Piris Labs

Roadmap & Position in Photonic Computing

Builds photonic AI inference hardware, 5x lower latency and 10x lower power than GPUs.

Company Overview

Builds a vertically integrated AI inference platform combining proprietary photonic (optical) interconnects with a custom software stack to deliver ultra-fast, scalable, and cost-effective inference for trillion-parameter AI models.

What They're Building

The company's public product roadmap & what they're committed to building.

Piris Labs has signaled plans for modular photonic interconnect hardware, a vertically optimized inference software stack, and partnerships with chip makers and ODMs to integrate optical data movement into existing data center architectures. Public positioning: 5x lower latency, 10x lower power per bit, and 2x lower cost per token versus GPU clusters. Working π Conversion Engine prototype (October 2025). SBIR government partnership secured. Early access waitlist stage.

Latest Intelligence

Zeitgeist tracks private signals to determine where the company is heading strategically.

Photonic Interconnects Target Inference Economics

May 13, 2026

Confidence:

Medium

New Intel: Piris Labs is building photonic interconnects and an inference stack aimed at lower latency, power, and cost per token. If the hardware demo holds up, it targets one of the biggest AI data center bottlenecks.

Founder and Key Execs

Ali Khalatpour, Co-Founder & CEO

Co-Founder & CEO

Keyvan Rezaei Moghadam, Co-Founder & President

Co-Founder & President

Founder Force Multiplier

Ali Khalatpour holds multiple MIT degrees (BSc EE/Math, MASc, MSc Physics, PhD EE), is a Harvard- and Stanford-trained optical scientist who developed the first room-temperature terahertz semiconductor laser, and led the development of GUSTO's optical engine for NASA.

Funding History

2025 | Ali Khalatpour and Keyvan Rezaei Moghadam co-found Piris Labs.
2025 | Working π Conversion Engine prototype completed.
2025 | SBIR government partnership secured.
2026 | Accepted into Y Combinator Winter 2026 batch.

Competitors

Wafer-Scale / Custom ASIC Inference

Cerebras Systems, Groq, SambaNova Systems.

GPU Cloud Inference

Together AI, Fireworks AI, Anyscale.

Optical/Photonic AI Hardware

Lightmatter, Luminous Computing, Ayar Labs (optical interconnects), Celestial AI.

Hyperscaler Inference

AWS Inferentia, Google TPU, Azure Maia.

Piris Labs

's Moat:

Photonic interconnects operate on different physics than electrical data movement: speed-of-light transmission with near-zero heat generation. Multiple MIT degrees in EE and physics, plus NASA optical systems experience, represent a talent and IP moat. If the technology works at scale, the cost-per-inference advantage compounds at trillion-parameter model sizes.

How They're Leveraging AI

Photonic inference optimization

Photonic interconnect-optimized inference engine that minimizes data movement bottlenecks to deliver ultra-low-latency token generation for trillion-parameter LLMs.

Scalable inference orchestration

Modular, horizontally scalable inference architecture that enables seamless scaling from single-node to multi-rack deployments without degradation in per-token latency or throughput.

Automated model adaptation

Hardware-software co-designed model optimization pipeline that automatically adapts large AI models for photonic inference, maximizing throughput and minimizing latency without manual model surgery.

AI Use Overview:

Using photonic inference optimization for optical data movement, scalable inference orchestration, and automated model adaptation for photonic hardware.

More Similar Companies

Etched

Transformer-only ASIC and inference supercomputer stack positioned against Nvidia B200.

If transformers remain dominant, hardwiring the graph into silicon plus owning the pod and rack layer gives Etched a cost-per-token edge NVIDIA cannot match without abandoning generality.

Aurorin CAD

Lets mechanical engineers design parts in seconds instead of hours with AI-native CAD.

SolidWorks has not been meaningfully challenged in 30 years. Aurorin built a parametric kernel from scratch with an AI chat interface, and a 3x SpaceX intern with GPU engineering experience is exactly the right person to do it.

BaseFrame

Helps hardware engineers find components in seconds with AI search inside Slack and Teams.

Hardware engineers waste hours cross-referencing datasheets across distributor sites. BaseFrame embeds a parts copilot directly in Slack and Teams, which is where the buying conversations already happen.

Chasi

Deploys 24/7 AI agents that automate sales, rental, and service for equipment dealers.

Industrial equipment dealers run sales, rental, and service across phone, email, and web chat with no automation. Chasi deploys 24/7 AI agents across all channels, and the founder led AI deployments at Cummins and Harley-Davidson, so he already has the Rolodex.

Back To All Companies >