Physical AI · 14 Aug 2026

2026-08-14 — Task 1 + Task 2: Physical AI / Model → Matter Scan

☉ SolPhysical AI14 Aug 2026EN16 min

This report exists in English only.

Run metadata:

  • ran_at: 2026-08-14
  • 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
  • evidence standard: web-current sources fetched in this run only
  • portfolio handoff: provisional; final action requires Eth IB current-state overlay

EXECUTIVE READ

Physical AI is moving from “humanoid demo cycle” into a three-layer deployment stack:

  1. Infrastructure layer — NVIDIA / Google DeepMind / robot-foundation-model companies are trying to own simulation, world models, embodied reasoning, and edge inference.
  2. Deployment layer — Figure, Agility, Boston Dynamics/Hyundai, Apptronik, 1X, Walden, Unitree and others are moving from videos into factories, logistics facilities, robot parks, pilot fleets, and pre-commercial manufacturing lines.
  3. Sovereign/defense layer — counter-UAS, attritable autonomous interceptors, autonomous surveillance, undersea autonomy, and defense AI software are already contracting faster than civilian humanoids.

The strongest portfolio-relevant signal is not “humanoids are ready everywhere.” It is: capital is consolidating around embodied AI platforms with either real deployment data, manufacturing paths, or defense procurement pull. Figure’s BMW data, Agility’s SPAC/order book, Unitree’s IPO path, NVIDIA’s physical-AI stack, and the U.S. counter-drone contract flow are the cleanest signals this cycle. (figure.ai)


SIGNAL_LAYER

Humanoid deployment signal: Figure moved from pilot narrative to measurable industrial runtime

Figure published BMW deployment metrics for Figure 02: 10-hour shifts Monday-Friday, 90,000+ parts loaded, 1,250+ runtime hours, and contribution to production of 30,000+ BMW X3 vehicles; the disclosed use case involved sheet-metal loading with cycle-time, placement-accuracy, and intervention KPIs. This is materially higher-quality evidence than a staged humanoid demo because it reports hours, parts, factory context, and failure learnings. URL: https://www.figure.ai/news/production-at-bmw (figure.ai)

Figure then announced Figure 03 at BMW Plant Spartanburg on June 30, 2026, moving into a harder “sequencing” logistics workflow requiring loco-manipulation, part handling, cart movement, and visual-motor correction through Helix 02. The signal: Figure is pushing from fixed pick/place toward dynamic material manipulation, which is closer to the economic frontier of humanoids. URL: https://www.figure.ai/news/f-03-at-bmw (figure.ai)

Figure also signed a commercial agreement with Catalyst Brands for deployment into its distribution and logistics network, starting at the Reno, Nevada Distribution Logistics Center; Catalyst operates brands including JCPenney, Aéropostale, and Brooks Brothers. This broadens Figure’s deployment surface from automotive manufacturing toward retail logistics. URL: https://www.figure.ai/news/figu…REDACTED (figure.ai)

Interpretation: Figure remains private, but it is becoming a leading benchmark for “deployment evidence.” If confirmed by customers rather than only company posts, this supports the thesis that factory/logistics humanoids will commercialize before home humanoids.


Agility Robotics: SPAC path + contracted orders create a public-market watchpoint

Agility Robotics and Churchill Capital Corp XI announced a business combination on June 24, 2026, with expected ticker AGLT, a $2.5B pre-money equity value, more than $620M expected gross proceeds, and roughly $200M PIPE financing. Agility disclosed more than $300M of multi-year contracted Digit v5 orders, subject to contractual milestones, and a pipeline of over 30 customers. URL: https://www.sec.gov/Archives/edgar/data/2074973/000121390026071290/ea029548401ex99-1.htm (sec.gov)

Digit is positioned for manufacturing, distribution, and logistics environments, not consumer use. This matters because the near-term humanoid ROI path is labor augmentation in semi-structured industrial spaces, not general home autonomy. (sec.gov)

Interpretation: If/when AGLT lists, it becomes one of the few liquid humanoid pure-play exposures. Watch order conversion, margin structure, deployment uptime, customer concentration, and whether “contracted orders” convert into recognized revenue.


Unitree: China humanoid supply-chain signal moving toward public listing

The Shanghai Stock Exchange announced a June 1, 2026 review meeting for Unitree Robotics’ IPO application, with a fundraising plan above RMB 4.2B, while noting that Chinese humanoid manufacturers are accelerating listing plans amid embodied-AI competition. URL: https://english.sse.com.cn/news/newsrelease/voice/c/c_20260526_10819716.shtml (english.sse.com.cn)

TechNode reported that Unitree cleared the listing committee review on June 2, 2026 and that the company said it shipped more than 5,500 humanoid robots during the referenced period. URL: https://technode.com/2026/06/02/unit…REDACTED/ (technode.com)

Interpretation: Unitree is the clearest China-side embodiment/manufacturing signal. For public-market strategy, it is less a direct U.S./EU buy signal and more a pressure signal on Western humanoid valuations, component BOMs, and time-to-market assumptions.


