safety-monitoring
(4)
Archetype AI
Archetype AI is a San Mateo-based physical AI company founded by former Google Advanced Technology and Projects (ATAP) leaders. It builds Newton, a multimodal foundation model that fuses real-time data from cameras, radar, microphones, accelerometers and dozens of other sensors with natural language, so machines can perceive, understand and reason about the physical world. On top of Newton, the company ships an Archetype Platform, an Agent Toolkit and prebuilt Physical Agents that turn raw sensor streams into actionable insight for manufacturing, construction, energy, telecom and city infrastructure. It raised a $35M Series A in November 2025, bringing total funding to $48M.
Invisible AI
Invisible AI builds edge-based computer vision systems for manufacturing floors. Its self-contained camera devices - each pairing an Intel RealSense 3D depth camera with an NVIDIA AI chipset and on-board storage - mount to existing factory infrastructure and turn manual assembly work into structured, cycle-level data without sending video to the cloud. Founded in 2018 by veterans of the autonomous-vehicle world, the company helps automakers and aerospace manufacturers such as Toyota spot cycle-time variation, ergonomic risk, and quality defects in real time. In 2026 it launched what it calls the world's first Vision Execution System, built on NVIDIA Metropolis and Cosmos Reason models, to run AI agents that analyze every production cycle at the edge.
SafeinHome
SafeinHome is a remote supports company that helps people with intellectual and developmental disabilities (I/DD) and older adults live more independently. It pairs a system of discreet home sensors and a two-way tablet with trained staff available around the clock, so individuals get reminders, check-ins and real-time help through the day and night without a caregiver physically present at all times. The service is delivered as a Medicaid Home and Community-Based Service (HCBS) across 17 states, aiming to cut per-hour support costs while preserving each person's dignity and autonomy.
Matroid
Matroid is a Palo Alto computer vision company that lets anyone build, train, and deploy custom object detectors without writing code. Founded in 2016 by Stanford professor and former Databricks founding-team member Reza Zadeh, its platform turns ordinary camera feeds into searchable, alert-generating systems that spot objects, people, defects, events, and safety violations in real time, across cloud, on-premise, and edge deployments. Matroid is used across manufacturing, aerospace, automotive, electronics, and video security to automate quality inspection and monitoring.