Run metadata:

• ran_at: 2026-09-25T08:01:54+03:00
• job_id: physical_ai_model_to_matter
• task: Task 1 — Physical AI / Embodiment Scan + Task 2 — Ethan Portfolio Engine Report
• scope: physical AI / embodiment / robotics / bionics / AI materials / BCI / defense robotics / capital-flow signals
• verification: web-current, sources fetched in this run only
• output intent: full report for sol/reports/2026…REDACTED.md; compact index for sol/rooms/portfolio.md

---

EXECUTIVE SIGNAL

Physical AI moved further from “demo race” into infrastructure race: compute, data pipelines, deployment hours, manufacturing lines, autonomous labs, and defense procurement are now stronger signals than viral humanoid videos.

Highest-conviction signals this run:

1. Robot-brain/data/compute layer is becoming the scarce asset. Figure signed a strategic compute partnership with Nscale for up to 100,000 NVIDIA Vera Rubin GPUs, described by Figure as an initial $3.5B compute commitment with intent to scale above $6B; Figure says this is needed because Helix is now data-and-compute bound. URL: https://www.figure.ai/news/figu…REDACTED (figure.ai)
2. Deployment claims are becoming more measurable but still uneven. Figure reports Figure 03 at BMW Spartanburg and a Catalyst Brands logistics agreement; Skild says it crossed $100M ARR ten months after first commercial deployment; Vecna raised $31M to scale AMR deployment teams. URLs: https://www.figure.ai/news/f-03-at-bmw, https://www.figure.ai/news/figu…REDACTED, https://skild.ai/blogs/skil…REDACTED, https://www.vecnarobotics.com/news/vecn…REDACTED/ (figure.ai)
3. Defense autonomy is converting faster than consumer humanoids. AeroVironment announced a $464.8M U.S. Army counter-UAS contract, Shield AI was selected to compete for up to $800M in U.S. Navy ISR task orders, American Rheinmetall received a $7.28M U.S. Marine Corps Mission Master contract, and Sea Machines secured a five-year SOF IDIQ for autonomy kits. URLs: https://www.avinc.com/2026/09/02/avs-…REDACTED/, https://shield.ai/shie…REDACTED/, https://www.rheinmetall.com/en/media/news-watch/news/2026/09/2026…REDACTED, https://sea-machines.com/sea-…REDACTED/ (avinc.com)
4. BCI/bionics capital flow is accelerating at the human-machine boundary. Precision Neuroscience closed a $250M Series D on September 24, 2026, bringing total capital raised to $430M and reporting more than 100 patient procedures across 18 institutions; BIOS/University of Utah announced an at-home LUKE Arm trial using a peripheral nerve interface for control and touch. URLs: https://www.globenewswire.com/news-release/2026/09/24/3368254/0/en/prec…REDACTED.html, https://www.biologicinputoutputsystems.com/media/neur…REDACTED (globenewswire.com)
5. AI materials/autonomous labs are crossing from academic tools to national infrastructure. NSF announced a $400M national network of AI-programmable cloud laboratories; Rice, Texas A&M, UChicago/Argonne, and Johns Hopkins each announced major NSF-backed autonomous/robotic lab nodes. URLs: https://www.nsf.gov/tip/updates/nsf-…REDACTED, https://news.rice.edu/news/2026/acce…REDACTED, https://stories.tamu.edu/news/2026/08/05/texa…REDACTED/, https://news.uchicago.edu/story/ai-p…REDACTED, https://engineering.jhu.edu/news/buil…REDACTED/ (nsf.gov)

---

SIGNAL_LAYER

Humanoid ecosystem: production + deployment claims separating from demos

Figure AI is the strongest current humanoid signal this run because its announcements connect compute, data, production, and commercial deployment. Figure reported: a strategic Nscale partnership for up to 100,000 NVIDIA Vera Rubin GPUs; Helix 2.5 zero-shot home generalization across 30 homes; a Catalyst Brands logistics deployment agreement; Figure 03 arrival at BMW Spartanburg; and BotQ production ramp from one Figure 03 per day to one per hour after delivering more than 350 third-generation units. These are company-reported claims, not independently audited deployment/revenue data, but they are more portfolio-relevant than isolated demos because they map to the full stack: data → compute → model → factory → customer site. URLs: https://www.figure.ai/news/figu…REDACTED, https://www.figure.ai/news/heli…REDACTED, https://www.figure.ai/news/figu…REDACTED, https://www.figure.ai/news/f-03-at-bmw, https://www.figure.ai/news/ramp…REDACTED (figure.ai)

