How The Sector Breaks Down

Eight companies, four sub-segments. The cohort splits cleanly on what each company actually owns as its primary asset — humanoid robot hardware plus an embedded AI policy, a hardware-agnostic robotics foundation model, an end-to-end driving or trucking AI stack with a named commercial surface, or the simulation-and-tooling substrate that any autonomous-systems developer buys. Within each sub-segment the competitive dynamic is distinct; across sub-segments the companies rarely compete head-to-head on the same buyer.

Humanoid robotics — Figure AI, 1X Technologies, Apptronik. The core asset is a vertically-integrated bipedal humanoid platform combined with an in-house or strategic-partner AI policy stack. Figure AI runs the Helix vision-language-action model in-house since the February 2025 OpenAI exit, with BMW Spartanburg as the headline industrial-production reference. 1X runs an in-house general-purpose AI policy under Eric Jang (ex-Google Brain) with NEO targeting consumer-home deployment at $20,000 and the December 2025 EQT 10,000-robot commercial partnership as a parallel channel. Apptronik runs Apollo against a commercial-pilot lattice (Mercedes-Benz, GXO Logistics, John Deere, Jabil) with Google as a Series A extension co-lead anchoring the AI-policy strategic relationship. All three sit on the same competitive frame against Tesla Optimus and Boston Dynamics.

Robotics foundation models — Physical Intelligence, Skild AI. The core asset is a cross-embodiment foundation model designed to ship across diverse robot hardware rather than to a single vertically-integrated platform. Physical Intelligence runs the π0 / π0.5 generalist-policy line on a research-team substrate drawn from Google DeepMind, UC Berkeley and Stanford; Skild AI runs the “Skild Brain” line on a CMU Robotics Institute substrate with the “omni-bodied” positioning, deployed on Foxconn × NVIDIA Blackwell lines plus the ABB / Universal Robots / Mobile Industrial Robots platform-partnership lattice. Both compete head-to-head on the foundation-model layer; both flank the humanoid players from above.

Autonomous driving and trucking — Wayve, Waabi. The core asset is an end-to-end learned driving or trucking AI deployed to a named commercial surface. Wayve ships the AV2.0 end-to-end neural-network “AI Driver” foundation model to Nissan (ProPILOT FY2027), Mercedes-Benz and Stellantis on the OEM-production track, and to Uber for the London 2026 and Tokyo late-2026 robotaxi pilots; the silicon supplier base spans NVIDIA, Qualcomm, AMD and Arm after the March 2026 Series D extension. Waabi runs the Waabi Driver simulation-first generative-AI stack on Class-8 commercial trucking through Uber Freight, with Lior Ron (ex-Uber Freight founder-CEO) as COO from August 2025 and the Volvo VNL Autonomous platform as the lead OEM-equipment partnership.

Autonomous-systems tooling — Applied Intuition. The core asset is the simulation, validation, vehicle-software-development and autonomy-specific foundation-model substrate that any autonomous-systems developer buys. Applied Intuition is the picks-and-shovels reference point in the cohort — not a humanoid platform, not a robotics foundation-model lab, not an end-to-end driving stack, but the toolchain that 18 of the top 20 global OEMs (Toyota, Volkswagen, GM, Porsche, Stellantis, Audi, Hyundai, Isuzu, TRATON) plus the US Department of Defense (Army RCV, CDAO production, Army BPA) buy to develop autonomous systems on top of. The placement is deliberate: Applied Intuition’s buyers include every other sub-segment’s buyers (OEM software organisations, defence-prime integrators, robotics-platform engineering teams) and its competitive set sits against NVIDIA DRIVE, dSPACE and AVL rather than against Figure, Wayve or Skild.

The four sub-segments are mutually exclusive across the cohort. A humanoid-platform vendor (Figure, 1X, Apptronik) does not compete head-to-head with a robotics foundation-model lab (Physical Intelligence, Skild AI) on the same buyer — the humanoid vendors are vertically-integrated against industrial-deployment customers and the foundation-model labs ship cross-platform AI; the two flank each other from above and below. An autonomous-driving stack (Wayve) does not compete with a humanoid platform on the OEM design-win or robotaxi distribution surfaces. And Applied Intuition’s autonomy-tooling platform sells to all of them as customers and prospects rather than competing with them on outputs.

