In April 2025, California-based startup nEye Systems secured $58 million in Series B funding to accelerate its mission of transforming AI data center infrastructure. This round brought the company’s total capital to $72.5 million. Backing came from major industry leaders, including Alphabet’s CapitalG, Microsoft’s M12, Micron Ventures, NVIDIA, and Socratic Partners. nEye Systems now stands at the forefront of a new era in data networking by focusing on optical circuit switching—a technology that promises to overcome severe communication bottlenecks in high-performance computing and AI systems.
Founded in 2020 by researchers from the University of California, Berkeley, including the renowned Professor Ming Wu, nEye Systems set out to challenge the limitations of traditional electrical data center infrastructure. Their flagship product, SuperSwitch, aims to replace legacy electrical switches with highly efficient wafer-scale optical switching solutions. These innovations could reduce power consumption, accelerate data transfer, and enhance GPU efficiency for next-generation AI workloads.
Problems in Modern Data Centers
AI models continue to increase in size and complexity. With these advances, AI training systems must move immense volumes of data between thousands of GPUs, memory modules, and processing units. Current data centers rely heavily on electrical switching technologies, which struggle to keep up with this demand. These limitations cause network congestion, reduce the overall utilization of GPU clusters, and create energy inefficiencies that drive up operational costs.
Traditional switches require significant power, occupy large physical space, and introduce latency. As companies scale their infrastructure, these limitations result in underused computing resources and longer AI model training cycles. Data center architects increasingly recognize the need for a new communication model—one that supports rapid data movement, maintains scalability, and consumes minimal energy.
nEye’s Optical Solution: SuperSwitch
nEye Systems designed its SuperSwitch to meet these exact demands. This switch uses MEMS-based silicon photonics to create optical connections between GPUs and memory components. Instead of converting light signals into electrical pulses and back again—a costly and inefficient process—nEye’s SuperSwitch maintains end-to-end optical communication. The result is a dramatically more efficient and faster method of data transfer inside high-density AI infrastructure.
The SuperSwitch offers groundbreaking advantages over conventional systems:
- Size Efficiency: The switch occupies 100 times less space than traditional optical switching racks. Its compact form factor enables deployment inside server racks rather than in dedicated, bulky frames.
- Energy Consumption: SuperSwitch technology consumes 1,000 times less power than standard switching equipment. This efficiency enables massive energy savings in large-scale deployments and supports global sustainability goals.
- Reconfiguration Speed: The device reconfigures data pathways up to 10,000 times faster than legacy systems. This responsiveness allows AI workloads to adapt instantly to changing computational needs.
- Cost Efficiency: On a per-bit basis, the SuperSwitch operates at one-tenth the cost of traditional switching methods, making it viable for broad-scale deployment without requiring extravagant capital expenditures.
These improvements flatten the data center network architecture, eliminate unnecessary layers, and create direct GPU-to-GPU communication channels. Such a configuration enhances throughput, reduces latency, and increases fault tolerance.
Industry Backing and Strategic Impact
The strength of nEye Systems’ investor roster demonstrates the importance of this technology. NVIDIA joined the round with a clear interest in reducing GPU underutilization. High-speed optical switching directly supports NVIDIA’s focus on maximizing the computational efficiency of their AI accelerators.
Microsoft’s Azure division also expressed strong interest. Cloud service providers like Azure face enormous demands from enterprises deploying AI-driven applications. Faster and more efficient data movement inside data centers gives Azure a distinct advantage in performance, cost control, and sustainability.
Micron Ventures, the investment arm of memory giant Micron Technology, contributed because of the complementary nature of their business. Efficient switching allows faster memory access, improving the performance of DRAM and NAND-based storage in AI systems.
Alphabet’s CapitalG led the round and praised the innovation as a critical enabler of future AI scalability. Their representatives stressed that solving data movement inefficiencies unlocks the full potential of existing AI hardware and opens new architectural possibilities for hyperscale computing.
Research and Development Foundations
nEye Systems built its core technology on years of university research in silicon photonics. Professor Ming Wu, a pioneer in micro-electromechanical systems (MEMS) and optical integration, helped found the company with a vision to commercialize academic breakthroughs. The founding team brought together experts in photonic circuit design, high-speed networking, and AI infrastructure.
Since its inception, nEye Systems developed several prototypes of the SuperSwitch. Engineers fabricated wafers using advanced CMOS-compatible processes and tested the optical MEMS arrays for performance, reliability, and scalability. Early results showed exceptional signal integrity and consistent operation under data center conditions.
The engineering team now focuses on scaling manufacturing, integrating the technology into standard racks, and testing deployments with select data center partners. This path ensures that customers can adopt the solution without completely redesigning existing infrastructure.
Production Plans and Future Outlook
nEye Systems plans to release production samples of the SuperSwitch by late 2025. Engineers currently work with contract manufacturers and chip fabrication partners to prepare for small-batch pilot production. Once validated at scale, the team will move toward commercial manufacturing for hyperscalers, cloud providers, and enterprise AI clients.
Several potential customers already expressed interest in pilot testing. While the company has not publicly disclosed its customer list, insiders suggest that multiple cloud providers and AI startups began discussions on integration planning and performance benchmarking.
The startup also investigates new use cases in supercomputing, real-time robotics, autonomous vehicles, and edge AI inference, where low-latency and high-bandwidth communication remains essential. As AI models push deeper into mission-critical applications, demand for high-speed, low-power communication infrastructure will continue to grow.
The Broader Impact on AI and Data Center Infrastructure
If successfully deployed at scale, nEye Systems’ technology could redefine the physical architecture of AI data centers. By replacing multiple layers of electrical switches with a single optical switching layer, data centers can reduce equipment footprint, eliminate excessive cabling, and simplify network management.
The impact extends beyond hardware efficiency. Optical switching also enables new types of AI training topologies, where models distribute more evenly across available compute nodes. This balance increases training speeds and reduces idle cycles, leading to more predictable performance and faster development cycles for cutting-edge machine learning models.
Environmental benefits also play a critical role. With global data centers consuming vast amounts of electricity, a shift toward energy-efficient optical switching aligns with broader ESG goals. Corporations and governments alike seek infrastructure solutions that reduce carbon footprints without compromising performance.
Conclusion
nEye Systems stands at the cusp of a major technological transformation in AI data infrastructure. By introducing a compact, ultra-fast, and energy-efficient optical switching platform, the company addresses some of the most pressing problems facing data centers today. The $58 million funding round reflects deep investor confidence in both the technology and the team behind it.
As AI models grow in complexity and power demands continue to rise, innovations like the SuperSwitch will become essential. With the right execution, nEye Systems could play a pivotal role in shaping the next decade of high-performance computing, making AI systems faster, greener, and more efficient than ever before.