report.object
Embodied AI infrastructure
Expert-reviewed from public records, platform data, primary sources, and private research beyond AI search. Use /api/v1/reports/RPT_01KZJRA9PR2J9V0EMBAIINFRA26/sample before purchase, then download PDF, JSON, or Markdown after checkout.
page.markdown
Embodied AI infrastructure.md
# Embodied AI infrastructure
Embodied AI infrastructure is not the robot brain; it is the evidence system that turns simulation, data, VLA policies and fleet logs into approvable physical work.
## Object
| field | value |
| ---------------- | ------------------------------- |
| report_id | RPT_01KZJRA9PR2J9V0EMBAIINFRA26 |
| schema_version | 2026-05-28 |
| last_updated_at | 2026-06-05 |
| category | INFRA |
| sector | Cross-industry |
| region | GLOBAL |
| confidence_score | 0.84 |
| source_count | 88 |
| evidence_count | 88 |
## Endpoints
| action | endpoint |
| ------------ | ---------------------------------------------------------- |
| summary | /api/v1/reports/RPT_01KZJRA9PR2J9V0EMBAIINFRA26 |
| sample | /api/v1/reports/RPT_01KZJRA9PR2J9V0EMBAIINFRA26/sample |
| schema | /api/v1/reports/RPT_01KZJRA9PR2J9V0EMBAIINFRA26/schema |
| quote | /api/v1/reports/RPT_01KZJRA9PR2J9V0EMBAIINFRA26/quote |
| purchase | /checkout/RPT_01KZJRA9PR2J9V0EMBAIINFRA26?view=agent |
| versions | /api/v1/reports/RPT_01KZJRA9PR2J9V0EMBAIINFRA26/versions |
| changelog | /api/v1/reports/RPT_01KZJRA9PR2J9V0EMBAIINFRA26/changelog |
| monitor | /api/v1/monitors?report_id=RPT_01KZJRA9PR2J9V0EMBAIINFRA26 |
| mcp_resource | colophon://reports/RPT_01KZJRA9PR2J9V0EMBAIINFRA26/sample |
## Coverage
| field | value |
| -------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| category | INFRA |
| sector | Cross-industry |
| region | GLOBAL |
| topics | robotics, embodied AI, physical AI, ROS 2, simulation, synthetic data, accelerated perception, mapping, manipulation, fleet operations, robot foundation models, NVIDIA Isaac, Physical Intelligence, Skild AI, industrial automation |
| source_classes | open_source_governance, developer_docs, open_source_project, vendor_docs, vendor_release, research_release, research_paper, standard, technical_specification, security_advisory, operator_case_study, customer_case_study, market_statistics, standard_update, benchmark, government_guidance, security_framework, regulation, vendor_case_study, technical_blog |
| update_cadence | Monthly |
## Sample JSON
```json
{
"report_id": "RPT_01KZJRA9PR2J9V0EMBAIINFRA26",
"slug": "embodied-ai-infrastructure",
"schema_version": "2026-05-28",
"title": "Embodied AI infrastructure",
"abstract": "Embodied AI infrastructure is not the robot brain; it is the evidence system that turns simulation, data, VLA policies and fleet logs into approvable physical work.",
"category": "INFRA",
"sector": "Cross-industry",
"region": "GLOBAL",
"topics": [
"robotics",
"embodied AI",
"physical AI",
"ROS 2",
"simulation",
"synthetic data",
"accelerated perception",
"mapping",
"manipulation",
"fleet operations",
"robot foundation models",
"NVIDIA Isaac",
"Physical Intelligence",
"Skild AI",
"industrial automation"
],
"last_updated_at": "2026-06-05T00:00:00Z",
"source_count": 88,
"evidence_count": 88,
"confidence_score": 0.84,
"web_url": "/r/embodied-ai-infrastructure/RPT_01KZJRA9PR2J9V0EMBAIINFRA26",
"sample_endpoint": "/api/v1/reports/RPT_01KZJRA9PR2J9V0EMBAIINFRA26/sample",
"inventory_endpoint": "/api/v1/reports/RPT_01KZJRA9PR2J9V0EMBAIINFRA26/inventory",
"full_endpoint": "/api/v1/reports/RPT_01KZJRA9PR2J9V0EMBAIINFRA26/full",
"schema_endpoint": "/api/v1/reports/RPT_01KZJRA9PR2J9V0EMBAIINFRA26/schema",
"redaction": {
"status": "sample",
"paid_sections_redacted": true,
"preview_limits": [
"summary",
"table_of_contents",
"evidence_profile"
]
},
"coverage": {
"category": "INFRA",
"sector": "Cross-industry",
"region": "GLOBAL",
"topics": [
"robotics",
"embodied AI",
"physical AI",
"ROS 2",
"simulation",
"synthetic data",
"accelerated perception",
"mapping",
"manipulation",
"fleet operations",
"robot foundation models",
"NVIDIA Isaac",
"Physical Intelligence",
"Skild AI",
"industrial automation"
],
"source_classes": [
"open_source_governance",
"developer_docs",
"open_source_project",
"vendor_docs",
"vendor_release",
"research_release",
"research_paper",
"standard",
"technical_specification",
"security_advisory",
"operator_case_study",
"customer_case_study",
"market_statistics",
"standard_update",
"benchmark",
"government_guidance",
"security_framework",
"regulation",
"vendor_case_study",
"technical_blog"
]
},
"confidence": {
"overall_confidence": 0.84,
"source_quality": 0.86,
"source_recency": 0.88,
"source_density": 0.84,
"claim_verifiability": 0.83,
"geographic_coverage": 0.78,
"known_gaps": [
"Public sources rarely publish comparable robot-fleet uptime, incident, recovery or intervention denominators.",
"Robot foundation-model evidence remains strongest in research and benchmark settings rather than audited production workcells.",
"Simulation-platform claims are vendor-controlled unless paired with private field-delta records.",
"Most workcell and fleet case studies do not publish independent ROI, safety or support evidence.",
"Cross-region applicability depends on local labour, safety, cybersecurity, building and sector rules that are outside a global stack map."
