The Center for Data Innovation recently spoke with Daniel Pérez-López, co-founder of Ipronics, a company based in Spain developing programmable silicon photonics. Pérez-López explained how the company’s technology is transforming data center operations by accelerating data flows while reducing the energy costs of AI infrastructure.
David Kertai: What inspired the creation of iPronics?
Daniel Pérez-López: iPronics began as a university spin-out with the goal of turning research in programmable silicon photonics into practical solutions. Photonics is the use of light, rather than electricity, to move and process information. We wanted to bring to photonics the same breakthrough that transformed electronics: the ability to program a chip’s function after it’s manufactured. Within three years, we built the world’s first programmable photonic processor, which is essentially a piece of hardware that can control light signals in different ways. This opens an opportunity in the fast-growing market for data centers, which need faster, more flexible, and more energy-efficient ways to move information.
Kertai: How does your technology change how data moves in networks?
Pérez-López: Our technology brings three layers into one device. Light carries data at very high speed through tiny optical channels, electronics control and reconfigure those channels while also tracking performance, and software gives operators a simple way to manage flows. Together, these layers let networks move data more flexibly and efficiently, adjusting in real-time as demand changes.
Kertai: What advantages does this bring to AI data centers?
Pérez-López: The amount of data moving through AI centers, on average, doubles every two years, but traditional electronic switches can’t keep pace. They are tied to fixed data rates and use more power with every new generation. Our optical switches, by contrast, act like wide-open pipes. They handle any data rate or format, use far less power, and keep energy use steady even as traffic grows. Instead of replacing electronics, our switches work alongside them, taking on the heaviest data tasks. This lowers costs, reduces power needs, and makes systems more reliable, helping data centers grow in a way that’s both practical and sustainable
Kertai: How do your offerings advance data infrastructure?
Pérez-López: The ONE-32, our Optical Network Engine, brings all the control and monitoring features of programmable photonics into a compact, data center–ready device. It supports 32 ports today, with capacity set to double each year, and offers faster reconfiguration at lower cost. For data centers linking thousands of processors and memory units, this means data can move more intelligently and avoid bottlenecks that slow AI workloads. The ONE-32 shows that programmable photonics is ready for real-world deployment, not just research labs.
Kertai: How could programmable photonics shape the future of data innovation?
Pérez-López: Data centers already use between 3 and 10 percent of the world’s electricity, and this share will only grow with wider AI adoption. Cost and energy are now the biggest barriers to digital progress. Our approach addresses both by making networks faster, more adaptable, and more efficient. Beyond data centers, programmable photonics will help industries like healthcare, finance, and climate science move and analyze data at light speed. We see it as the invisible engine that will allow data-driven industries to keep advancing sustainably, without exhausting resources.