Tesla / Optimus remains investable only through Tesla equity, but the current signal is still “pre-revenue capex/R&D option,” not validated robot revenue. Tesla’s Q2 2026 materials and SEC filing say it is preparing and investing for large-scale Optimus production; third-party reporting on the Q2 update says Fremont Model S/X lines were being converted for Optimus production and that capex/R&D spending is rising around Optimus, robotaxi, AI compute, Cybercab, and other initiatives. URL: https://www.sec.gov/Archives/edgar/data/1318605/000162828026049270/tsla-20260630.htm, https://www.fool.com/earnings/call-transcripts/2026/08/05/tesl…REDACTED/, https://apnews.com/article/976c…REDACTED (sec.gov)

Boston Dynamics / Hyundai opened a dedicated humanoid testing ground at Hyundai’s Georgia Metaplant, indicating a shift toward structured industrial validation environments rather than pure showcase robotics. The signal is deployment-preparation, not commercial scale. URL: https://www.axios.com/local/boston/2026/09/21/bost…REDACTED (axios.com)

Unitree produced a major public-market signal: AP reported that Unitree shares initially surged as much as 629% in its Shanghai STAR Market debut after raising about 6.1B yuan / $904M, making it the first publicly traded mainland Chinese humanoid robotics maker. This is a capital-flow/exuberance signal, not necessarily a validated deployment-profit signal. URL: https://apnews.com/article/f33f…REDACTED (apnews.com)

Agility Robotics remains a warehouse-deployment signal: AP reported the planned SPAC transaction at a $2.5B valuation and described Digit as commercially operational in warehouse/industrial facilities. This remains closer to near-term deployment than general home humanoids. URL: https://apnews.com/article/39f2…REDACTED (apnews.com)

---

Robot-brain layer / portable embodied cognition

Skild AI is the clearest “robot brain” signal. NVIDIA reported on September 10, 2026 that Skild’s S1 uses NVIDIA infrastructure, synthetic data generation, simulation, model training, and real-world deployment; Skild separately claimed it crossed $100M ARR ten months after first commercial deployment. Skild’s collaborations with ABB Robotics, Universal Robots, MiR, and NVIDIA point toward a hardware-agnostic embodied intelligence layer across industrial robots, AMRs, and humanoids. URLs: https://blogs.nvidia.com/blog/skil…REDACTED/, https://skild.ai/blogs/skil…REDACTED, https://www.globenewswire.com/news-release/2026/03/17/3256839/0/en/Skil…REDACTED.html (blogs.nvidia.com)

NVIDIA is strengthening the platform layer rather than betting on one robot maker. NVIDIA announced physical AI models/frameworks at CES 2026, including Isaac GR00T N1.6 and Cosmos-related tooling; NVIDIA also reported that Skild built S1 on NVIDIA infrastructure and that the broader collaboration targets synthetic data, simulation, model training, and real-world physical AI deployment. URLs: https://investor.nvidia.com/news/press-release-details/2026/NVID…REDACTED/, https://blogs.nvidia.com/blog/skil…REDACTED/ (investor.nvidia.com)

ABB Robotics / RobotStudio matters because it gives physical AI a path into existing industrial users. ABB’s robotics news via A3 reported that RobotStudio HyperReality integrates NVIDIA Omniverse libraries, aims to narrow virtual-to-real deployment, and is expected to serve ABB’s RobotStudio customer base, with select customers testing ahead of a second-half 2026 release. URL: https://www.automate.org/robotics/news/abb-…REDACTED (automate.org)

