What Great Sky does

Great Sky is a Boulder, Colorado-based deep-tech startup founded in late 2024 that designs and manufactures AI hardware based on Superconducting Optoelectronic Networks (SOENs) — a brain-inspired computing architecture spun out of 12 years of research at the National Institute of Standards and Technology. The platform combines superconducting Josephson-junction circuits for analog computation, co-located superconducting memory that stores synaptic weights directly beside the synaptic circuit (eliminating the Von Neumann bottleneck), and integrated single-photon photonic communication between neurons. Chips operate at 4 Kelvin using mature foundry nodes (standard as of 2005) and helium-4 cooling rather than millikelvin refrigeration. The company has taped out three first-generation wafers demonstrating image classification at up to 10 million frames per second, speech processing at more than a million spoken words per minute, and a reprogrammable 1,008-parameter chip, and has published a multi-generation roadmap through multi-chip modules (~100M parameters in ~3 years) and wafer-scale systems (~40 billion parameters per 300mm wafer).

The company is pre-revenue and commercializing through direct enterprise engagement — a contact form routes prospects into Partnerships, Investment, Research Collaboration, Press, and a "Contact us to try our chip" CTA on the technology page. Stated target customers are concentrated in defense (real-time signal processing and video intelligence), energy and fusion (plasma instability prediction at microsecond cadence), AI infrastructure providers (hyperscalers seeking lower cost-per-token and energy use), and research collaborators. Distribution is high-touch and US-domestic, with a Palo Alto commercial presence in addition to the Boulder headquarters and a dedicated Business Development Manager role. The $14 million seed round (March 2026, led by Bison Ventures) plus a $400k DARPA Spark Tank grant fund engineering expansion and market entry; pricing is not publicly disclosed and is quote-based. Revenue mechanics going forward are expected to be enterprise hardware sales and, potentially, research licensing — Great Sky is currently building its commercial motion rather than operating it.

Name

Great Sky

Legal name

Great Sky AI, Inc.

Website

https://greatsky.ai

Company type

Private

Founded year

2024

Operating status

Operating

Headcount range

1–10 employees

Ownership category

akta.pro rank

Product category

Neuromorphic AI Compute Hardware

NAICS

Semiconductor and Related Device Manufacturing (334413), Semiconductor and Other Electronic Component Manufacturing (33441)

SIC

Semiconductors & Related Devices (3674)

akta.pro primary industry

Compute & Acceleration ICs (CPU/GPU/AI/FPGA/ASIC) (HDAHABAC)

Keywords

• Superconducting computing hardware
• Neuromorphic AI chips
• Optoelectronic neural networks
• Brain-inspired accelerators
• Photonics AI infrastructure

Headquarters

HQ city

Boulder

HQ country

United States

HQ region

North America

Offices2 records

Markets served

GTM type

B2B

Offering type

Hardware or Manufacturing

Cost components

Technology or R&D, Personnel, Infrastructure, Operations, Marketing or Sales

Revenue model

1. Chip / hardware sales (pre-revenue / future): Pre-revenue. Company invites prospects to "Contact us to try our chip" and is commercializing first-generation SOEN chips targeted at defense, energy, and AI-infrastructure customers. Revenue model not publicly disclosed; expected to be enterprise hardware sales and/or access to compute based on SOEN systems.
2. Research collaboration and licensing (potential): Originated from 12 years of NIST research with 8 patents licensed out to the company. The Contact form includes a "Research collaboration" inquiry type, suggesting potential research partnerships, joint development, and/or IP licensing as adjacent revenue channels.

Pricing tiers

Go-to-market motion3 records

Distribution channels3 records

Marketing channels10 records

Core offering

Great Sky designs, fabricates, and commercializes Superconducting Optoelectronic Networks (SOENs), a brain-inspired AI compute architecture that combines superconducting Josephson-junction circuits for analog computation, co-located superconducting memory for synaptic weights, and integrated single-photon photonic communication. The company has taped out three first-generation SOEN chips (image classification at up to 10M frames/sec, speech processing at over 1M spoken words/min, and a reprogrammable 1,008-parameter chip) and plans to scale to multi-chip modules, wafer-scale modules, and multi-wafer trillion-parameter cognitive systems. Access is offered to enterprise and research customers on a partnership basis via direct engagement.

Product overview

Great Sky offers a single unified hardware platform — the Superconducting Optoelectronic Networks (SOENs) architecture — together with a multi-generation chip and system roadmap. The core SOEN platform combines superconducting Josephson-junction compute, integrated photonic (single-photon) communication, and semiconductor control circuitry, and it underpins all of Great Sky's product generations: the First-Generation SOEN Chips (three tape-outs for image, speech, and reprogrammable inference), the near-term Multi-Chip Modules (MCMs), the longer-term Wafer-Scale SOEN Modules, and the future Multi-Wafer Trillion-Parameter Cognitive Systems built by tiling wafer-scale networks with an advanced fiber-optic interconnection system. The platform is sold on a partnership basis rather than as packaged consumer software; prospects are invited to "Contact us to try our chip."