]
},
"document": {
"object": "report_document",
"schemaVersion": "2026-05-28",
"title": "Embodied AI infrastructure",
"subtitle": "Simulation, synthetic data, robot learning and foundation-model interfaces for physical work.",
"abstract": "Embodied AI infrastructure is not the robot brain; it is the evidence system that turns simulation, data, VLA policies and fleet logs into approvable physical work.",
"category": "INFRA",
"sector": "Cross-industry",
"region": "GLOBAL",
"topics": [
"robotics",
"embodied AI",
"physical AI",
"ROS 2",
"simulation",
"synthetic data",
"accelerated perception",
"mapping",
"manipulation",
"fleet operations",
"robot foundation models",
"NVIDIA Isaac",
"Physical Intelligence",
"Skild AI",
"industrial automation"
],
"sources": [
{
"id": "S1",
"label": "ROS - REP 2000 releases and target platforms",
"publisher": "Open Robotics",
"sourceClass": "open_source_governance",
"accessedAt": "2026-06-05",
"status": "verified",
"access": "sample",
"originalUrl": "https://docs.ros.org/en/independent/api/rep/html/rep-2000.html",
"transformChain": [],
"url": "https://docs.ros.org/en/independent/api/rep/html/rep-2000.html"
},
{
"id": "S2",
"label": "ROS 2 - DDS implementations",
"publisher": "Open Robotics",
"sourceClass": "developer_docs",
"accessedAt": "2026-06-05",
"status": "verified",
"access": "sample",
"caveat": "Developer documentation; verify version, configuration and deployment behaviour during diligence.",
"originalUrl": "https://docs.ros.org/en/rolling/Installation/RMW-Implementations/DDS-Implementations.html",
"transformChain": [],
"url": "https://docs.ros.org/en/rolling/Installation/RMW-Implementations/DDS-Implementations.html"
},
{
"id": "S3",
"label": "ros2_control - controller manager and hardware abstraction",
"publisher": "ros2_control",
"sourceClass": "developer_docs",
"accessedAt": "2026-06-05",
"status": "verified",
"access": "sample",
"caveat": "Developer documentation; verify version, configuration and deployment behaviour during diligence.",
"originalUrl": "https://control.ros.org/foxy/doc/getting_started/getting_started.html",
"transformChain": [],
"url": "https://control.ros.org/foxy/doc/getting_started/getting_started.html"
},
{
"id": "S4",
"label": "Open-RMF - robot interoperability framework",
"publisher": "Open-RMF",
"sourceClass": "open_source_project",
"accessedAt": "2026-06-05",
"status": "verified",
"access": "sample",
"originalUrl": "https://www.open-rmf.org/",
"transformChain": [],
"url": "https://www.open-rmf.org/"
},
{
"id": "S5",
"label": "NVIDIA Isaac - AI robot development platform",
"publisher": "NVIDIA",
"sourceClass": "vendor_docs",
"accessedAt": "2026-06-05",
"status": "verified",
"access": "sample",
"originalUrl": "https://developer.nvidia.com/isaac/",
"transformChain": [],
"url": "https://developer.nvidia.com/isaac/"
},
{
"id": "S6",
"label": "NVIDIA Isaac Sim - robotics simulation and synthetic data",
"publisher": "NVIDIA",
"sourceClass": "vendor_docs",
"accessedAt": "2026-06-05",
"status": "verified",
"access": "locked",
"originalUrl": "https://developer.nvidia.com/isaac/sim",
"transformChain": [],
"url": "https://developer.nvidia.com/isaac/sim"
},
{
"id": "S8",
"label": "Google DeepMind - Open X-Embodiment",
"publisher": "Google DeepMind",
"sourceClass": "research_release",
"accessedAt": "2026-06-05",
"status": "verified",
"access": "locked",
"originalUrl": "https://deepmind.google/blog/scaling-up-learning-across-many-different-robot-types/",
"publicationDate": "2023-10-03",
"transformChain": [],
"url": "https://deepmind.google/blog/scaling-up-learning-across-many-different-robot-types/"
},
{
"id": "S9",
"label": "OpenVLA - open-source vision-language-action model",
"publisher": "arXiv",
"sourceClass": "research_paper",
"accessedAt": "2026-06-05",
"status": "verified",
"access": "locked",
"originalUrl": "https://arxiv.org/abs/2406.09246",
"publicationDate": "2024-06-13",
"transformChain": [],
"url": "https://arxiv.org/abs/2406.09246"
},
{
"id": "S12",
"label": "Foxglove - robotics observability documentation",
"publisher": "Foxglove",
"sourceClass": "vendor_docs",
"accessedAt": "2026-06-05",
"status": "verified",
"access": "locked",
"originalUrl": "https://docs.foxglove.dev/docs",
"transformChain": [],
"url": "https://docs.foxglove.dev/docs"
},
{
"id": "S18",
"label": "Universal Robots - PolyScope 5 CVE-2026-8153 advisory",
"publisher": "Universal Robots",
"sourceClass": "security_advisory",
"accessedAt": "2026-06-05",
"status": "verified",
"access": "locked",
"originalUrl": "https://www.universal-robots.com/articles/ur/cybersecurity/cve-2026-8153-command-injection-in-the-polyscope-5-dashboard-server/",
