Sometimes a story lands that feels like it marks the shift from theoretical promise to industrial inevitability, and this one sits squarely in that zone. NcodiN, a deep-tech startup working on optical interposer technology with integrated nanolasers, has closed an oversubscribed €16 million Seed round—serious capital for a company that, until now, was still primarily in the R&D phase. The new funding is a bridge from research into industrial scale: product development, hiring specialized engineering talent, validating its CMOS production on 300 mm wafers, and shaping the supply chain and ecosystem needed for commercialization. MIG Capital led the round, joined by Maverick Silicon, PhotonVentures, Verve Ventures, and returning investors Elaia, Earlybird, and OVNI—meaning people who already knew the tech wanted more exposure.
The core of NcodiN’s work is NConnect, a photonic interposer platform designed to overcome the physical and energy barriers of electrical data movement—what industry insiders now casually call the “copper wall.” Modern AI computing is hungry, and scaling large models is already straining energy budgets and architectural design. NcodiN’s approach places the world’s smallest fully integrated laser directly on silicon, enabling dense optical links and chip-to-chip bandwidth that electrical interconnects simply can’t match at scale. The premise feels simple but seismic: packaging supercomputer-level capability inside a single processor footprint without having to redesign existing architectures. The new financing will accelerate industrial pilot runs, validate compatibility with advanced packaging approaches, and lay the groundwork for large-scale deployment.
There’s confidence behind the announcement, almost a bit of swagger, because the company has already demonstrated proof points most startups in this space only dream of: fully working optical links on silicon, nanolasers operating below 0.1 pJ/bit, integrated detectors, and a standalone cleanroom serving as a playground for rapid prototyping alongside early industry collaborators. They’ve also stacked a heavyweight advisory board, including Eli Yablonovitch, a pioneer in photonic crystals; Gus Yeung, former ARM executive and Intel Foundry Services CTO; and Peter de Dobbelaere, one of the key engineering leaders behind Luxtera’s rise. Their language is blunt: this isn’t incremental improvement—this is the missing piece for wafer-scale and chiplet-based AI architectures.
Investor commentary drew a straight line to market context. Memory bandwidth has become a defining constraint in next-generation AI systems, and the current reliance on copper interconnects is already showing its limits as model complexity grows exponentially. NcodiN’s tech positions itself as the enabling infrastructure for the next generation of GPUs, AI accelerators, and hyperscale compute fabrics. With generative AI workloads expanding, specialized hardware proliferating, and data exploding across distributed environments, the timing seems almost uncannily aligned with what the market now demands. And with this funding, the company is also putting down roots in Silicon Valley while expanding its European industrial footprint—aligning technology leadership with the centers of semiconductor manufacturing and AI adoption.
There’s a sense here that the photonics landscape is entering a second act. For years, the focus was about getting light reliably on silicon. Now, the challenge has shifted to system-level integration for multi-chip architectures. NcodiN is positioning itself not just as another optical component innovator but as an infrastructure redefinition company—a platform that could reshape how AI hardware is built. If they execute, this won’t just be another funding announcement; it will be one of those turning points people reference later with casual hindsight: that was when the post-copper future started feeling real.