NVIDIA physical-AI stack: ecosystem control is becoming the more investable layer than individual robot OEMs

NVIDIA announced partnerships with robotics ecosystem players, including robot-brain developers and humanoid pioneers, around Cosmos world models, Isaac simulation, and production-scale physical AI. URL: https://investor.nvidia.com/news/press-release-details/2026/NVID…REDACTED/ (investor.nvidia.com)

NVIDIA also framed Cosmos 3, Isaac GR00T N1.7, and Alpamayo 1.5 as models/frameworks for physical AI reasoning and action across robots and autonomous vehicles. URL: https://investor.nvidia.com/news/press-release-details/2026/NVID…REDACTED/ (investor.nvidia.com)

NVIDIA’s GTC material says ABB Robotics, FANUC, KUKA, and Yaskawa — with a combined installed base above 2M robots — are using Omniverse libraries and Isaac simulation frameworks to validate robot applications and production lines through digital twins. URL: https://blogs.nvidia.com/blog/gtc-…REDACTED/ (blogs.nvidia.com)

Interpretation: The highest-quality public equity signal remains infrastructure: simulation, synthetic data, robot policy training, edge compute, and physical-AI developer tooling. This supports overweighting the picks-and-shovels layer versus trying to pick a single humanoid winner too early.


Google DeepMind: embodied reasoning is being productized through Gemini Robotics-ER

Google DeepMind introduced Gemini Robotics-ER 1.6 in 2026 as an embodied-reasoning model for robots, available via Gemini API and Google AI Studio; its model card describes the system as a VLM that enhances spatial and physical reasoning capabilities. URLs: https://deepmind.google/blog/gemi…REDACTED/ and https://deepmind.google/models/model-cards/gemi…REDACTED/ (deepmind.google)

Apptronik’s press page states that fleets of Apollo 2 robots collect real-world data across Robot Park locations and customer sites in partnership with Google DeepMind, and that Apptronik has a strategic partnership with Google DeepMind’s robotics team. URL: https://apptronik.com/company/press-releases (apptronik.com)

Interpretation: Google/Alphabet is a serious robot-brain participant, but portfolio exposure is diluted inside Alphabet. The important watchpoint is whether Gemini Robotics becomes a cross-embodiment control layer or mainly a strategic data partnership with select robotics companies.


Skild AI / Physical Intelligence / Apptronik / Walden: mega-rounds confirm capital crowding into robot brains and deployment systems

Skild AI announced a $1.4B Series C on January 14, 2026 and describes its system as a unified robotics foundation model generalizing across tasks and robot hardware. URL: https://www.skild.ai/blogs/series-c (skild.ai)

TechCrunch reported Apptronik raised its Series A total to $935M at a $5B+ valuation in February 2026. URL: https://techcrunch.com/2026/02/11/huma…REDACTED/ (techcrunch.com)

TechCrunch reported Physical Intelligence was in talks to raise about $1B at an $11B+ valuation in March 2026. This is not yet equivalent to a closed financing based on this source, so it remains a watchlist signal, not a confirmed capital event. URL: https://techcrunch.com/2026/03/27/phys…REDACTED/ (techcrunch.com)

Walden Robotics launched from stealth on July 15, 2026 with $300M in seed funding, a $1.1B valuation, and participation from Toyota, NVIDIA, Boeing, Samsung Ventures, Prologis Ventures, CoreWeave Ventures, and others. URL: https://www.waldenrobotics.com/news/wald…REDACTED (waldenrobotics.com)

Interpretation: Capital is flowing into two types of company: generalized robot intelligence and full-stack “robots doing real work now.” The private-market valuations are aggressive; public-market proxies should remain disciplined.


Tesla Optimus: still strategically important, but manufacturing-risk signal increased

Tesla’s Q1 2026 update said first-generation Optimus production lines were being installed in anticipation of volume production. URL: https://ir.tesla.com/_flysystem/s3/sec/000162828026026551/tsla-20260422-gen.pdf (ir.tesla.com)

TechCrunch reported that Tesla’s Q2 2026 shareholder materials removed earlier language about Optimus reaching “volume production,” and quoted Musk saying Optimus would be Tesla’s hardest product to scale because “everything on the robot is new.” URL: https://techcrunch.com/2026/07/22/tesl…REDACTED/ (techcrunch.com)

Interpretation: TSLA remains the most liquid public humanoid narrative, but this cycle’s evidence argues against treating Optimus as a near-term revenue anchor. It is a long-duration optionality component, not a confirmed 2026 conversion signal.