"publicationDate": "2026-05-11",
"transformChain": [],
"url": "https://www.universal-robots.com/articles/ur/cybersecurity/cve-2026-8153-command-injection-in-the-polyscope-5-dashboard-server/"
}
],
"evidence": [
{
"id": "E1",
"sourceId": "S1",
"claimIds": [
"C2"
],
"summary": "REP 2000 defines ROS 2 release cadence, LTS support and target-platform rules; Jazzy Jalisco is listed as a May 2024 to May 2029 LTS release, while Kilted Kaiju is a May 2025 to November 2026 non-LTS release.",
"status": "verified",
"confidence": 0.86,
"access": "sample"
},
{
"id": "E2",
"sourceId": "S2",
"claimIds": [
"C2"
],
"summary": "ROS 2 documentation describes DDS as the default communication middleware and lists multiple RMW implementation choices, which makes middleware selection a deployment and support decision rather than a mere developer preference.",
"status": "verified",
"confidence": 0.86,
"access": "sample"
},
{
"id": "E3",
"sourceId": "S3",
"claimIds": [
"C2",
"C3"
],
"summary": "ros2_control describes the Controller Manager as connecting controllers and hardware-abstraction sides, with a Resource Manager abstracting physical hardware and drivers behind state and command interfaces.",
"status": "verified",
"confidence": 0.86,
"access": "sample"
},
{
"id": "E4",
"sourceId": "S4",
"claimIds": [
"C3"
],
"summary": "Open-RMF describes itself as a free, open-source, modular system for sharing and interoperability between robot fleets and physical infrastructure such as doors, elevators and building management systems.",
"status": "verified",
"confidence": 0.86,
"access": "sample"
},
{
"id": "E5",
"sourceId": "S5",
"claimIds": [
"C1",
"C4"
],
"summary": "NVIDIA Isaac presents a workflow that combines synthetic data generation in Isaac Sim and Cosmos Transfer, model training and post-training in Isaac Lab, and deployment on Jetson Orin or Thor.",
"status": "verified",
"confidence": 0.86,
"access": "sample"
},
{
"id": "E6",
"sourceId": "S6",
"claimIds": [
"C4"
],
"summary": "NVIDIA Isaac Sim is described as an open-source reference framework on Omniverse libraries for robotics simulation, testing and synthetic data generation, with CAD, URDF and real-world capture ingestion into USD.",
"status": "verified",
"confidence": 0.86,
"access": "sample"
},
{
"id": "E8",
"sourceId": "S8",
"claimIds": [
"C5"
],
"summary": "Google DeepMind's Open X-Embodiment release says the dataset pooled data from 22 robot types across more than 20 institutions, with more than 500 skills, 150,000 tasks and over one million episodes.",
"status": "verified",
"confidence": 0.86,
"access": "sample"
},
{
"id": "E9",
"sourceId": "S9",
"claimIds": [
"C5"
],
"summary": "The OpenVLA paper introduces a 7B-parameter open-source vision-language-action model trained on about 970,000 real-world robot demonstrations and reports evaluation across 29 tasks and multiple robot embodiments.",
"status": "verified",
"confidence": 0.86,
"access": "sample"
},
{
"id": "E12",
"sourceId": "S12",
"claimIds": [
"C6"
],
"summary": "Foxglove describes a platform to record, upload, organize and visualize multimodal log data, including MCAP and ROS Bag, and to index data by device, time and topic.",
"status": "verified",
"confidence": 0.86,
"access": "sample"
},
{
"id": "E18",
"sourceId": "S18",
"claimIds": [
"C7"
],
"summary": "Universal Robots' May 2026 advisory says PolyScope 5 versions before 5.25.1 are affected by an unauthenticated OS command-injection flaw in Dashboard Server, scored CVSS 3.1 9.8, and recommends upgrading and network isolation.",
"status": "verified",
"confidence": 0.86,
"access": "sample"
}
],
"claims": [
{
"id": "C1",
"statement": "Embodied AI infrastructure is becoming the durable value layer because it connects robot bodies, simulation, synthetic data, policies, mapping, telemetry, safety evidence and site operations.",
"confidence": 0.88,
"evidenceIds": [
"E5"
],
"access": "sample"
},
{
"id": "C2",
"statement": "ROS 2 is a strong portability layer for robotics development, but production teams still need explicit middleware, release, support and hardware-control decisions.",
"confidence": 0.9,
"evidenceIds": [
"E1",
"E2",
"E3"
],
"access": "sample"
},
{
"id": "C3",
"statement": "Hardware abstraction is a control surface: ros2_control, Open-RMF, MassRobotics and OEM SDKs solve different parts of the body, fleet and infrastructure interface.",
"confidence": 0.88,
"evidenceIds": [
"E3",
"E4"
],
"access": "sample"
},
{
"id": "C4",
"statement": "Simulation has moved from design aid to approval evidence because synthetic data, USD scenes, digital twins and sim-to-real tests increasingly carry deployment decisions.",