---

Dexterity: still the economic frontier

The strongest dexterity evidence this run is not walking; it is whole-room manipulation, home generalization, hand redesign, and nerve-connected prosthetic feedback. Figure says Figure 03 includes a redesigned sensory suite and hand system for Helix, and its Helix 02/2.5 posts focus on long-horizon loco-manipulation and zero-shot generalization in homes. BIOS/University of Utah’s LUKE Arm trial is a separate but important embodiment signal because real-time nerve interface plus touch feedback addresses the same control/feedback problem that limits both prosthetics and dexterous robotics. URLs: https://www.figure.ai/news/intr…REDACTED, https://www.figure.ai/news/helix-02, https://www.figure.ai/news/heli…REDACTED, https://www.biologicinputoutputsystems.com/media/neur…REDACTED (figure.ai)

Thesis state: walking is commoditizing; manipulation remains the bottleneck. Portfolio-relevant exposure should prefer companies that can prove manipulation in messy workflows, not just locomotion demos.

---

Defense robotics / autonomous systems

Defense is converting from experiment to procurement faster than humanoid consumer robotics. AeroVironment’s LOCUST was selected for a $464.8M U.S. Army Enduring-High Energy Laser counter-UAS contract; Shield AI was selected to compete for up to $800M in Navy ISR services with V-BAT; Rheinmetall delivered Mission Master under a $7.28M U.S. Marine Corps support contract; Sea Machines secured a five-year IDIQ with U.S. Special Operations Forces for autonomy kits; Hermeus picked Anduril autonomy for Quarterhorse under a funded partnership. URLs: https://www.avinc.com/2026/09/02/avs-…REDACTED/, https://shield.ai/shie…REDACTED/, https://www.rheinmetall.com/en/media/news-watch/news/2026/09/2026…REDACTED, https://sea-machines.com/sea-…REDACTED/, https://breakingdefense.com/2026/09/herm…REDACTED/ (avinc.com)

Aitan’s $41M emergence from stealth with “robotic sovereignty” and “edge AI weapon systems” is a smaller but strategically loud signal: military customers are buying autonomy stacks shaped by Ukraine/Middle East battlefield lessons, not just drone hardware. URL: https://www.axios.com/2026/09/02/aita…REDACTED (axios.com)

The U.S. Army xTechHumanoid competition explicitly lists baseline humanoid systems and advanced autonomy AI for autonomous mission planning/execution, showing that humanoid form factors are entering defense experimentation channels even if near-term defense conversion remains stronger in drones, ground vehicles, maritime autonomy, and counter-UAS. URL: https://xtech.army.mil/competition/xtechhumanoid/ (xtech.army.mil)

---

BCI / bionics / human-machine boundary

Precision Neuroscience’s September 24, 2026 $250M Series D is the largest human-machine-boundary financing event in this scan. The company says the round brings total capital raised to $430M and follows more than 100 patient procedures across 18 major medical institutions. URL: https://www.globenewswire.com/news-release/2026/09/24/3368254/0/en/prec…REDACTED.html (globenewswire.com)

DARPA’s Firefox program is a strategic government signal: it seeks high-speed, scalable, non-invasive neural interfaces using optical sensing and analog computing to reduce the precision/invasiveness trade-off in BCI. URL: https://www.darpa.mil/work-with-us/opportunities/darpa-ps-26-112 (darpa.mil)

Synchron’s public trial page still describes its BCI as investigational and not commercially approved, which keeps the category in clinical/regulatory watch mode rather than near-term revenue certainty. URL: https://synchron.com/join (synchron.com)

BIOS/University of Utah’s LUKE Arm at-home trial is important because it combines prosthetic control and touch feedback in everyday use. This is a direct embodiment signal: human nervous-system feedback loops are converging with robotic limb design. URL: https://www.biologicinputoutputsystems.com/media/neur…REDACTED (biologicinputoutputsystems.com)

---

AI materials / autonomous labs / Model → Matter

The AI-to-matter bridge is now receiving national-scale funding. NSF announced a $400M investment into a national network of AI-programmable cloud laboratories aligned with the U.S. Genesis Mission and autonomous laboratory infrastructure. URL: https://www.nsf.gov/tip/updates/nsf-…REDACTED (nsf.gov)

