NVIDIA-Backed iPronics Raises $125M to Rewire AI's Future with Light

VALENCIA, Spain & SANTA CLARA, Calif. – September 02, 2026 – In a move that signals a seismic shift in the architecture of artificial intelligence, iPronics, a pioneer in programmable optical networking, has secured $125 million in a Series B funding round. The investment, co-led by Maverick Silicon and Light Street Capital, notably includes the world's preeminent AI chipmaker, NVIDIA, as a strategic investor. The round, which brings the company’s total funding to a formidable $177 million, is a massive vote of confidence in a technology poised to solve one of the most critical bottlenecks holding back the next generation of AI: the network itself.

The capital infusion is earmarked to accelerate the commercial deployment of iPronics' flagship product, the iPronics Optical Networking Engine (ONE). This isn't just another incremental upgrade; it’s a fundamental rethinking of how data travels inside the massive data centers that power our AI-driven world. As Christian Dupont, CEO of iPronics, stated, “With this investment, we can significantly accelerate our commercialization, delivering the scale our customers need to drive broad adoption and ensuring data centers can fully utilize their GPUs without compromising performance.”

The AI Data Center's Achilles' Heel

The explosive growth of generative AI has created an insatiable demand for computational power, leading to the construction of sprawling AI factories filled with tens of thousands of GPUs. But a critical vulnerability has emerged in this gold rush. While GPUs have advanced at a blistering pace, the traditional electronic networks that lash them together are straining under the load. This networking layer has become the Achilles' heel of modern AI infrastructure.

Traditional networks, built on copper wiring and electronic switches, consume enormous amounts of power and introduce latency that can cripple the performance of large-scale AI training models. As AI clusters grow, the physical and electrical limitations become stark. Power bills soar, cooling challenges mount, and fixed network topologies mean expensive GPUs often sit idle, waiting for data. This inefficiency is untenable for hyperscalers who measure progress in fractions of a percentage point. Failures within these vast clusters are no longer rare events; they are a constant operational reality, making network resilience a paramount concern.

This is the problem iPronics was founded to solve. The industry is rapidly transitioning from copper to optics, and the company is positioning itself at the forefront of this change with a new generation of high-density switching solutions. As Manish Muthal, senior managing director at Maverick Silicon, put it, “The explosive growth of AI workloads has pushed traditional network infrastructure to its power and bandwidth limits, making optical circuit switching more critical than ever.”

A New Architecture Built on Light

At the heart of iPronics' innovation is silicon photonics—a technology that manufactures optical components directly onto silicon chips, leveraging the scale and precision of the semiconductor industry. The company's iPronics ONE platform is a rack-mounted Optical Circuit Switching (OCS) system that routes data using light instead of electrons.

Unlike older mechanical optical switches, iPronics' solution is entirely solid-state, with no moving parts, dramatically increasing reliability. It boasts near-zero latency (under 30 nanoseconds) and can reconfigure network pathways in less than 300 microseconds. This allows an AI cluster's network topology to be changed on the fly, dynamically allocating bandwidth to different workloads, whether for training a new foundation model or running millions of inference queries. The result is a dramatic increase in GPU utilization and a significant reduction in power consumption—by as much as 50% for the switching fabric alone.

Founded in 2019 as a spin-off from the renowned photonics lab at Universitat Politècnica de València, iPronics has spent years perfecting this technology. “This Series B is a testament to our team's relentless execution,” said Dr. Daniel Pérez-López, the company’s founder and CTO. What makes their approach particularly powerful is not just the hardware, but the software that controls it. The iPronics ONE comes with a full suite of APIs, enabling data center operators to program the optical layer and automate connectivity without needing specialized photonics expertise. This accessibility is key to translating a hardware breakthrough into tangible business value.

A Constellation of Silicon Valley All-Stars

The composition of this funding round reads like a who's who of the semiconductor and AI infrastructure world, underscoring its strategic importance. NVIDIA’s participation is the most telling endorsement. As the dominant provider of AI accelerators, NVIDIA has a vested interest in ensuring the surrounding infrastructure can keep pace. Its investment is a clear signal that programmable optical networking is not just a promising technology, but a necessary component for the future of AI.

The human capital joining iPronics is just as impressive. Manish Muthal of Maverick Silicon, who brings deep experience from Intel and Broadcom, now joins the board. He is joined by Geoffrey Tate, an advisor to co-lead investor Light Street Capital. Tate is a legend in the chip industry, having served as a senior VP at AMD, the founding CEO of Rambus, and more recently the CEO of an AI hardware firm acquired by Analog Devices. His presence on the board of Ayar Labs, another optical interconnect pioneer, highlights his deep belief in the technology's trajectory.

Adding to this brain trust is Young Sohn, who becomes an advisor to the company. A founding member of Catalight Capital and former President of Samsung, Sohn also sits on the boards of Arm and Cadence Design Systems. His perspective is clear: “Next-generation AI infrastructure will require a more dynamic, optical-first network fabric,” he stated, validating iPronics’ entire mission. Shef Osborn, a partner at Light Street Capital, echoed this sentiment, calling the company's platform “the most scalable and cost-effective approach to solid-state optical switching at the interconnect layer.”

From Lab to Hyperscaler

With $125 million in new capital, iPronics is shifting from development to full-scale commercialization. The company has already been shipping its ONE platform to “major AI infrastructure providers” for the past year, proving its technology in real-world environments. The funds will be used to scale its operations and manufacturing to meet what it describes as rapidly increasing market demand.

To spearhead this growth, iPronics has opened a new office in Santa Clara, California, placing it at the epicenter of the AI industry. The company is actively hiring for roles focused on customer deployment and partnerships, a clear sign that its focus is on market penetration and helping hyperscalers and enterprises integrate its technology into their next-generation AI clusters.

By creating a programmable, resilient, and power-efficient optical fabric, iPronics is not just selling a piece of hardware. It is offering a new architectural blueprint for AI data centers—one that promises to unlock greater performance, manage spiraling costs, and provide the dynamic flexibility needed to power the next decade of artificial intelligence. As Dr. Pérez-López affirmed, “Together with our tier-one partners we are successfully demonstrating that we can build and deploy this technology at scale for AI infrastructure.”