"confidence": 0.86,
"evidenceIds": [
"E5",
"E6"
],
"access": "sample"
},
{
"id": "C5",
"statement": "Robot foundation models are useful for transfer, but public evidence still points to dataset, embodiment, simulation and validation constraints rather than a universal production brain.",
"confidence": 0.85,
"evidenceIds": [
"E8",
"E9"
],
"access": "sample"
},
{
"id": "C6",
"statement": "Fleet observability and operations software decide production economics because downtime, incidents, remote actions, logs and per-site configuration become the operating denominator.",
"confidence": 0.86,
"evidenceIds": [
"E12"
],
"access": "sample"
},
{
"id": "C7",
"statement": "Safety, cybersecurity and update control must be part of software-stack approval, not a separate compliance memo after autonomy has been selected.",
"confidence": 0.9,
"evidenceIds": [
"E18"
],
"access": "sample"
}
],
"colophon": {
"checkedAt": "2026-06-05",
"latestRevisionAt": "2026-06-05T00:00:00Z",
"reviewStandard": "expert_human_reviewed",
"kinetics": {
"speed": "moderate",
"impact": 0.93,
"momentum": 0.84,
"businessImpact": "Robot procurement is shifting from isolated hardware pilots to embodied-AI approval across simulation, synthetic data, mapping, policies, fleet telemetry, safety and update control.",
"rateOfChange": "Physical-AI model releases and open robot-learning projects can move eval, data-rights and reapproval gates faster than annual robotics cycles.",
"from": "robot programmes approved body by body and site by site",
"toward": "approval paths with simulator provenance, data rights, private evals, fleet telemetry and reapproval triggers"
},
"watch": [],
"gapAnalysis": "Missing public artefacts include comparable production cost, site-level uptime, private safety cases, simulator-fidelity deltas, mapping failure denominators, fleet incident records, model-update regressions and cross-vendor integration-labour evidence."
},
"body": {
"sections": [
{
"id": "executive-summary",
"title": "Executive summary",
"access": "sample",
"dek": "Evidence before autonomy.",
"claimIds": [
"C1",
"C5",
"C6",
"C7"
],
"evidenceIds": [
"E3",
"E5",
"E6",
"E8",
"E9",
"E12",
"E18"
],
"blocks": [
{
"id": "executive-summary-narrative",
"evidenceIds": [
"E3",
"E6",
"E9",
"E12",
"E18"
],
"type": "section_narrative",
"lead": "Embodied AI infrastructure is becoming a production-record problem.",
"body": [
"Robot density shows hardware adoption has broadened. IFR reported 162 robots per 10,000 manufacturing employees globally in 2023, more than double the level seven years earlier [E24]. The scarce operating asset is the record that joins simulated scenes, synthetic data, robot-learning datasets, mapping state, policy updates, fleet telemetry and safety evidence.",
"A body, simulator or foundation model becomes production intelligence only when the buyer can prove site fit, update control and recovery. Recent physical-AI releases now package synthetic-data, simulation-evaluation, VLA-control and orchestration components as infrastructure [E84]. Skild shows omni-bodied policy evidence [E82]. Physical Intelligence's pi0.7 shows steerable model evidence [E88]. Private proof still has to happen on the buyer's robot, site and work.",
"Approve the robot body and OEM interface first [E3, E23]. Bind simulator and mapping state before policy approval [E6, E86]. Keep telemetry, audit logs, update gates and controller exposure in the same release record [E12, E18].",
"The first approval decision is therefore evidence before autonomy: decide which proof must survive a body, simulator, map, model or fleet change before expanding the pilot. The approval gates come before model comparison."
]
},
{
"id": "executive-decision-table",
"title": "Approval gates choose the scale lane",
"caption": "Public sources can identify the proof families; the buyer still has to rerun them on its own body, simulator, model and fleet.",
"evidenceIds": [
"E1",
"E2",
"E3",
"E6",
"E8",
"E9",
"E12",
"E18"
],
"type": "table",
"columns": [
{
"key": "decision",
"label": "Decision",
"align": "left"
},
{
"key": "false_shortcut",
"label": "False shortcut",
"align": "left"
},
{
"key": "proof_to_demand",
"label": "Proof to demand",
"align": "left"
},
{
"key": "buyer_move",
"label": "Buyer move",
"align": "left"
}
],
"rows": [
{
"decision": "Standardize middleware",
"false_shortcut": "Assume ROS 2 support means production portability.",
"proof_to_demand": "Distro support and RMW/DDS choice [E1, E2]; controller timing and hardware-interface ownership [E3].",
"buyer_move": "Approve a supported runtime lane before adding robots."