The Information Matters Compass — Robotics AI Sector

The Information Matters Compass plots every covered company on two axes — how defensible the business looks (left–right) and how much disruption potential it carries (bottom–top). The dashed lines at 7.5 split the chart into four equal quadrants. The robotics-AI cohort sits almost entirely below the Defensibility bar: most of the cohort is pre-revenue or thinly revenue-bearing, the deployment unit-economics are unproven at scale across every sub-segment, and the regulatory perimeter for embodied AI in industrial, consumer, and on-road settings is still consolidating. Disruption Potential is where the variation runs — led by Skild AI (the highest Disruption composite in the cohort on the strength of the deployed Foxconn deployment plus the strategic-investor lattice), Wayve (the deepest OEM and ride-hail distribution book) and Physical Intelligence (the cross-embodiment research signal).

The 8 Companies

Defensibility and Disruption are scored 0–10; Overall is the weighted combination. The numbers are Information Matters’ assessments, applied consistently across the cohort, and audited before publication.

What This Tells Us About Robotics AI In 2026

Robotics is the longest-horizon and most-capital-intensive segment in the IM universe, and the cohort is largely pre-revenue. Of the eight covered companies, only Applied Intuition (~$830M ARR per Sacra coverage) and Skild AI (~$30M annualised through 2025 per named-press triangulation) have published meaningful revenue signals; Figure, 1X, Apptronik, Physical Intelligence, Waabi and Wayve are all either pre-commercial or on commercial-pilot economics that have not yet been disclosed as recurring revenue. Cumulative external capital across the cohort is materially north of $13B — with Figure ($1.9B), Wayve ($2.5B+), Skild ($1.83B+), Physical Intelligence ($1.1B+), Applied Intuition ($1.2B+) and Apptronik ($935M+) each above the $900M mark — against a revenue base that is structurally years from converging on the implied valuations. No company has reached the top-right Dominant Innovator quadrant. The proximate reason is consistent across sub-segments: the moat in robotics sits with the deployed-at-scale physical asset, not with the model release; the deployed-at-scale physical asset takes years to compound; and the regulatory clock binds directly through factory-floor safety standards, AV approval pathways, consumer-product liability and defence-export controls. Readers should not collapse model-capability release cadence into commercial-deployment compounding.

The humanoid-vs-foundation-model question is itself a real strategic question, and the cohort is currently testing both answers. Figure, 1X and Apptronik are betting that vertically-integrated humanoid hardware plus an in-house or strategic-partner AI policy stack is the path to scale — Helix in-house at Figure since February 2025, in-house general-purpose policy at 1X under Eric Jang, Google-anchored AI-policy strategy at Apptronik. Physical Intelligence and Skild AI are betting on the inverse: that a cross-embodiment foundation model trained on diverse robot data ships across humanoid, mobile-manipulation and industrial-robotics platforms faster than any single vertically-integrated stack can compound. The strategic question over the 2026-2028 window is whether robots need their own purpose-built foundation models, or whether they ride on the cross-embodiment generalist substrate that Physical Intelligence and Skild are training, or whether they ride on the horizontal foundation-model providers (Google Gemini Robotics, OpenAI VLA work, NVIDIA Isaac GR00T) that operate above all of them. The cohort does not yet support a confident analytical call on which answer wins; the two robotics-foundation-model labs and the three vertically-integrated humanoid platforms in the cohort are each running a credible test of their respective theses.

Three reader takes by audience. For industrial buyers and corp-dev: read the sub-segments as four distinct procurement processes with four distinct buying centres — humanoid-platform commercial pilots (Mercedes-Benz, GXO Logistics, John Deere, Jabil, BMW, EQT-portfolio companies), robotics-foundation-model platform-partnership procurement (ABB Robotics, Universal Robots, Mobile Industrial Robots), autonomous-driving OEM design-wins plus ride-hail operator partnerships (Nissan, Mercedes-Benz, Stellantis, Uber), and autonomy-tooling enterprise-software contracts (Applied Intuition into 18-of-top-20 OEMs and the Pentagon). For investors: the valuation benchmarks span an unusually wide range — from Apptronik at $5B and Physical Intelligence at $5.6B (the in-cohort discount tier) through Skild at $14B and Applied Intuition at $15B (the deployed-traction tier) up to Figure at $39B and Wayve at $8.6B (the frontier-and-distribution premium tier). The Watchlist mechanism is appropriate for any vendor currently lacking disclosed commercial traction; the IM Framework Compass below treats every covered name as Disruptive Challenger or Emerging Player, with no top-tier label asserted in this cohort. For sector strategists: the watch-list is whether any cohort member converts a model-capability lead into a procurement-grade industrial-deployment ramp inside the next 18 months — the first vendor to do so will likely reset the Defensibility composite for its sub-segment.

