Invest, Bpifrance Digital Venture, Innovacom, Bosch Ventures, and AVITIC.

The funding will enable the company to expand its workforce in France and internationally, including the United States, and to increase production capacity as it advances the deployment of its integrated photonics platform.

The platform, known as SHIP (Scintil Heterogeneous Integration Photonics), integrates laser light sources directly onto silicon photonic circuits, allowing each optical fibre to carry multiple data channels, a technique known as Dense Wavelength Division Multiplexing (DWDM).

Scaling for AI growth

Scintil's technology is targeted at meeting the bandwidth, efficiency, and density requirements for next-generation AI infrastructure, aiming to serve large GPU clusters in data centres. Its LEAF Light product, based on SHIP, is entering high-volume production and is designed to support the increasing interconnect demands of AI data centres by enabling more data to be transmitted with reduced power consumption and thermal output.

Matt Crowley, Chief Executive Officer of Scintil Photonics, said, "This investment marks a pivotal moment for Scintil as we move to full-scale deployment. Our SHIP technology enables integrated photonic solutions with the scalability, energy efficiency, and integration density required to power next-generation compute infrastructure."

"This efficiency not only reduces data centre operating costs but also contributes to lowering the carbon footprint of AI infrastructure. With LEAF Light entering high-volume production, we're expanding from our base in Grenoble into the international markets, including the U.S., to support the world's most advanced AI factories."

Sylvie Menezo, Founder and Chief Technology Officer at Scintil Photonics, explained the technical collaboration behind LEAF Light.

She said, "We developed our LEAF Light integrated circuit in close collaboration with our customers. Used as an external laser source for Co-Packaged Optics (CPO) transmissions, it is a key component for the next generation of AI data centres. Its uniqueness lies in a single-chip solution that monolithically integrates precisely spaced DFB lasers on silicon photonic circuits, produced through a robust commercial supply chain. Our SHIP technology provides a strong and disruptive foundation to expand our portfolio of integrated photonic circuit solutions for the coming decades."

Investor perspectives

The Series B round was led by Yotta Capital Partners and NGP Capital, with participation from new and existing backers. Vincent Deltrieu, Managing Partner at Yotta Capital Partners, commented, "Scintil exemplifies the kind of innovation leaders we look for, combining advanced manufacturing, deep-tech leadership, and meaningful impact on the energy demands of AI infrastructure. Scintil's integrated photonics platform is essential to scale the next generation of AI factories. We're excited to support their global growth as they move to high-volume shipments."

Bo Ilsoe, Managing Partner at NGP Capital, added, "Integrated photonics is becoming a foundation of all AI infrastructure, and Scintil is turning that future into reality. Their technology delivers the bandwidth density and energy efficiency AI factories require with global scalability. We're excited to support Scintil as they scale deployments and become a leading player in building the next wave of compute and data infrastructure."

Industry outlook

Industry analysts see integrated photonics and co-packaged optics as critical to the evolution of data centre and AI computing infrastructure. Alan Weckel, Co-Founder and Technology Analyst at 650 Group, said, "As the number of XPUs and AI bandwidth grows, scale-up fabrics will need to transition away from copper to optical-based solutions. AI data centres will move toward co-packaged optics and external DWDM light sources.

Scale-up networking is an incremental opportunity for vendors, with the market set to exceed $25 billion by 2029. By the end of the decade, the whole AI Networking opportunity will approach $100 billion per year. Scintil stands out for bringing a manufacturable, foundry-aligned platform into the mix. Reliability and quickly scaling to volume are the only way hyperscalers can scale to the AI demand they plan to deploy this decade."

Daryl Inniss, Thought Lead of Optical Components and Advanced Fibre at Omdia, noted, "Dense Wavelength Division Multiplexing (DWDM) co-packaged optics is a differentiator that supports connecting thousands of GPUs at the bandwidth density requirements of next-gen AI infrastructure. Scintil's single-chip DWDM laser source, based on a heterogeneous integrated photonics process, demonstrates how this can be delivered in a manufacturable and scalable platform. This capability is important as AI factories push toward larger clusters and higher throughput."

Technology focus

Scintil's proprietary SHIP process enables the integration of multiple optical devices, such as lasers, photodiodes, and modulators, onto a single chip, replacing several traditionally discrete components. This integration is achieved using commercial foundry manufacturing, which the company states is capable of scaling to higher volumes required by large clients. The monolithic approach is intended to offer both performance improvements and facilitate cost-effective mass production.

The company's LEAF Light product is based on this platform and designed to deliver 6.4 terabits per second per millimetre (Tbps/mm) in edge bandwidth density while using roughly one-sixth the power of conventional solutions. The technology is aimed at scaling up GPU clusters and supporting emerging AI workloads.

Expansion and hiring

The new investment will allow Scintil to recruit additional personnel in engineering, operations, and customer integration roles as it increases production to meet demand. In addition to further developing its base in Grenoble, the company plans to grow its presence in the United States to better serve hyperscale data centre clients and AI infrastructure partners.