Bay Area Robotics Supply Chain

Facilities, components, and deployment data. A sourced review of the cluster, constraints, and the opportunity for robot-deployment spatial intelligence.

Four evidence classes

Official — Company/regulator published
Reported — News/research org
Analyst — Investment bank estimate
Assessment — AI Aerial interpretation

The overlooked supply-chain input is facility data.

The strongest near-term opportunity for AI Aerial is not manufacturing actuators, roller screws, motors, batteries, or sensors. It is robot-deployment spatial intelligence: capturing, structuring, and maintaining the facility data that robotics teams need before installation, commissioning, and expansion.

Factory capture is a useful entry project. The scalable business is broader: capture a robotics factory → convert to controlled spatial baseline → add robot-specific annotations → follow the OEM into customer deployment sites → maintain the record as layouts change.

7.6M sq ft and growing.

7.6M
Sq ft of Bay Area robotics leases (JLL, 2026)
220+
Robotics leases in Bay Area (JLL, 2026)
<500K
Sq ft in 2020 — 15× growth
46
Priority companies in our workbook

Interpreted correctly.

On Tesla's July 22, 2026 earnings call, Musk described Tesla as having to build the Optimus supply chain rather than relying on a mature, existing supplier base. This should be interpreted narrowly: humanoid architectures require new combinations of motors, transmissions, sensors, batteries, electronics, structures, hands, and software. Many components must be redesigned for humanoid size, weight, cost, duty cycle, and volume.

Figure qualified hundreds of suppliers against incoming inspection criteria while developing dedicated manufacturing lines and 50+ in-process inspection points. 1X takes a highly integrated approach — manufacturing motors, batteries, structures, transmissions, soft goods, sensors, and other critical systems internally.

Who's expanding, and how fast.

Official · Company disclosure
1X Technologies
80K sq ft Palo Alto HQ · 58K sq ft Hayward factory (200+ employees) · 10K NEOs/yr capacity · San Carlos facility online 2026 · Target: 100K units/yr by end 2027
Official · Company disclosure
Figure AI — BotQ
350+ Figure 03 robots produced · Cycle time: 1 robot/day → 1/hour · 150+ networked workstations · 50+ in-process inspection points · 9,000+ actuators across 10+ SKUs
Official · Company disclosure
Agility Robotics
60K sq ft Fremont facility (Jul 2026) · ~200 hires planned · $300M+ Digit v5 orders · 30+ customer pipeline · Deployments: GXO, Schaeffler, Toyota, Mercado Libre
Official · Partnership
Figure × Brookfield
Real-world data collection for pretraining · 100K residential units · 500M+ sq ft office · 160M sq ft logistics. Validates real-world deployment access.

What goes into a humanoid robot.

SystemRepresentative InputsTypical Challenge
ActuationMotors, drives, reducers, screws, bearings, encoders, brakesTorque density, weight, heat, efficiency, cost, life
Hands & ManipulationMicro-motors, tendons, gears, tactile sensors, force sensingDexterity, durability, manufacturability
PerceptionCameras, lenses, depth sensing, IMUs, microphonesOcclusion, glare, latency, calibration
Compute & ControlsEdge compute, motor controllers, safety controls, firmwarePower, heat, reliability, cybersecurity
PowerCells, pack structure, BMS, charging, thermal managementRuntime, mass, cycle life, fire risk
StructureCastings, machined parts, composites, covers, fastenersLow mass, strength, repeatability, appearance
WiringHigh-flex harnesses, connectors, shielding, strain reliefRouting, flex life, serviceability
ManufacturingTooling, fixtures, winding, molding, calibration, testYield, takt time, traceability
Software & DataFoundation models, teleop data, simulation, fleet mgmt, OTAData quality, generalization, long-tail failures
Deployment InfrastructureCurrent facility model, routes, charging, network, workflowsSite variation, stale drawings, coordination

Where AI Aerial fits.

Strategic assessments for AI Aerial Imagery LLC — not market-size statistics.

OpportunityFit NowRevenue PathRiskScore
Robot deployment facility dataHighProject + multi-site programMedium9.2/10
Recurring change captureHighSubscription / scheduled updatesMedium8.6/10
Factory spatial baseline + scan-to-BIMHighOne-time project + refreshesMedium8.2/10
Simulation handoff preparationMediumProject add-on through partnerHigh6.8/10
Facility evidence for safety reviewMediumPartner-led project add-onHigh6.0/10
Incoming inspection / supplier QALowSpecialist partnershipHigh3.5/10
Core humanoid component manufacturingVery lowLong qualification cycleVery high1.5/10
Teleoperation staffingLowLabor-based serviceHigh2.0/10

Strategic targets.

Agility Robotics Field Operations
Fremont expansion, field-ops hiring, active deployments, 30+ customer pipeline. Proposed: Digit Deployment Facility Pilot — versioned spatial-readiness package for one representative test zone or customer site.
Robust AI
Phased ShipLab deployment model creates potential need for pre-deployment baselines, constraints, commissioning evidence, and change tracking.
InOrbit Robot Space
Mountain View demonstration environment for multi-robot route annotations, shared zones, layout-change detection, facility-data handoff, partner-produced simulation example.
1X San Carlos
Lighthouse facility target. Entry: reality capture, current-condition docs, scan-to-BIM, contractor coordination, versioned change records. Not a replacement for 1X Factory OS.
Additional Design Partners
Chef Robotics · Dexterity · Bear Robotics · Collaborative Robotics · Figure — repeated deployment environments needing facility baselines.

Opportunity and threat.

NVIDIA's NuRec workflow can reconstruct a scene from smartphone imagery and export it into a USD-based workflow. But NVIDIA states the reconstructed scene begins as visual geometry without inherent collision properties — additional preparation is required before it becomes useful for physical interaction or robot testing.

Basic capture will become easier and less valuable. Residual value remains in: complete controlled coverage, project-specific accuracy, coordinate and version governance, clean collision geometry, semantic labeling, articulated assets, physics assumptions, scenario definition, validation against the physical site, and secure handling of customer data.

Not a green field — but a specific opening.

Established providers: Hexagon Digital Factory, NavVis, Matterport Capture Services, Cintoo, Siemens, Autodesk, Bentley, Accenture, local surveyors, scan-to-BIM firms, robotics integrators.

The opening is a fast, local, robotics-specific facility-data service that delivers neutral files, documented limitations, deployment-oriented annotations, and recurring updates.

The plan.

Days 1–15
Finalize Site BOM taxonomy. Create intake, capture, evidence, QA, security, limitations templates. Establish partners for higher-accuracy capture, robot safety, OpenUSD/ROS work.
Days 16–30
Capture one representative warehouse, lab, or robot-test environment. Produce the full Robot-Ready Facility Data Pack. Change part of the layout and demonstrate version comparison.
Days 31–60
Approach Agility Field Operations. Approach Robust AI. Approach InOrbit. Contact focused group of OEM, integrator, facility, and industrial-property teams.
Days 61–90
Complete at least one paid pilot. Measure site visits avoided, constraints found, engineering questions answered, update time, files reused. Publish redacted case study. Launch recurring change service.

The facility is part of the robot system.

Building the robot is only half the problem. Every deployment also requires current information about the environment where the robot will work. That's what we capture.

Explore Robotics Facility Readiness →