# Edgescale AI

> Edgescale AI builds Physical AI Infrastructure: integrated hardware and
> software that runs AI on-site in industrial and mission-critical
> environments — factories, utilities, hospitals, energy sites, and
> logistics networks. Its flagship appliance, the Cube, deploys in under
> an hour, runs fully offline, and keeps all data, models, and workflows
> on-premises for complete data sovereignty.

Edgescale bridges the gap between cloud software and the physical edge.
The Cube connects to existing machines, sensors, PLCs, historians, and
industrial systems, providing local sovereign AI inference (vLLM runtime,
OpenAI-compatible API, embedded MCP servers) without sending data to the
cloud. Key components include the Data Manifold (zero-ETL data
integration across hundreds of sources), Continuous Agents (always-on
operational monitoring and diagnostics), Hybrid Inference (semantic
routing between local and cloud models), and Edgeport (one-click app and
agent deployment). Partners include Red Hat, NVIDIA, Palantir, Confluent,
and Hitachi. Typical customer value ranges from $500K to $15M per site
per year in downtime reduction, yield gains, energy savings, and safety
outcomes. Founded by Marc Rouanne, Brian Mengwasser, and David La Placa.

## Product

- [Meet the Cube](https://edgescaleai.com/product/): Full product
  overview — deployment model, Data Manifold, Edgeport, security
  architecture (zero-trust, TPM 2.0, AES-256), and hardware
  specifications (Intel Xeon 6520P, NVIDIA RTX Pro 5000 Blackwell,
  128GB DDR5, 12TB NVMe, 3U form factor, 5G/Wi-Fi 6 connectivity).
- [Physical AI Infrastructure](https://edgescaleai.com/physical-ai-infrastructure/):
  Definitional guide to Physical AI Infrastructure and the five-layer
  stack: Continuous Agents, Hybrid Inference, Data Manifold, Local
  Intelligent Edge Layer, and Software Distribution Fabric.
- [Request a Custom Demo](https://edgescaleai.com/custom-demo/): Demo
  request form.

## Company

- [About Edgescale](https://edgescaleai.com/about/): Mission, founding
  team, and company background.
- [Partner Ecosystem](https://edgescaleai.com/partner-ecosystem/):
  Technology and AI partners including Red Hat, NVIDIA, Palantir,
  Confluent, and Hitachi.
- [Careers](https://edgescaleai.com/careers/): Open roles.
- [Contact](https://edgescaleai.com/contact/): Get in touch.
- [Trust Center](https://trust.edgescaleai.com/): Security and
  compliance documentation.

## Blog

- [Constitutional Governance for Physical AI](https://edgescaleai.com/constitutional-governance-for-physical-ai/):
  Governance framework for AI operating in physical environments.
- [8 Lessons Learned from Deploying Physical AI at Scale](https://edgescaleai.com/8-lessons-learned-from-deploying-physical-ai-at-scale/):
  Field lessons from real-world Physical AI deployments.
- [Ringfencing AI for US Industrial Advantage](https://edgescaleai.com/ringfencing-ai-for-us-industrial-advantage/):
  Sovereign AI and American re-industrialization.
- [Optimize Manufacturing with AI](https://edgescaleai.com/optimize-manufacturing-with-ai/):
  AI-driven yield, quality, and uptime gains on the factory floor.
- [Winning the Energy Race with AI](https://edgescaleai.com/winning-the-energy-race-with-ai/):
  Edge AI for energy and oil & gas operations.
- [Hitachi: Where No Cloud Has Gone Before](https://edgescaleai.com/hitachi-where-no-cloud-has-gone-before/):
  Partnership perspective on extending cloud capability to the edge.
- [Revolutionizing Edge AI](https://edgescaleai.com/revolutionizing-edge-ai/):
  The shift from cloud-centric to edge-native AI.
- [Extending the Cloud to Scale AI in the Real World](https://edgescaleai.com/edgescale-ai-extending-the-cloud-to-scale-ai-in-the-real-world/):
  Edgescale's founding thesis.

## Optional

- [Blog index](https://edgescaleai.com/blog/): All posts.
- [Sitemap](https://edgescaleai.com/sitemap_index.xml): XML sitemap index.