What's the difference between robotics brain providers and robot makers?

The three-tier robotics value chain framework

The robotics industry in 2026 operates across three structurally distinct tiers. Naming these tiers explicitly is editorially substantive because conflating them produces misframed competitive analysis. A reader asking whether Figure AI competes with Skild AI is asking a question that the three-tier framework clarifies: they operate at different tiers, not as direct head-to-head competitors.

The three tiers:

1. Brain-provider tier: companies building foundation models for robotics without making hardware. Examples: Skild AI, Physical Intelligence, Covariant, Google DeepMind, OpenAI Robotics, NVIDIA Project GR00T, Meta AI.
2. OEM-platform tier: companies building robot hardware platforms with integrated brains. Examples: Figure AI, Apptronik, 1X Technologies, Tesla, Agility Robotics, Boston Dynamics, Unitree, UBTech, Sanctuary AI, PAL Robotics.
3. Deployment tier: real-world operation at customer facilities producing the verification outcomes that prove or disprove capability claims. Examples: BMW Spartanburg (Figure 02 chassis assembly), GXO Flowery Branch (Agility Digit 100,000-tote throughput), Mercedes-Benz Berlin-Marienfelde and Hungarian plant (Apptronik Apollo), Catalyst Brands Reno (Figure 03 logistics).

Per DEPLOY's verified-vs-claimed framework, each tier operates against different verification surfaces. Brain-provider tier verification operates on research-and-demonstration scale plus integration-partner depth. OEM-platform tier verification operates on the five-tier availability framework plus four-tier capability framework. Deployment tier verification operates on customer-facility throughput plus contractual outcomes.

What each tier competes on

The competitive dynamic per tier is structurally different:

Brain-provider tier competes on model architecture, training data quality, benchmark performance, integration-partner depth, and (eventually) commercial-scale deployment outcomes. The competitive question is whether the brain transfers across robot platforms reliably and whether OEMs license brain-provider models versus building internal brains.

OEM-platform tier competes on form factor (humanoid vs quadruped vs specialized industrial), customer relationships (enterprise contracts; consumer commerce), deployment scope (single-customer depth vs multi-customer breadth), pricing structure (enterprise-confidential vs consumer-direct vs research-tools), and verification posture (verified consumer-deployed vs verified enterprise-deployed vs research-and-demonstration vs claimed future per DEPLOY's frameworks).

Deployment tier is not a competitive arena per se; it is the verification surface that brain-provider tier and OEM-platform tier outcomes ultimately resolve at. BMW Spartanburg is a Figure verification anchor. GXO Flowery Branch is an Agility verification anchor. The deployment tier customer is not competing with anyone; the customer is the verification surface that the upstream tiers compete to produce favorable outcomes at.

Cross-tier competition and collaboration

The three tiers operate complementarily on verification surfaces but compete and collaborate at strategic levels:

Some OEMs build their own brains rather than licensing brain-provider models. Tesla operates an internal autonomy stack across Robotaxi (vision-only FSD) and Optimus (Tesla-developed humanoid AI). 1X Technologies operates internal brain development with explicit teleop disclosure as part of the data acquisition layer per the NEO teleop explainer. The internal-brain pattern means brain-provider tier value is constrained by which OEMs choose to license vs build.

Some OEMs license brain-provider models rather than building internal brains. The verification depth on which OEMs license what brains varies per disclosure depth. Frontier AI labs entering robotics is partly about this question: which OEMs integrate Google DeepMind's Gemini Robotics, OpenAI Robotics models, or peer brain-provider stacks rather than developing internal alternatives.

Brain-provider tier value depends on which path dominates. If OEMs commoditize their own brain development, brain-provider tier value decreases. If OEMs license brain-provider models broadly, brain-provider tier value increases. The outcome is unsettled; framework readings should preserve this strategic uncertainty rather than collapsing into either-or claims.

Why deployment tier outcomes matter most

The framework's verification discipline anchors at the deployment tier because customer-facility outcomes resolve the upstream tier claims. A brain-provider tier model that performs well on academic benchmarks but fails at customer-facility deployment is at lower verification depth than a brain-provider tier model with verified customer-facility integration. An OEM-platform tier company with consumer-promised pricing but no deployed customers is at lower verification depth than an OEM-platform tier company with multi-year multi-customer commercial deployments.

The deployment tier verification anchor explains why DEPLOY's framework operates the five-tier availability framework and four-tier capability framework at OEM-platform tier specifically: that is where consumer-evaluation outcomes accumulate. Brain-provider tier outcomes anchor through OEM partnerships at customer facilities; brain-provider tier evaluation is one step removed from the deployment tier verification anchor.

What the framework cap-flags across tiers

Per the framework's cap-flag application, each tier carries distinct verification gaps:

• Brain-provider tier: per-model commercial-deployment counts; per-partner integration depth; cross-platform transfer at commercial scale; financial sustainability of brain-provider business models.
• OEM-platform tier: per-unit pricing for enterprise-deployed cohort; multi-customer scaled-throughput breadth; consumer-deployment timeline for consumer-promised tier (Tesla Optimus); long-term operational reliability data.
• Deployment tier: long-horizon customer-relationship outcomes; per-vertical generalization; financial-sustainability of robots-as-a-service contracting structures.

The cap-flags are editorial truths per DEPLOY's framework, not gaps. Each tier operates at the verification depth it operates at; the cap-flag tier surfaces the boundaries explicitly.

Where to go for context

For the foundation-model-for-robotics category that brain-provider tier companies operate within, see what is a foundation model for robotics. For Skild AI as a brain-provider tier exemplar, see what is Skild AI. For the broader brain-provider landscape comparison, see brain-provider landscape comparison.

For the OEM-platform tier across humanoid manufacturers, see the leading humanoid robot makers, can I buy a humanoid robot in 2026, and what can humanoid robots actually do today.

For deployment-tier worked examples, see the foundational signals anchoring customer-facility outcomes: Figure 02 at BMW Spartanburg 30,000 vehicles, Agility Digit at GXO Flowery Branch 100,000 totes, Apptronik Apollo enterprise pilots at Mercedes-Benz GXO Jabil, and OpenAI Robotics relaunch. For methodology canonical references applicable to brain-provider vs OEM platform tier distinction: captive vs third-party brain providers (CANONICAL tier distinction worked example) + verified-vs-claimed at within-entity granularity (platform tier vs brain tier verification posture distinct) + the 9-tier source-quality rubric.