nEye.ai Raises $80M to Scale AI Optical Switching

Why It Matters

Optical switching is becoming essential for AI training at hyperscale, and nEye.ai’s capital raise positions it to meet the power‑efficiency and latency demands of next‑generation data centers. The backing of top cloud‑provider investors signals industry confidence in the technology’s commercial viability.

Key Takeaways

• nEye.ai secured $80M Series C, total funding $152M.
• Investors include Sutter Hill, CapitalG, M12, Micron, boosting credibility.
• SuperSwitch chip uses silicon photonics, MEMS, CMOS for optical switching.
• Optical circuit switches cut power and latency for gigawatt-scale AI data centers.
• Company targets $3B market, scaling wafer‑level production for composable infrastructure.

Pulse Analysis

The surge in generative‑AI models has pushed data‑center architects to confront a "gigawatt factory" dilemma: thousands of GPUs demanding bandwidth far beyond the limits of copper‑based networking. Conventional electrical switches not only consume excessive power but also introduce latency that can stall training cycles. nEye.ai’s SuperSwitch integrates silicon photonics, MEMS actuation, and CMOS logic onto a single foundry‑ready chip, enabling light‑based data paths that dramatically reduce heat and energy use while delivering sub‑nanosecond switching speeds. This shift from electrical to optical interconnects is poised to become a foundational layer for AI infrastructure.

The $80 million Series C, led by Sutter Hill Ventures and supported by Alphabet’s CapitalG, Microsoft’s M12, and Micron, underscores the strategic importance of optical switching to the cloud ecosystem. The capital will accelerate wafer‑scale production of OCS modules, allowing nEye.ai to meet rising demand for composable infrastructure that can dynamically reconfigure CPU, GPU, and memory resources at runtime. With the optical circuit switching market projected to exceed $3 billion within three years, the infusion positions nEye.ai to capture a sizable share by offering the lowest cost per unit through existing foundry partnerships, sidestepping the massive capex of building proprietary fabs.

Competitive pressure is intensifying as traditional hardware vendors scramble to integrate photonic solutions, yet nEye.ai’s approach of leveraging standard silicon foundries grants it speed and flexibility unmatched by rivals tied to custom fabs. As AI models from OpenAI and Google grow in parameter count, the data‑center “plumbing” will be as critical as the chips themselves. The involvement of the two largest cloud providers signals that nEye.ai’s technology could become the unseen backbone of the next AI wave, driving both performance gains and sustainability improvements across the industry.