Specific nodes show the stack: Rice received $19.9M for AI-powered materials labs focused on electronic and quantum materials; Texas A&M received $24.9M for an Autonomous Robotic Metallurgist Materials Innovation Platform; UChicago/Argonne received $20M for PoLARIS robotic cloud labs for polymers; Johns Hopkins announced a $20M NSF-backed AI-powered lab network for extreme-condition materials. URLs: https://news.rice.edu/news/2026/acce…REDACTED, https://stories.tamu.edu/news/2026/08/05/texa…REDACTED/, https://news.uchicago.edu/story/ai-p…REDACTED, https://engineering.jhu.edu/news/buil…REDACTED/ (news.rice.edu)

Commercial early-stage signal: Azulene Labs raised $3.4M pre-seed to apply quantum-mechanical data, physics-based AI, and high-accuracy molecular simulation to drug/materials development. Small round, but thesis-relevant because it sits exactly at the model-to-matter interface. URL: https://www.azulenelabs.com/news/azul…REDACTED (azulenelabs.com)

---

CAPITAL_FLOW_MAP

Large / strategic capital flow

Capital-flow interpretation

Capital is clustering around five layers:

1. Compute/platform: NVIDIA, Nscale/Figure, Isaac/Cosmos/Omniverse.
2. Robot brains: Skild, Figure Helix, Physical Intelligence-like category, ABB/NVIDIA industrial simulation.
3. Deployment-first robotics: warehouse AMRs, industrial humanoids, logistics.
4. Defense autonomy: counter-UAS, ISR, maritime autonomy, autonomous strike/mission autonomy.
5. Human-machine boundary: BCI, peripheral nerve interfaces, prosthetics, wearable physiological interfaces.

CONVERSION_LAYER

Strongest conversion signals

1. Defense procurement > demo. AeroVironment, Shield AI, Rheinmetall, Sea Machines, Hermeus/Anduril, and Aitan all show money/contracts around autonomous defense systems. This is the highest near-term conversion layer in the scan. (avinc.com)

2. Warehouse/logistics > home humanoids. Figure at BMW/Catalyst, Agility/Digit, Vecna funding, and Skild industrial collaborations all point to constrained commercial environments as the real proving ground. Home humanoids remain strategically huge but operationally less validated. (figure.ai)

3. Robot-brain partnerships > single robot form factor. Skild with ABB/UR/MiR/NVIDIA and NVIDIA’s physical AI stack imply that portable intelligence may monetize across multiple embodiments before one humanoid dominates. (globenewswire.com)

4. Autonomous labs are converting through government infrastructure first. NSF’s $400M cloud-lab network and multiple $20M-ish nodes are not immediate public-equity revenue signals, but they are durable demand creation for lab automation, robotics, sensors, AI infrastructure, materials informatics, and cloud-controlled experimentation. (nsf.gov)

---

TIME_MAP

0–6 months

• Watch Tesla Optimus production evidence: units produced, factory hours, internal task hours, failure rates, BOM, hand redesign, and whether “production soon” becomes a measurable line item. Current verified state: Tesla is investing/preparing for large-scale Optimus production; commercial robot revenue remains unverified. (sec.gov)
• Watch Figure’s compute/data/deployment claims for third-party validation: BMW/Catalyst operational hours, customer economics, safety record, and whether BotQ one-per-hour production translates into active deployed fleet. (figure.ai)
• Watch Precision Neuroscience for regulatory progress after the $250M Series D; the financing is strong, but BCI commercialization remains regulatory and clinical-outcome bound. (globenewswire.com)

6–18 months

• Expect more physical AI platform announcements from NVIDIA/ABB/robot OEMs as simulation-to-real tools move into industrial customer workflows. ABB’s second-half 2026 RobotStudio HyperReality plan is an immediate watch item. (automate.org)
• Defense autonomy should continue converting through IDIQs, ISR services, counter-UAS, maritime autonomy, and edge AI weapon systems. (avinc.com)
• Autonomous lab nodes should begin procurement/buildout, benefiting vendors in lab robotics, characterization, sensors, cloud orchestration, materials informatics, and AI infrastructure. (nsf.gov)

18–36 months

• The key humanoid proof window: whether top players produce reliable deployed hours at customer sites, not staged videos.
• BCI/bionics proof window: patient count, at-home durability, reimbursement, implant/explant safety, and regulatory approvals.
• Materials/autonomous lab proof window: whether robotic labs produce commercially relevant discoveries faster/cheaper than conventional R&D.

