IPronics, a chip startup, has pulled in $125 million in funding from Nvidia and other investors. The money will go toward developing chips that can rewire data centers in under a millisecond — a speed that could change how computing infrastructure handles heavy workloads.
The funding round
The round was led by Nvidia, with additional investors joining in. iPronics didn't disclose the full list of backers, but the size of the raise signals serious confidence in the technology. For a company working on optical or photonic switching, that kind of capital is a big step toward production.
Nvidia's involvement is notable. The GPU maker has been pushing into networking and data center interconnect, and iPronics' approach fits neatly into that strategy. By backing a startup that can physically reconfigure connections in microseconds, Nvidia gets a stake in a piece of infrastructure that could make its own chips more efficient.
What the chips do
Traditional data centers rely on fixed wiring and switches that take time to reroute traffic. iPronics' chips are designed to change the topology of a network on the fly. Instead of waiting for a switch to update routing tables, the chip can redirect data paths in under a millisecond.
That speed matters for workloads that shift constantly — think AI training, where data flows between thousands of GPUs. If a link becomes congested or a server fails, the network can rewire itself almost instantly, keeping the job moving without a hiccup.
Data centers are hungry for power, and a big chunk of that goes to moving data around. By making the network more flexible, iPronics says its chips can cut energy use and lower operating costs. The company also points to better GPU utilization — if GPUs aren't waiting on slow network paths, they spend more time actually computing.
Scalability is another selling point. As data centers grow, the ability to rewire connections without physical changes means operators can adapt to new demands without ripping out cables or adding hardware. That's a practical advantage for cloud providers and enterprises running large AI clusters.
The funding will be used to push the chips from prototype to deployment. iPronics hasn't announced a specific release date, but the company is now in a position to scale up manufacturing and work with early customers.
