TechniaHQRobot logoTechniaHQRobot

ANNOUNCED ≠ PILOT ≠ OPERATIONAL

Physical AI Deployment Tracker

Track documented Physical AI and embodied robotics deployments while separating pilots, agreements, operating sites, model disclosure and intervention evidence.

5 retained recordsLast verified: August 13, 2026Primary-source fields preserved per record
5 of 5 records shownFilters are local and do not create crawlable URLs

Not publicly disclosed in the deployment source

Official manufacturer deployment report
Physical AI model
Not publicly disclosed in the deployment source
Robot
Figure 02
Robot company
Figure AI
Customer
BMW Manufacturing
Deployment site
Spartanburg, South Carolina, United States
Industry
Automotive manufacturing
Task
Sheet-metal loading in BMW X3 production
Start
2025
Robot count
Not disclosed
Stage
Completed production deployment reported by Figure
Operating hours
>1,250 runtime hours reported
Autonomy evidence
Deployment source does not publish a complete autonomy/intervention protocol
Status
Completed reported deployment
Source
Figure — production at BMW

Not publicly disclosed in the deployment source

Official manufacturer deployment update
Physical AI model
Not publicly disclosed in the deployment source
Robot
Figure 03
Robot company
Figure AI
Customer
BMW Manufacturing
Deployment site
Hall 52, Spartanburg, South Carolina, United States
Industry
Automotive manufacturing
Task
Sequencing use case
Start
2026-06-30
Robot count
Not disclosed
Stage
Deployment update
Operating hours
Not disclosed
Autonomy evidence
Not disclosed as a full operational protocol
Status
Deployment update
Source
Figure — F.03 at BMW

Not publicly disclosed as a named foundation model

Official deployment announcement
Physical AI model
Not publicly disclosed as a named foundation model
Robot
Agility Digit
Robot company
Agility Robotics
Customer
GXO Logistics
Deployment site
Flowery Branch, Georgia, United States
Industry
Logistics
Task
Moving totes between an AMR and conveyor
Start
2024-06-05
Robot count
Not disclosed in cited update
Stage
Commercial deployment
Operating hours
Not disclosed
Autonomy evidence
Task automation; exact intervention rate not disclosed in cited source
Status
Operational deployment
Source
Agility Robotics — Digit at GXO

Not publicly disclosed in the deployment announcement

Official company/customer agreement
Physical AI model
Not publicly disclosed in the deployment announcement
Robot
Apptronik Apollo
Robot company
Apptronik
Customer
Mercedes-Benz
Deployment site
Mercedes-Benz manufacturing operations
Industry
Automotive manufacturing
Task
Manufacturing pilot
Start
2024-03-15
Robot count
Not disclosed
Stage
Pilot announced
Operating hours
Not disclosed
Autonomy evidence
Pilot announcement does not establish unrestricted autonomy
Status
Pilot
Source
Apptronik — Mercedes-Benz agreement

Not publicly disclosed as one named model in retained deployment record

Official announcement in retained deployment dataset
Physical AI model
Not publicly disclosed as one named model in retained deployment record
Robot
AgiBot A2 Ultra
Robot company
AgiBot
Customer
Longcheer Technology
Deployment site
Consumer-electronics factory, China
Industry
Electronics manufacturing
Task
Inspection and material handling
Start
2026-06-23 pilot date in retained source-linked dataset
Robot count
Not disclosed
Stage
Pilot / factory operation report
Operating hours
Not disclosed
Autonomy evidence
Do not infer from task count alone
Status
Pilot evidence
Source
AgiBot deployment source

Methodology

A partnership announcement is not upgraded into a deployment. When the robot is deployed but the exact Physical AI model is not public, the model field remains Not publicly disclosed instead of guessing which internal policy was used.

Included

Published model, robot, benchmark, deployment or failure evidence that can be tied to a retained source.

Not inferred

Unknown autonomy, latency, trial counts, deployment scale and failure causes remain unknown when the source does not disclose them.

Update policy

Versions and historical records remain distinguishable. A new release does not silently overwrite an older model or evaluation condition.