---

BOTTLENECK_INDEX

---

THESIS_STATE

Confirmed / strengthened

• Physical AI is becoming a stack market: compute + world models + simulation + data + robot hardware + deployment operations.
• NVIDIA remains the broadest liquid platform exposure across humanoids, industrial robotics, simulation, and physical AI infrastructure. (investor.nvidia.com)
• Defense autonomy has the clearest near-term conversion path, with contract-backed demand and urgent operational need. (avinc.com)
• Dexterity/manipulation remains the central bottleneck, not locomotion. (figure.ai)
• Model → Matter is expanding beyond robots into autonomous labs/materials discovery. (nsf.gov)

Modified

• Humanoid exposure should be split into: 1. deployment-proven industrial/logistics humanoids, and 2. general-purpose home humanoid moonshots.

The first deserves more near-term portfolio weight; the second remains high-upside but less verified.

Not yet confirmed

• Tesla Optimus commercial readiness.
• Figure’s long-term customer economics and independent deployment-hour validation.
• Skild’s ARR quality, gross margins, and retention.
• BCI reimbursement and approval timelines.
• Which vendors capture NSF autonomous lab procurement.

---

STRATEGIC INTERPRETATION

The most important shift is that physical AI is no longer one market. It is splitting into investable layers:

1. AI infrastructure layer: NVIDIA, cloud compute, simulation, world models, synthetic data.
2. Embodied intelligence layer: Skild, Figure Helix, VLA policies, robot brains.
3. Robot hardware/manufacturing layer: Tesla Optimus, Figure 03, Apptronik Apollo, Agility Digit, Unitree, Boston Dynamics.
4. Deployment operations layer: warehouse/logistics, industrial plants, defense ISR/counter-UAS, autonomous maritime systems.
5. Human-machine boundary: BCI, prosthetics, neural interfaces.
6. Model → Matter science layer: AI-driven autonomous labs, materials, chemistry, robotics for experimentation.