},
{
"decision": "Use simulation and synthetic data for approval",
"false_shortcut": "Treat a digital twin as the real operating envelope.",
"proof_to_demand": "Scene ingestion and synthetic/eval workflows [E6, E84]; mapping state from the GR00T workflow [E86]. Add physics caveats, failure replay and field-delta logs to the buyer record.",
"buyer_move": "Bind every release to a simulator state and field-retest rule."
},
{
"decision": "Adopt robot foundation models",
"false_shortcut": "Let public transfer results stand in for site readiness.",
"proof_to_demand": "Embodiment overlap, dataset and task coverage [E8, E9]; evaluation fit and a private-fixture test plan before site use.",
"buyer_move": "Pilot behind a narrow production record."
},
{
"decision": "Scale fleet operations",
"false_shortcut": "Count robots deployed rather than incidents absorbed.",
"proof_to_demand": "Telemetry, remote actions, audit logs and operator queue [E12, E13]; patch state and reliability-gap evidence [E18, E21].",
"buyer_move": "Fund RobOps before buying the next body."
}
]
},
{
"id": "executive-approval-sequence",
"title": "Body or model changes reopen approval",
"caption": "A model update, simulator change, map drift, controller exposure or new incident class sends the route back to the relevant question.",
"evidenceIds": [
"E2",
"E3",
"E6",
"E9",
"E12",
"E18"
],
"type": "decision_tree",
"lead": "A body, simulator, map, model or fleet change reopens substrate, evidence-register, telemetry and update-control gates before the route can stay approved.",
"steps": [
{
"prompt": "Can the runtime substrate and controller interface be supported as a production lane? [E2, E23]",
"tone": "neutral",
"branches": [
{
"label": "No",
"outcome": "Keep the build in lab or pilot status until middleware and controller-interface evidence is clear.",
"tone": "neutral"
},
{
"label": "Yes",
"outcome": "Move to simulator and map evidence.",
"next": "Q2 evidence register",
"tone": "neutral"
}
]
},
{
"prompt": "Does the evidence register include simulator state and map state? [E6, E86]",
"tone": "neutral",
"branches": [
{
"label": "No",
"outcome": "Limit the model to eval intake until the production state is versioned.",
"tone": "neutral"
},
{
"label": "Yes",
"outcome": "Move to safety, security and fleet telemetry.",
"next": "Q3 operating gates",
"tone": "neutral"
}
]
},
{
"prompt": "Can fleet telemetry and case evidence prove field recovery, and can update control explain controller exposure? [E12, E21] [E18]",
"tone": "neutral",
"branches": [
{
"label": "No",
"outcome": "Block scale; the route cannot explain or recover from physical exceptions.",
"tone": "neutral"
},
{
"label": "Yes",
"outcome": "Approve a narrow scale lane with named reapproval triggers.",
"tone": "neutral"
}
]
}
]
},
{
"id": "route-definition-callout",
"title": "Route definition",
"evidenceIds": [],
"type": "callout",
"tone": "neutral",
"body": "Route means the approved path from work to accepted output: the source set, data boundary, model or vendor channel, tools, reviewer, evidence record and reapproval trigger that make an AI result usable."
}
]
},
{
"id": "what-changed",
"title": "What changed",
"access": "sample",
"dek": "Hardware adoption broadened; software now carries the reusable evidence.",
"claimIds": [
"C1"
],
"evidenceIds": [
"E5",
"E6",
"E8",
"E9",
"E12"
],
"blocks": [
{
"id": "what-changed-narrative",
"evidenceIds": [
"E5",
"E6",
"E8",
"E9",
"E12"
],
"type": "section_narrative",
"lead": "Robotics is moving from scarce hardware to scarce integration memory.",
"body": [
"Robot adoption has become broad enough for the hard question to move into operations: each facility must operate, monitor, update and defend the system. The public evidence is a spread of software surfaces now competing to own the production record.",
"NVIDIA's Isaac stack makes simulation, synthetic data and model training feel like one accelerated path [E5, E6]. Its GR00T N1.6 workflow adds mapping and localization to the same policy-evidence path [E86]. Open X-Embodiment and DROID make multi-robot data a serious eval input [E8, E10]. OpenVLA and Skild Brain make cross-embodiment transfer plausible, not proven [E9, E82]. Physical Intelligence's pi0.5 adds another evidence line around heterogeneous co-training [E87]. Foxglove and InOrbit make robot operations look more like DevOps [E12, E13]. Formant adds fleet-level provisioning and configuration evidence [E14].",
"The stack has therefore split into two markets. One sells portability: middleware, simulation, data formats, observability and reusable skills. The other sells local certainty: OEM controller access, safety files, site maps, workcell calibration and operator exceptions.",
"The stronger architecture keeps each layer's claim tied to evidence that can survive the next body, site or safety review."
]
},
{
"id": "what-changed-table",
"title": "The stack is splitting into portable and local proof",
"caption": "The buyer should ask which evidence survives a new body, site, simulator or safety revision.",
"evidenceIds": [
"E1",
"E3",
"E4",
"E6",
"E8",
"E9",
"E12",
"E18"
],
"type": "table",
"columns": [
{
"key": "shift",
"label": "Shift",
"align": "left"
},
{
"key": "evidence",
"label": "Source signal",
"align": "left"
},
{
"key": "decision",
"label": "Decision changed",
"align": "left"
}
],
"rows": [
{
"shift": "Release cadence matters",
"evidence": "ROS 2 release support and distro targets are explicit governance artefacts [E1].",
"decision": "Standardize on support windows, not just package availability."
