Scintil Photonics, a French deep-tech firm specializing in heterogeneous built-in photonics, has raised $58 million (€50 million) in a Collection B spherical to speed up growth and commercialization of its optical applied sciences designed for AI infrastructure.
The funding spherical was led by Yotta Capital Companions and NGP Capital, with further backing from NVIDIA, BNP Paribas Développement, Supernova Make investments, Bpifrance Digital Enterprise, Innovacom, Bosch Ventures, and Utilized Ventures ITIC Innovation Fund. The funding underscores robust market confidence in Scintil’s capacity to deal with the bandwidth, latency, and effectivity challenges on the core of recent AI information facilities.
The corporate’s expertise revolves round its SHIP platform – Scintil Heterogeneous Integration Photonics – which integrates a number of optical gadgets, together with lasers and modulators, onto a single silicon chip. This method consolidates dozens of historically separate elements, delivering increased density and decrease energy consumption. On the coronary heart of Scintil’s product roadmap is its LEAF Mild engine, the primary trade single-chip dense wavelength division multiplexing (DWDM) mild engine constructed for next-generation co-packaged optics. By providing multi-wavelength laser integration on a silicon photonics circuit, the answer helps the large interconnect wants of GPU clusters that underpin AI coaching and inference.
Scintil argues that its single-chip DWDM approach considerably reduces power use, reducing the associated fee per bit whereas additionally serving to AI infrastructure suppliers scale back their carbon footprint. In real-world phrases, LEAF Mild delivers 6.4 Tbps/mm edge bandwidth density at about one-sixth the facility consumption of standard pluggable optical modules. This positions the expertise as a spine answer for the optical materials required by AI factories, information facilities designed particularly for large-scale GPU workloads.
The brand new funding would enable Scintil to scale hiring each in its Grenoble headquarters and internationally, together with a brand new U.S. presence to work extra carefully with hyperscale and enterprise AI companions. Grenoble’s semiconductor and photonics ecosystem, together with shut ties with the CEA-Leti analysis institute, has given the corporate a robust basis in semiconductor integration and entry to superior expertise. Increasing into the USA is seen as a crucial step to deepen relationships with cloud suppliers, chipmakers, and information middle operators searching for scalable and sustainable optical interconnect options.
AI Networking Market
Business observers view Scintil’s progress as well timed. Analysts on the 650 Group forecast that the AI networking market may surpass $100 billion yearly by the top of the last decade, with co-packaged optics and exterior DWDM mild sources changing into important to assist the explosion of GPUs and AI bandwidth necessities. Equally, specialists at Omdia have famous that DWDM-based options can be very important for connecting 1000’s of GPUs inside a single information middle cluster. Scintil’s manufacturable, foundry-aligned platform is taken into account by many as one of many few commercially viable choices that may ship such scale.
Traders highlighted each the technical promise and the broader strategic relevance of Scintil’s work. Yotta Capital Companions’ Vincent Deltrieu described the corporate as an “innovation chief combining superior manufacturing and power effectivity,” whereas NGP Capital’s Bo Ilsoe pointed to its potential as a basis for international AI infrastructure. NVIDIA’s participation provides additional validation, because the GPU chief itself drives demand for next-generation optical interconnects.
For CEO Matt Crowley and CTO Sylvie Menezo, the Collection B spherical marks a pivot from growth towards international deployment. Crowley emphasised the effectivity positive factors and sustainability advantages, whereas Menezo underscored the distinctive worth of Scintil’s single-chip DWDM integration in enabling AI information facilities to scale reliably. With LEAF Mild coming into high-volume manufacturing, Scintil is positioning itself not simply as a provider of elements however as a central enabler of the optical backbones that may outline AI infrastructure over the approaching decade.
By marrying scalability, power effectivity, and manufacturability, Scintil Photonics is betting that built-in photonics will change into as indispensable to AI information facilities as GPUs themselves. The most recent funding offers the assets to fulfill surging demand and to solidify its position within the international race to construct the infrastructure powering AI at scale.