Differentiator

Problem solved

Functional benefit

Products and services

• Superconducting Optoelectronic Networks (SOENs) Platform Great Sky's foundational brain-inspired AI compute architecture that combines superconducting Josephson-junction circuits for analog computation, co-located superconducting memory for synaptic weights, and integrated single-photon photonic communication. Physically instantiates neurons, synapses, and their connectivity in hardware. Sold on a partnership basis to enterprise and research customers.
• First-Generation SOEN Chips Taped-out first-generation SOEN chips demonstrating image classification (up to 10M frames/sec), speech processing (over 1M spoken words/min), and a reprogrammable 1,008-parameter inference network. For enterprise and research customers seeking access to novel AI hardware substrates.
• Multi-Chip Modules (MCMs) Near-term scalability unit that interconnects many SOEN die in multi-chip modules using fiber optics, enabling models with approximately 100 million reconfigurable parameters for video analysis and signal processing workloads.
• Wafer-Scale SOEN Modules Long-term product supporting approximately 40 billion parameters on a single 300mm wafer with continuous learning, delivering ~65x more operations per second than a human brain and 4,300x the bisection bandwidth of a state-of-the-art GPU system.
• Multi-Wafer Trillion-Parameter Cognitive Systems Future product envisioned as free-space optical and fiber-interconnected multi-wafer stacks supporting trillion-parameter multimodal models with lifelong learning, projected to sustain continuous conversations with a million people simultaneously at no perceptible latency.

Quantifiable outcome

• 300x lower LLM cost per token (projected $0.01 per 1M output tokens vs. $3 per 1M)
• +6 more outcomes

Segments5 records

Ideal customer profiles4 records

Technology focussed Yes

API detail

Has API

No

API docs

API detail

Core technology

AI maturity

App detail

AI capability13 records

Feature6 records

Partnerships

Two partnerships are on record, tiered foundational / origin partner and core / board-level adjacency.

• National Institute of Standards and Technology (NIST)foundational / origin partnerStrategic or Co-development Partner · 13 March 2026NIST is the U.S. national laboratory in Boulder, Colorado, where the core SOEN technology was developed over 12 years (2013–2024) by the Great Sky founding team. NIST licensed the underlying IP (8 patents, 23 peer-reviewed publications) to Great Sky at company founding and "remains our partner as we launch into a startup" per the CEO. Early R&D funding flowed from NIST, DARPA, IARPA, NASA, and NRC.
• Ayar Labs (via Mark Wade)core / board-level adjacencyStrategic or Co-development Partner · 13 March 2026Mark Wade, CEO and cofounder of Ayar Labs, joined Great Sky's Board of Directors as part of the March 13, 2026 announcement. Ayar Labs operates in adjacent optical interconnect technology, and the board adjacency suggests strategic technology overlap (photonic / interconnect) between the two companies.

Scale indicators8 records

Recent moves6 records

Expansion highlights6 records

Emerging players

• Groq: Groq develops LPU inference accelerators focused on low-latency, high-throughput AI inference, a similar positioning to Great Sky's projected performance advantages. It represents competing silicon approaches to the inference-economics problem Great Sky targets.
• Tenstorrent: Tenstorrent develops RISC-V-based AI accelerators targeting efficient training and inference, competing for the same alternative-hardware customer base as Great Sky. While not photonic, it is a comparable next-generation AI compute startup.
• SambaNova Systems: SambaNova builds dataflow AI accelerators for enterprise and government customers, with a similar focus on full-stack optimization for foundation model inference. It is a relevant peer in the alternative-AI-hardware category.
• Cerebras Systems: Cerebras builds wafer-scale AI accelerators targeting similar hyperscaler and inference workloads. While not photonic or superconducting, it is a comparable 'alternative architecture' AI hardware peer competing for the same customer mindshare.

Broad incumbents

• Intel (Loihi / Hala Point): Intel's Loihi neuromorphic research program and broader foundry services are relevant context: Loihi competes in brain-inspired computing research, while Intel's foundry could manufacture competing cryogenic-compatible silicon. Intel is a broad incumbent with overlapping capabilities rather than a direct SOEN competitor.
• IBM Research (Superconducting / NorthPole): IBM has decades of superconducting and neuromorphic computing research and is commercializing analog-AI accelerators (NorthPole). It represents the most credible incumbent working on adjacent superconducting/neuromorphic AI hardware.

Direct peers

• Lightmatter: Lightmatter builds photonic AI accelerators that use light for compute and interconnect, targeting similar hyperscaler and inference workloads as Great Sky's SOENs. It is the most directly comparable photonic AI compute startup and is further along in commercialization.
• Lightelligence: Lightelligence builds optical computing systems for AI, using photonic chips to accelerate matrix multiplication and inference. It addresses the same fundamental efficiency bottleneck Great Sky targets, but with a more silicon-photonics-centric stack.
• Neurophos: Neurophos is developing optical metasurface-based AI compute using light for both computation and memory, an approach that intersects Great Sky's photonic + co-located memory thesis. It is an early-stage peer in photonic AI hardware.
• Celestial AI: Celestial AI develops photonic interconnect and in-memory compute fabrics for AI accelerators, overlapping with Great Sky's photonics-plus-memory approach. Both target orders-of-magnitude efficiency gains over GPU-based systems.

Market position

Strengths5 records

Weaknesses5 records

Competitive moat4 records

Key risks6 records

Key highlights6 records

Customer concentration

• Xtwitter.com/greatskyai
• LinkedInlinkedin.com/company/great-sky-ai

Compliance2 records

Revenue estimate

Valuation estimate

Number of profiles

Profiles10 records

Funding overview

Funding rounds3 records

Investors11 records

Funding detail is available on the Subscription and Enterprise plan.Contact sales →

M&A

Investments

M&A and investment is available on the Subscription and Enterprise plan.Contact sales →

Frequently asked questions about Great Sky