},
{
"shift": "Abstraction has become procurement",
"evidence": "ros2_control and Open-RMF expose different interface boundaries [E3, E4]. MassRobotics and ABB add fleet-status and OEM-controller boundaries [E17, E23].",
"decision": "Buy the interface rights needed for the fleet, not only the robot."
},
{
"shift": "Simulation is now evidence",
"evidence": "Isaac Sim and Intrinsic Flowstate both connect digital workcells to test and deployment workflows [E6, E22].",
"decision": "Require simulator provenance and field-delta records."
},
{
"shift": "Data is moving across embodiments",
"evidence": "Open X-Embodiment and OpenVLA show cross-robot learning, but still through bounded datasets and tasks [E8, E9].",
"decision": "Treat foundation models as candidates for private evals, not production proof."
},
{
"shift": "Update risk is physical",
"evidence": "The 2026 PolyScope advisory tied a controller service to unauthenticated command execution [E18].",
"decision": "Make patch state and network reachability part of stack approval."
}
]
}
]
},
{
"id": "stack-control-frame",
"title": "Stack control frame",
"access": "sample",
"dek": "A layer is portable only when the evidence survives the next robot and the next site.",
"claimIds": [
"C2",
"C3",
"C4"
],
"evidenceIds": [
"E1",
"E2",
"E3",
"E4",
"E6"
],
"blocks": [
{
"id": "stack-control-frame-narrative",
"evidenceIds": [
"E2",
"E3",
"E4"
],
"type": "section_narrative",
"lead": "Buy embodied AI as separable evidence lanes.",
"body": [
"The first split is between software that can be governed as a reusable route and software that is inseparable from a robot body, controller, safety standard or site condition. Middleware may be portable. A calibrated gripper, stair gait, AMR traffic rule or emergency-stop integration is not.",
"ROS 2 is the shared language for many teams, while DDS, domain, lifecycle, RMW, real-time and controller choices still need approval. OEM SDKs often carry the real controller authority and warranty boundaries.",
"The stack-control frame should sit before vendor selection. It makes teams name which evidence travels across bodies, which evidence must be re-run for a new site and which evidence reopens safety or cybersecurity approval.",
"The practical test is simple. If changing the robot body, workcell, lighting, floor, network, payload, operator role or controller firmware would invalidate the claim, the layer is not portable yet."
]
},
{
"id": "stack-control-matrix",
"title": "Production evidence chooses the layer",
"caption": "Cells show the approval work implied by each layer: reusable standard, body or site proof, or formal reapproval before scale.",
"evidenceIds": [
"E2",
"E3",
"E4",
"E6",
"E18"
],
"type": "visual_matrix",
"columns": [
"Portable route",
"Body or site proof",
"Reapproval gate"
],
"rows": [
{
"label": "ROS 2 and middleware",
"note": "Topics, services, actions, DDS and release support.",
"cells": [
{
"label": "Standardize distro and RMW",
"detail": "Useful across teams when support windows and network policy are fixed.",
"tone": "info",
"strength": "high"
},
{
"label": "Test timing and domain rules",
"detail": "DDS and controller jitter still depend on deployment conditions.",
"tone": "warning",
"strength": "medium"
},
{
"label": "Reopen on distro change",
"detail": "LTS moves or RMW swaps need compatibility testing.",
"tone": "warning",
"strength": "gate"
}
]
},
{
"label": "OEM SDKs and controller APIs",
"note": "RobotStudio, PC SDKs, FlexPendant, Robot Web Services and similar surfaces.",
"cells": [
{
"label": "Reusable wrappers",
"detail": "API clients can be standardized.",
"tone": "info",
"strength": "medium"
},
{
"label": "Warranty boundary",
"detail": "Motion authority, controller state and support rights remain OEM-specific.",
"tone": "warning",
"strength": "high"
},
{
"label": "Reopen on firmware",
"detail": "Firmware, safety function or exposed service changes need review.",
"tone": "warning",
"strength": "gate"
}
]
},
{
"label": "Simulation and digital twin",
"note": "Isaac Sim, Flowstate, Gazebo and proprietary twins.",
"cells": [
{
"label": "Scene records travel",
"detail": "USD, URDF and scenario files can support repeatable tests.",
"tone": "info",
"strength": "medium"
},
{
"label": "Physics must be checked",
"detail": "Sensor, contact, lighting and material assumptions are local.",
"tone": "warning",
"strength": "high"
},
{
"label": "Reopen on field delta",
"detail": "Unexpected real-world failures invalidate simulator trust.",
"tone": "warning",
"strength": "gate"
}
]
},
{
"label": "Foundation models and policies",
"note": "OpenVLA, pi0, GR00T-style and task-specific policies.",
"cells": [
{
"label": "Use for candidate skills",
"detail": "Data and model artefacts can start evals faster.",
"tone": "info",
"strength": "medium"
},
{
"label": "Embodiment is local",
"detail": "Gripper, task, camera, latency and fixtures decide pass rate.",
"tone": "warning",
"strength": "high"
},
{
"label": "Reopen on model update",
"detail": "Weights, prompts, policy heads and datasets need change control.",
"tone": "warning",
"strength": "gate"
}
]
},
{
"label": "Fleet operations",
"note": "Logs, incidents, remote actions, maps and configuration.",
"cells": [
{
"label": "Standardize telemetry",
"detail": "Topic, device, event and incident schemas can travel.",
"tone": "info",
"strength": "high"
},
{
"label": "Operator queue is local",
"detail": "Teleoperation, escalation and maintenance depend on site roles.",
"tone": "warning",
"strength": "medium"
},
{
"label": "Reopen on incident class",
"detail": "A new failure mode changes monitoring and response rules.",
"tone": "warning",
"strength": "gate"
}
]
}
],
"legend": [
{
"label": "Reusable route evidence",
"tone": "info",
"strength": "high"
},
{
"label": "Private proof still needed",
"tone": "warning",
"strength": "medium"
},
{
"label": "Clear before scale",
"tone": "warning",
"strength": "gate"
}
]
}
]
},
{
"id": "owner-approval-lanes",
"title": "Owner approval lanes",
"access": "locked",
"dek": "Each layer needs the owner who can stop the rollout.",
"claimIds": [],
"evidenceIds": [],
"blocks": [
{
"id": "owner-approval-lanes-locked",
"title": "Locked section",
"evidenceIds": [],
"type": "callout",
"tone": "locked",
"body": "Purchase or subscribe to unlock this section."
