Skip to main content

Neurophos

Category: AI Chips / Semiconductors

Neurophos is developing high-speed, energy-efficient optical AI accelerators that utilize proprietary metamaterial technology to overcome the physical and energy limitations of traditional silicon-based processors. Neurophos was founded in 2020. The company is led by Patrick Bowen. Based in Austin, USA. Team size: 11-50. Total funding raised: $110.0M. Latest round: Series A ($110M, Jan 2026). Key investors include ["Gates Frontier","MetaVC Partners","M12 (Microsoft)","Bosch Ventures","Aramco Ventures","Mango Capital","Antecedent Ventures"].

Founded
2020
Headquarters
Austin, USA
Team size
11-50
Total funding
$110.0M

Value proposition

Breaks the 'memory wall' of traditional semiconductors by delivering up to 100x-200x improvements in energy efficiency and compute density compared to state-of-the-art GPUs (e.g., Nvidia B200).

Products and solutions

["Optical Processing Unit (OPU) - High-density AI inference accelerator","Micron-scale Metasurface Modulators - 10,000x smaller than traditional photonic elements","PyTorch-compatible Software Compiler - Allows direct compilation of standard AI code for optical hardware","Exaflop-scale Photonic AI Chips - Targeted for large-scale data center deployments"]

Unique value

Neurophos is the first to apply metamaterials to optical computing, enabling a 10,000x miniaturization of optical modulators compared to standard silicon photonics, allowing for the world's densest optical in-memory compute platform.

Target customer

Hyperscale data center operators, AI infrastructure providers, and enterprise clients focused on sustainable/green AI computing.

Industries served

["Semiconductors","Artificial Intelligence","Data Centers","Sustainable/Green Computing"]

Technology advantage

Combines the speed of light-based computing with the manufacturability of standard CMOS processes, allowing for massive scaling without the energy overhead of traditional electronic data movement.

How they differentiate

Neurophos utilizes proprietary metamaterials to create optical modulators that are 10,000x smaller than traditional silicon photonics, enabling the world's densest optical in-memory compute platform and breaking the 'memory wall' that limits electronic chips.

Main competitors

["Lightmatter","Celestial AI","Nvidia"]

Key partnerships

["Terakraft (Pilot partner for Green AI Factory data centers)","Silicon Catalyst (Semiconductor incubator and strategic accelerator)","Duke University (Research and IP spin-out partner)","Metacept Inc. (Technology incubator and founding partner)","Microsoft (Identified as a strategic interest/potential early adopter)"]

Notable customers

["Terakraft (Pilot partner for Green AI Factory data centers)"]

Major milestones

["Secured $7.2M Seed funding led by MetaVC Partners in 2023","Achieved 10,000x miniaturization of optical elements compared to standard silicon photonics","Closed $110M Series A round in January 2026 to scale production of Optical Processing Units (OPUs)","Established strategic partnership with Terakraft for sustainable AI infrastructure"]

Growth metrics

Scaling from R&D to production following a $110M Series A; team size estimated between 11-50 employees.

Market positioning

Deep-tech semiconductor challenger targeting the high-performance AI inference and data center market.

Geographic focus

Global (Headquartered in Austin, Texas, with focus on North American and European hyperscale data centers)

Patents and IP

Leverages a portfolio of over 300 patents derived from decades of research at Duke University and Metacept Inc., including exclusive licenses for metamaterial IP from MetaVC Partners.

About Patrick Bowen

Dr. Patrick Bowen is a leading expert in metamaterials and nanophotonics with over a decade of experience in translating complex physics into commercial hardware. Prior to founding Neurophos, he served as the CTO of Metacept Inc. and held pivotal research and engineering roles at Seattle-based metamaterials leaders Lumotive and Kymeta Corporation. He earned his PhD in Electrical and Computer Engineering from Duke University and a Master's in Micro and Nano Systems from ETH Zurich.

Official website: