Great Sky Gains $14 Million in Seed Funding
Great Sky, a startup based in Boulder, Colorado, has successfully raised $14 million in seed funding. The round was led by Bison Ventures, with additional participation from Matchstick Ventures and Range Ventures. The company is known for pushing the physical limits of intelligence through its development of a brain-inspired computing platform designed to enhance artificial intelligence capabilities.
Innovative AI Computing Approach
Great Sky is pioneering a novel approach to AI computing by integrating superconducting computation with optical communication. This architecture, termed Superconducting Optoelectronic Networks (SOENs), aims to overcome the limitations of traditional AI systems that predominantly rely on transformers and GPUs. The platform combines superconducting circuits operating at cryogenic temperatures with optical links that can transmit signals as faint as a single photon. This advanced system mimics biological neural networks, allowing for continuous learning and on-device adaptation to new data streams.
Founders and Leadership
The company was founded by a team of experts in the field, including Jeffrey Shainline, who serves as CEO, Jeffrey Chiles, the CTO and co-founder, and Bryce Primavera, the co-founder and VP of Architecture. Their collective experience is rooted in over a decade of research at the U.S. National Institute of Standards and Technology (NIST), where they developed the technology's underlying components.
Use of Funds
The newly acquired funds will support the advancement of Great Sky's AI systems. The company plans to scale its technology by connecting wafer-scale neural networks using advanced fiber-optic interconnections. This scaling up is expected to enhance the capability of AI systems to handle high-throughput, low-latency workloads, such as real-time video analysis and other multimodal AI applications.
Future Developments
Great Sky has already demonstrated significant improvements in energy efficiency and performance with their chip technology, which can process over 60 million video frames per second. This performance far exceeds the typical capabilities of conventional GPU-based systems, which handle approximately 30 frames per second. Looking ahead, the company envisions building larger AI systems with more neurons and synapses, potentially revolutionizing the efficiency and capability of AI technologies.