}
]
},
{
"id": "body-site-route-atlas",
"title": "Body and site route atlas",
"access": "locked",
"dek": "The body decides which proof cannot be reused.",
"claimIds": [],
"evidenceIds": [],
"blocks": [
{
"id": "body-site-route-atlas-locked",
"title": "Locked section",
"evidenceIds": [],
"type": "callout",
"tone": "locked",
"body": "Purchase or subscribe to unlock this section."
}
]
},
{
"id": "layer-build-buy",
"title": "Build, buy and standardize",
"access": "locked",
"dek": "The useful stack map is a procurement map.",
"claimIds": [],
"evidenceIds": [],
"blocks": [
{
"id": "layer-build-buy-locked",
"title": "Locked section",
"evidenceIds": [],
"type": "callout",
"tone": "locked",
"body": "Purchase or subscribe to unlock this section."
}
]
},
{
"id": "accepted-output-economics",
"title": "Accepted-output economics",
"access": "locked",
"dek": "Count accepted robot-hours and workcell output, not licences alone.",
"claimIds": [],
"evidenceIds": [],
"blocks": [
{
"id": "accepted-output-economics-locked",
"title": "Locked section",
"evidenceIds": [],
"type": "callout",
"tone": "locked",
"body": "Purchase or subscribe to unlock this section."
}
]
},
{
"id": "private-eval-design",
"title": "Private eval design",
"access": "locked",
"dek": "Benchmarks choose tests; private fixtures choose production.",
"claimIds": [],
"evidenceIds": [],
"blocks": [
{
"id": "private-eval-design-locked",
"title": "Locked section",
"evidenceIds": [],
"type": "callout",
"tone": "locked",
"body": "Purchase or subscribe to unlock this section."
}
]
},
{
"id": "simulation-data-flywheel",
"title": "Simulation and data flywheel",
"access": "locked",
"dek": "A simulator is valuable only when it improves the next field deployment.",
"claimIds": [],
"evidenceIds": [],
"blocks": [
{
"id": "simulation-data-flywheel-locked",
"title": "Locked section",
"evidenceIds": [],
"type": "callout",
"tone": "locked",
"body": "Purchase or subscribe to unlock this section."
}
]
},
{
"id": "foundation-models",
"title": "Foundation models and validation",
"access": "locked",
"dek": "Transfer evidence is not production acceptance.",
"claimIds": [],
"evidenceIds": [],
"blocks": [
{
"id": "foundation-models-locked",
"title": "Locked section",
"evidenceIds": [],
"type": "callout",
"tone": "locked",
"body": "Purchase or subscribe to unlock this section."
}
]
},
{
"id": "fleet-operations",
"title": "Fleet operations",
"access": "locked",
"dek": "RobOps is where pilots become production systems.",
"claimIds": [],
"evidenceIds": [],
"blocks": [
{
"id": "fleet-operations-locked",
"title": "Locked section",
"evidenceIds": [],
"type": "callout",
"tone": "locked",
"body": "Purchase or subscribe to unlock this section."
}
]
},
{
"id": "safety-cyber-update-control",
"title": "Safety, cybersecurity and update control",
"access": "locked",
"dek": "Robot software risk is physical when it reaches the controller.",
"claimIds": [],
"evidenceIds": [],
"blocks": [
{
"id": "safety-cyber-update-control-locked",
"title": "Locked section",
"evidenceIds": [],
"type": "callout",
"tone": "locked",
"body": "Purchase or subscribe to unlock this section."
}
]
},
{
"id": "procurement-diligence",
"title": "Procurement diligence",
"access": "locked",
"dek": "The contract should buy evidence export, not only software access.",
"claimIds": [],
"evidenceIds": [],
"blocks": [
{
"id": "procurement-diligence-locked",
"title": "Locked section",
"evidenceIds": [],
"type": "callout",
"tone": "locked",
"body": "Purchase or subscribe to unlock this section."
