iPronics provides a rack‑ready silicon photonic optical switch (iPronics ONE) that can reconfigure network topologies in sub‑millisecond intervals, delivering lossless, low‑latency optical circuit switching for AI training and inference. The switch’s programmable photonic unit cells, built‑in telemetry, and Python APIs enable software‑defined control, reducing power consumption by up to three times and improving cost‑to‑capacity by 25× compared with traditional electronic switches, thereby increasing GPU utilization and lowering data‑center operating costs.
Funding
$11.7M raised to dateRaised to date based on public sources. This may differ from the amount the company actually raised and is based only on what is publicly available on the internet.
Founders
Product
Problem
AI training and inference workloads demand ever‑increasing bandwidth, low latency, and high GPU utilization, but existing data‑center networking relies on copper interconnects and electronic packet switches that are power‑hungry, inflexible, and become bottlenecks as scale grows.
Solution
iPronics offers a rack‑ready silicon photonic optical switch (iPronics ONE) that can be reconfigured in sub‑millisecond intervals to dynamically adjust network topology for AI workloads. The switch uses programmable photonic unit cells and integrated driving and monitoring electronics to provide lossless, low‑latency optical circuit switching with built‑in telemetry. Software‑defined control and Python APIs enable seamless integration into existing SDN environments, allowing real‑time routing, fault avoidance, and automatic performance optimization without physical rewiring. The compact design delivers up to three times lower power consumption and a 25× better cost‑to‑capacity ratio compared with traditional electronic switches, improving GPU utilization and reducing overall data‑center operating costs.
Target Audience
Primary customers are hyperscale data‑center operators, AI cloud service providers, and enterprise HPC clusters that require high‑bandwidth, low‑latency interconnects for GPU‑intensive workloads.
Features
- Sub‑millisecond real‑time reconfiguration of optical paths via programmable silicon photonic unit cells
- Integrated driving unit and on‑chip monitoring photodetectors for per‑channel gain control and lossless operation
- Built‑in telemetry and automatic calibration algorithms for continuous performance optimization
- Software‑defined networking integration with Python APIs, supporting dynamic routing and orchestration
- Rack‑ready, stackable form factor offering 32 to 256 ports with ultra‑flat O‑band spectral response
- Up to 3× lower power consumption and 25× better cost‑to‑capacity ratio versus traditional electronic switches
- Lossless optical circuit switching eliminates transceiver conversions, reducing latency and heat