Watch lines over the next 90 days. Five signals would shift the picture. First, the Apptronik post-Series-A-extension Apollo production scale-up against the commercial-pilot ramp at Mercedes-Benz, GXO Logistics, John Deere and Jabil — the only humanoid cohort member with a disclosed multi-customer commercial-partner lattice. Second, the Figure BotQ factory ramp from the disclosed ~240 units/month in April 2026 toward the stated 12,000-robot first-line annual capacity, plus the cadence of consumer-pilot disclosures Brett Adcock has signalled. Third, the Physical Intelligence follow-on model-release cadence (π0 in October 2024, π0.5 in April 2025; a next-generation release through 2026 would compress the foundation-model gap to Skild and to horizontal frontier labs) plus any movement on the reported $1B round at $11B valuation. Fourth, the Wayve London robotaxi-trial launch in 2026 with Uber, the Tokyo pilot in late 2026 (subject to Japanese regulatory approval) and the Nissan ProPILOT FY2027 production start — the three commercial-revenue conversion milestones embedded in the $8.6B Series D valuation framing. Fifth, the regulatory cadence: the November 9 2026 CMMC Level 2 compliance deadline binding on Applied Intuition’s DoD-contractor revenue line, the EU AI Act high-risk-AI obligations on industrial-AI deployments coming into force August 2 2026 (binding on European Apptronik and Wayve customers), and any high-profile safety incident in commercial humanoid, robotaxi or autonomous-trucking deployments.

The Verticals, Frontier and Rewire futures dominate the structural overlay. The Information Matters Framework names eight futures — the patterns the next ten years of AI will resolve into. The robotics-AI cohort is most legible under three. Verticals captures the humanoid platforms (Figure, 1X, Apptronik), the autonomous-driving and trucking stacks (Wayve, Waabi) and the autonomy-tooling specialist (Applied Intuition) — each betting that vertical depth on a defined physical-world deployment surface beats horizontal foundation-model capability at the workload level. Frontier captures the robotics foundation-model labs (Physical Intelligence, Skild AI) and Wayve’s end-to-end neural-network AV2.0 thesis — each betting that release-by-release capability margin on the cross-embodiment or end-to-end-learning substrate is where the disruption composite is earned. Rewire sits behind the cohort as the long-horizon overlay — industrial labour markets reorganising around humanoid robots, freight reorganising around driverless trucks, OEM and DoD software organisations reorganising around vendor-supplied autonomy platforms — on a 3-7 year horizon that no cohort member has yet validated at scale. Trust shows up as the rate-limiter rather than the differentiator: the regulatory perimeter (factory-floor safety standards, AV approval pathways, CMMC Level 2, EU AI Act and Machinery Regulation, ITAR / EAR) caps how aggressively each cohort member can move without resetting it as a moat.

How To Read These Scores

Every company is scored on nine plain-English dimensions. Defensibility covers how sticky the customers are, what proprietary knowledge or data the company holds, the strength of its distribution channels, its strategic resilience to shocks, and whether it benefits from platform-style network effects. Disruption Potential covers momentum, how novel the capability is, how fast the team executes, and how much category leadership the company commands. Each dimension is scored from 0 to 10. A sector-appropriate weighting produces the Defensibility and Disruption composites that drive the Compass position.

The competitive position labels — Dominant Innovator, Disruptive Challenger, Established Incumbent, Emerging Player — come from where the composites place a company on the Compass, not from analyst judgment. A separate Wound-Down label is used for residual entities post-acquisition or wind-down; no vendors in this cohort carry that status. For the full methodology, including how each dimension is broken down further, see the Information Matters Framework Scoring methodology. Every score on this page has been through Information Matters’ two-layer audit before publication.