}
]
},
{
"id": "recovery-controls",
"title": "Recovery controls",
"access": "locked",
"dek": "An embodied-AI stack is approved only when recovery is rehearsed.",
"claimIds": [],
"evidenceIds": [],
"blocks": [
{
"id": "recovery-controls-locked",
"title": "Locked section",
"evidenceIds": [],
"type": "callout",
"tone": "locked",
"body": "Purchase or subscribe to unlock this section."
}
]
},
{
"id": "deployment-economics",
"title": "Deployment burden",
"access": "locked",
"dek": "The bill hides in rollout work, support and recovery.",
"claimIds": [],
"evidenceIds": [],
"blocks": [
{
"id": "deployment-economics-locked",
"title": "Locked section",
"evidenceIds": [],
"type": "callout",
"tone": "locked",
"body": "Purchase or subscribe to unlock this section."
}
]
},
{
"id": "case-evidence",
"title": "Case evidence",
"access": "locked",
"dek": "Public cases show where the software layer earns its keep.",
"claimIds": [],
"evidenceIds": [],
"blocks": [
{
"id": "case-evidence-locked",
"title": "Locked section",
"evidenceIds": [],
"type": "callout",
"tone": "locked",
"body": "Purchase or subscribe to unlock this section."
}
]
},
{
"id": "vendor-directory",
"title": "Vendor and stack directory",
"access": "locked",
"dek": "Representative software layers to include in an embodied-AI shortlist.",
"claimIds": [],
"evidenceIds": [],
"blocks": [
{
"id": "vendor-directory-locked",
"title": "Locked section",
"evidenceIds": [],
"type": "callout",
"tone": "locked",
"body": "Purchase or subscribe to unlock this section."
}
]
},
{
"id": "operating-model",
"title": "Operating model",
"access": "locked",
"dek": "How to stand up an embodied-AI approval process.",
"claimIds": [],
"evidenceIds": [],
"blocks": [
{
"id": "operating-model-locked",
"title": "Locked section",
"evidenceIds": [],
"type": "callout",
"tone": "locked",
"body": "Purchase or subscribe to unlock this section."
}
]
},
{
"id": "known-unknowns",
"title": "Known unknowns",
"access": "locked",
"dek": "The public evidence gaps that should shape private diligence.",
"claimIds": [],
"evidenceIds": [],
"blocks": [
{
"id": "known-unknowns-locked",
"title": "Locked section",
"evidenceIds": [],
"type": "callout",
"tone": "locked",
"body": "Purchase or subscribe to unlock this section."
}
]
},
{
"id": "watch-and-reapproval",
"title": "Standing reapproval triggers",
"access": "locked",
"dek": "Internal tripwires that reopen an approved lane.",
"claimIds": [],
"evidenceIds": [],
"blocks": [
{
"id": "watch-and-reapproval-locked",
"title": "Locked section",
"evidenceIds": [],
"type": "callout",
"tone": "locked",
"body": "Purchase or subscribe to unlock this section."
}
]
},
{
"id": "source-ledger",
"title": "Source ledger",
"access": "locked",
"dek": "Primary sources and diligence anchors behind the analysis.",
"claimIds": [],
"evidenceIds": [],
"blocks": [
{
"id": "source-ledger-locked",
"title": "Locked section",
"evidenceIds": [],
"type": "callout",
"tone": "locked",
"body": "Purchase or subscribe to unlock this section."
}
]
},
{
"id": "glossary",
"title": "Glossary",
"access": "sample",
"claimIds": [],
"evidenceIds": [],
"blocks": [
{
"id": "glossary-table",
"title": "Terms used in this report",
"caption": "Definitions are reader aids, not separate evidence claims.",
"evidenceIds": [],
"type": "table",
"columns": [
{
"key": "term",
"label": "Term",
"align": "left"
},
{
"key": "meaning",
"label": "Plain meaning",
"align": "left"
}
],
"rows": [
{
"term": "Route",
"meaning": "The approved path from work to accepted output: source set, data boundary, model or vendor channel, tools, reviewer, evidence record and reapproval trigger."
},
{
"term": "API",
"meaning": "Application programming interface: a governed software connection used by an app or workflow to call a model or service."
},
{
"term": "SDK",
"meaning": "Software development kit: packaged code and tools developers use to connect a product to an API or service."
}
]
}
]
}
]
}
},
"gap_analysis": "Missing public artefacts include comparable production cost, site-level uptime, private safety cases, simulator-fidelity deltas, mapping failure denominators, fleet incident records, model-update regressions and cross-vendor integration-labour evidence.",
"kinetics": {
"speed": "moderate",
"impact": 0.93,
"momentum": 0.84,
"businessImpact": "Robot procurement is shifting from isolated hardware pilots to embodied-AI approval across simulation, synthetic data, mapping, policies, fleet telemetry, safety and update control.",
"rateOfChange": "Physical-AI model releases and open robot-learning projects can move eval, data-rights and reapproval gates faster than annual robotics cycles.",
"from": "robot programmes approved body by body and site by site",
"toward": "approval paths with simulator provenance, data rights, private evals, fleet telemetry and reapproval triggers"
},
"watch_signals": [],
"sample_findings": [
"The durable value layer is the production record that connects robot bodies, simulators, synthetic data, mapping, policies, telemetry, safety files and site exceptions."
]
}
```