iPronics announced a $125 million Series B on September 2, 2026, co-led by Maverick Silicon and Light Street Capital. NVIDIA joined the round, along with Triatomic Capital, Bosch Ventures, Catalight Capital, the European Innovation Council Fund, The Tate Family Trust, Fine Structure Ventures, Amadeus Capital Partners, Build Collective, and Criteria Venture Tech. The raise takes the Universitat Politècnica de València spin-off to $177 million in total funding.

The company sells iPronics ONE, a rack-mounted optical circuit switch built on silicon photonics. The pitch is that AI clusters now hit power and bandwidth limits on copper scale-up fabrics, and that a programmable optical layer can reconfigure connectivity in real time for training and inference without rebuilding the network. The box ships with control, telemetry, and APIs so operators can treat the optical path as software.

Christian Dupont, the chief executive, said the money will speed commercial deployments so customers can keep GPUs busy without giving up performance. Manish Muthal of Maverick Silicon and Geoffrey Tate, an advisor to Light Street Capital, are joining the board. Young Sohn of Catalight Capital, formerly of Samsung and Inphi, becomes an advisor. iPronics said it has opened a Santa Clara office and is hiring there.

“This Series B is a testament to our team’s relentless execution. Together with our tier-one partners we are successfully demonstrating that we can build and deploy this technology at scale for AI infrastructure.” Daniel Pérez-López, founder and CTO, iPronics

Dupont’s companion note said the Series A was about 18 months earlier, and that the new capital is meant for operations, deployments, and a deeper Silicon Valley footprint rather than a research reboot.

Decoded Take

NVIDIA putting money into a Spanish optical-switch vendor is a tell that the next bottleneck after GPUs is the fabric that stitches them together. iPronics is selling a compact OCS, not another GPU, which is the right layer if hyperscalers want to reuse installed compute instead of overbuilding copper. The missing proof is a named production cluster. Board seats and a Santa Clara lease are cheaper than a public Microsoft or Amazon deployment. Watch whether iPronics ONE shows up in a disclosed rack count before the next round, and whether NVIDIA’s participation stays financial or turns into a qualified part on a reference AI factory design.