SOURCES

1. Figure — “Figure and Nscale Sign Strategic Partnership For Up to 100,000 GPUs on the NVIDIA Vera Rubin Platform” — https://www.figure.ai/news/figu…REDACTED (figure.ai)
2. Figure — “Helix 2.5: Zero-Shot 30-Home Generalization” — https://www.figure.ai/news/heli…REDACTED (figure.ai)
3. Figure — “Figure Signs Agreement with Catalyst Brands to Scale Humanoid Operations” — https://www.figure.ai/news/figu…REDACTED (figure.ai)
4. Figure — “F.03 Arrives at BMW” — https://www.figure.ai/news/f-03-at-bmw (figure.ai)
5. Figure — “Ramping Figure 03 Production” — https://www.figure.ai/news/ramp…REDACTED (figure.ai)
6. NVIDIA — “NVIDIA Releases New Physical AI Models as Global Partners Unveil Next-Generation Robots” — https://investor.nvidia.com/news/press-release-details/2026/NVID…REDACTED/ (investor.nvidia.com)
7. NVIDIA Blog — “Skild AI Taps NVIDIA Physical AI to Teach Robots New Tasks From a Single Video” — https://blogs.nvidia.com/blog/skil…REDACTED/ (blogs.nvidia.com)
8. Skild AI — “The Hidden Pillar of Robotics” — https://skild.ai/blogs/skil…REDACTED (skild.ai)
9. GlobeNewswire — “Skild AI Expands Generalized Robot Intelligence Across Industries With ABB Robotics, Universal Robots and NVIDIA” — https://www.globenewswire.com/news-release/2026/03/17/3256839/0/en/Skil…REDACTED.html (globenewswire.com)
10. Automate/A3 — “ABB Robotics partners with Nvidia to deliver industrial-grade Physical AI at scale” — https://www.automate.org/robotics/news/abb-…REDACTED (automate.org)
11. Tesla SEC 10-Q — https://www.sec.gov/Archives/edgar/data/1318605/000162828026049270/tsla-20260630.htm (sec.gov)
12. AP — “Earnings at Musk's car company fall as research spending cuts into profit from selling cars” — https://apnews.com/article/976c…REDACTED (apnews.com)
13. Apptronik — “Apptronik Closes Over $935 Million Series A” — https://apptronik.com/news-collection/appt…REDACTED (apptronik.com)
14. AP — “Agility Robotics heads to Wall Street…” — https://apnews.com/article/39f2…REDACTED (apnews.com)
15. AP — “Shares in Chinese humanoid robot maker Unitree soar…” — https://apnews.com/article/f33f…REDACTED (apnews.com)
16. Axios — “Boston Dynamics opens testing ground for humanoid robots” — https://www.axios.com/local/boston/2026/09/21/bost…REDACTED (axios.com)
17. Vecna Robotics — “Vecna Robotics Raises $31 Million…” — https://www.vecnarobotics.com/news/vecn…REDACTED/ (vecnarobotics.com)
18. AeroVironment — “LOCUST Selected for Nearly $500 million Army Counter-UAS Contract…” — https://www.avinc.com/2026/09/02/avs-…REDACTED/ (avinc.com)
19. Shield AI — “Shield AI selected by U.S. Navy to compete for $800M…” — https://shield.ai/shie…REDACTED/ (shield.ai)
20. Rheinmetall — “American Rheinmetall delivers Mission Master to the U.S. Marine Corps” — https://www.rheinmetall.com/en/media/news-watch/news/2026/09/2026…REDACTED (rheinmetall.com)
21. Sea Machines — “Sea Machines secures IDIQ Contract with US Special Operations Forces” — https://sea-machines.com/sea-…REDACTED/ (sea-machines.com)
22. Breaking Defense — “Hermeus picks Anduril autonomy for Quarterhorse drone” — https://breakingdefense.com/2026/09/herm…REDACTED/ (breakingdefense.com)
23. Axios — “Aitan, battle-tested by IDF, raises $41 million” — https://www.axios.com/2026/09/02/aita…REDACTED (axios.com)
24. U.S. Army xTechHumanoid — https://xtech.army.mil/competition/xtechhumanoid/ (xtech.army.mil)
25. Precision Neuroscience — $250M Series D — https://www.globenewswire.com/news-release/2026/09/24/3368254/0/en/prec…REDACTED.html (globenewswire.com)
26. DARPA Firefox — https://www.darpa.mil/work-with-us/opportunities/darpa-ps-26-112 (darpa.mil)
27. Synchron trial page — https://synchron.com/join (synchron.com)
28. BIOS / LUKE Arm trial — https://www.biologicinputoutputsystems.com/media/neur…REDACTED (biologicinputoutputsystems.com)
29. Beta Bionics — proposed $125M public offering — https://investors.betabionics.com/news-releases/news-release-details/beta…REDACTED (investors.betabionics.com)
30. NSF — $400M AI-programmable cloud labs — https://www.nsf.gov/tip/updates/nsf-…REDACTED (nsf.gov)
31. Rice — $19.9M AI-powered materials lab — https://news.rice.edu/news/2026/acce…REDACTED (news.rice.edu)
32. Texas A&M — $24.9M autonomous robotic metallurgist lab — https://stories.tamu.edu/news/2026/08/05/texa…REDACTED/ (stories.tamu.edu)
33. UChicago/Argonne — $20M PoLARIS robotic labs — https://news.uchicago.edu/story/ai-p…REDACTED (news.uchicago.edu)
34. Johns Hopkins — $20M materials lab network — https://engineering.jhu.edu/news/buil…REDACTED/ (engineering.jhu.edu)
35. Azulene Labs — $3.4M pre-seed — https://www.azulenelabs.com/news/azul…REDACTED (azulenelabs.com)