Great Sky Unveils Brain-like Computing Architecture Built for the Next Era of AI

Estimated reading time: 4 minutes

Great Sky is a company pioneering a fundamentally new computing architecture for AI. Today, the company announced the public debut of its technology, fundraising, and several major commercial milestones. Its superconducting optoelectronic technology is backed with a $14 million seed round led by Bison Ventures with participation from notable firms including Matchstick Ventures and Range Ventures, as well as angel investors like Mark Leslie, Adam Pritzker, and Ivan Vendrov. Great Sky has also passed major technical milestones including the tape-out of its chips that deliver orders-of-magnitude improvements in efficiency and performance over today’s silicon.

For decades, researchers have described an idealized computing paradigm—one that combines ultra-low-temperature superconducting computation, high-bandwidth optical communication, and brain-inspired architectural principles—as a theoretically superior alternative to conventional approaches to AI. The team has steadily made progress to prove the components, and they have now been integrated into complete neural networks. Advances in circuit concepts, manufacturing capabilities, and modeling breakthroughs have made the moment possible.

Great Sky is in a unique position in the world to bring this long-imagined architecture into deployment. The technology and company are premised on 12 years of research at the National Institute of Standards and Technology (NIST), where the team produced 23 peer-reviewed publications and eight patents. The technology arrives as a purpose-built compute platform for the next era of AI that requires high-throughput, low-latency computation for use cases like intelligent processing of spoken language, continuous video analysis, and large-scale multimodal models.

“AI’s current stack—transformers running on GPUs—has brought tremendous advances. But at the foundation, the current approach is mismatched to the needs of efficient, scalable AI,” said Jeff Shainline, co-founder and CEO of Great Sky. “By constructing new hardware that enables more sophisticated architectures and algorithms while performing operations near the physical limits of speed and energy efficiency, we can transition to a completely different roadmap for scaling that doesn’t require hundreds of billions in capex and gigawatt data centers. There’s a vast, new space to explore.”

Breaking the Performance Plateau

Across the industry, researchers and practitioners have pointed out the looming limits of GPU scaling: diminishing returns on model size, unsustainable energy demands, and latency ceilings that impede agentic and multimodal workloads. Morgan Stanley analysts have cautioned that AI demand could leave the US with a power deficit of as much as 13 gigawatts by as early as 2028.

Great Sky’s neuroscience-inspired architecture overcomes those limits by integrating three technological pillars:

Superconducting compute

Enables extreme energy efficiency and unprecedented performance at cryogenic temperatures. Instead of digitally computing what neurons do, Great Sky builds circuits that simply behave like neurons.

High-bandwidth optical connectivity

Achieves communication speed and efficiency far surpassing what electrical interconnects allow, eliminating one of the largest bottlenecks in AI systems. Great Sky sends optical signals as faint as a single photon, achieving the lowest latency and energy usage physically possible.

Semiconductor circuitry

Bridges the gap between electrons and photons, performing amplification, light emission, and control logic.

The result is a system that behaves more like a brain than a conventional computer and with performance characteristics that unlock entirely new classes of AI applications. Architectures supported by these pillars map naturally to multimodal and continuous-stream AI workloads.

“AI is a domain with tremendous promise, while quantum computing is exciting because it introduces new devices and information-processing concepts. We use many of the same devices as quantum technologies – superconducting circuits, single-photon detectors – but we use them for classical, analog computation in a parallel, distributed architecture, which is ideal for AI,” said Shainline. “And unlike quantum computers, our systems are highly resilient to noise, can operate at much higher temperatures than superconducting qubits, and are broadly applicable to any application that benefits from intelligence.”

Great Sky pushes brain-inspired computing further than prior approaches. By co-locating memory and processing, Great Sky enables high-speed programmability as well as learning algorithms that support on-device adaptation, so systems can learn from new data streams without constant retraining cycles. This tighter coupling of memory and learning also improves stability and alignment, making it easier to shape behavior through reinforcement signals and constraints as models update over time. Additionally, by leveraging superconducting and photonic device physics rather than relying on continuing to scale down to few-nanometer transistors, this hardware can be manufactured at much lower cost than today’s GPUs while taking advantage of the capabilities of US foundries.

Built for the Future of AI: Real-Time Multimodal and AI-Native Infrastructure

Great Sky’s first-generation of chips has shown the company’s performance gains with immediate and long-term breakthrough use cases in commercial, defense, and energy sectors. For instance, when performing video analysis, Great Sky’s approach processes over 60 million video frames per second with orders of magnitude less energy per video analyzed than a conventional model running on today’s GPUs, which typically process video at 30 frames per second or less.

“Great Sky is rethinking compute from first principles by building hardware inspired by the architecture of the human brain,” said Ari Wright, principal at Bison Ventures. “They are setting out to enable forms of intelligence that feel inherently human, while achieving speeds and energy efficiencies traditional architectures can’t approach. At Bison, we look for visionary founders pursuing non-incremental, deep technical innovation with the potential to transform the world as we know it. Great Sky exemplifies that.”

The company is manufacturing its technology in a robust, scalable process already evidenced through multiple successful tape-outs. Long term, Great Sky will develop multi-modular cognitive systems created by tiling wafer-scale networks with an advanced fiber-optic interconnection system. These will unlock supercomputers with more neurons and synapses than the human brain, operating a million times faster than biologics and far beyond today’s largest GPU clusters.

Great Sky’s progress reflects over a decade of foundational research spanning superconducting materials, optical networking, cryogenic system engineering, and novel AI-native architectures. The company was founded by CEO Jeff Shainline, CTO Jeff Chiles, VP of Fabrication Saeed Khan, and VP of Architecture Bryce Primavera — electronics, optics, and physics PhDs and former NIST researchers who have spent the last 12 years building this technology. As part of the news, Tom Biegala, Founding Partner at Bison Ventures, and Mark Wade, CEO and cofounder of Ayar Labs, will join its Board.