Boston Dynamics / Hyundai: strongest incumbent manufacturing pathway, but deployment timing is later

Boston Dynamics unveiled the product version of the electric Atlas at CES 2026 and disclosed Hyundai’s broader U.S. robotics-factory plan capable of producing 30,000 robots per year. URL: https://bostondynamics.com/blog/bost…REDACTED/ (bostondynamics.com)

Hyundai said Atlas robots will begin operations at HMGMA by 2028 and that the Robotics Metaplant Application Center opens in 2026 for training and validation. URL: https://www.hyundaimotorgroup.com/en/story/hyun…REDACTED (hyundaimotorgroup.com)

Boston Dynamics states that Hyundai is its first Atlas customer, with a completed deployment last year and a fleet scheduled to ship to RMAC in 2026. URL: https://bostondynamics.com/blog/ente…REDACTED/ (bostondynamics.com)

Interpretation: Hyundai/Boston Dynamics is less hype-prone than many humanoid names because it has manufacturing discipline and internal deployment sites. The drawback is indirect exposure through Hyundai and longer timelines.


Defense robotics / autonomous systems: procurement is already ahead of civilian humanoid revenue

Joint Interagency Task Force 401 awarded a three-year IDIQ contract with a $500M ceiling to Perennial Autonomy for AI-enabled counter-UAS systems, including Merops interceptors, Bumblebee quadcopters, and Hornet midrange strike drones; the systems integrate detection, tracking, engagement, computer vision, RF sensing, jam-resistant communications, and autonomous targeting while retaining warfighter decision authority over lethal force. URL: https://www.jbsa.mil/News/News/Article/4496078/join…REDACTED/ (jbsa.mil)

Kratos announced a roughly $156M single-award contract from the U.S. Department of Energy Office of Secure Transportation for mobile counter-UAS. URL: https://www.kratosdefense.com/newsroom/krat…REDACTED (kratosdefense.com)

Anduril announced a June 2026 expansion with U.S. Customs and Border Protection for 200 additional Extended Range Sentry Towers, and a May 2026 British Army down-select for Land Autonomous Collaborative Platforms. URLs: https://www.anduril.com/news/andu…REDACTED and https://www.anduril.com/news/brit…REDACTED (anduril.com)

Reuters reported through Investing.com that Anduril was in talks for a funding round at about a $100B valuation in July 2026; this is a reported discussion, not a closed round based on the source. URL: https://www.investing.com/news/economy-news/excl…REDACTED (investing.com)

Interpretation: Defense autonomy has clearer procurement pull than most civilian humanoid robotics. Public-market attention should track C-UAS, drones, autonomy software, sensing, edge compute, and defense primes with autonomy exposure.


AI materials / autonomous labs: “Model → Matter” is becoming funded infrastructure, not only research rhetoric

Nature Communications Materials published a July 2026 article on managing autonomous materials labs with multi-agent AI, arguing that as autonomous experimentation systems increase, AI-driven science can move toward labs that optimize resources locally and across lab networks. URL: https://www.nature.com/articles/s43246-026-01219-5 (nature.com)

The NSF awarded $20M to Johns Hopkins-led AIMD-Net for an AI-powered lab network for autonomous materials research. URL: https://engineering.jhu.edu/news/buil…REDACTED/ (engineering.jhu.edu)

CUNY received an $18.1M NSF award to build a cloud-programmable national lab using AI and robotics to speed discovery, design, and production of bio-inspired materials. URL: https://www.eurekalert.org/news-releases/1137370 (eurekalert.org)

UChicago/Argonne received a $20M NSF grant for PoLARIS, a remotely accessible, robot-operated polymer lab, with the team expecting end-to-end autonomous experiments for initial science drivers within 18 months. URL: https://www.computerscience.uchicago.edu/news/remo…REDACTED/ (computerscience.uchicago.edu)

Medra launched its AI Experimentalist and announced a DARPA-funded project; Medra says its Physical AI Scientist platform couples AI reasoning with wet-lab execution, and that its ML001 autonomous lab is running closed-beta projects across academia, biopharma, and government. URL: https://www.medra.ai/darpa-collaboration (medra.ai)

Interpretation: AI materials and autonomous labs are slower to monetize directly in public equities, but they strengthen the long-term thesis that AI is moving from prediction into experiment execution, manufacturing, and materials validation.


BCI / neural interface: clinical milestones continue, but commercialization remains regulatory-medical, not consumer

Neuralink’s updates page says “Two Years of Telepathy” announced 21 Neuralink participants in January 2026; Neuralink describes its BCI as fully implanted, wireless, and high-channel-count for people with paralysis. URL: https://neuralink.com/updates/ (neuralink.com)

Neuralink’s device-control trial page lists PRIME for computer control and CONVOY for control of a computer and robotic arm through thought; it states that multiple ALS and spinal cord injury participants have received the N1 Implant after the first participant in January 2024. URL: https://neuralink.com/trials/device-control/ (neuralink.com)

Synchron raised a $200M Series D in November 2025 to advance non-surgical BCI technology; while the round is not 2026, it remains relevant to the current BCI capital stack because it is within the current clinical cycle. URL: https://www.businesswire.com/news/home/20251106150841/en/Synchron-Raises-%2420…REDACTED (businesswire.com)

Paradromics reported on July 7, 2026 that the University of Michigan completed the first Connexus BCI implantation for its FDA-approved Connect-One clinical study, evaluating long-term use to restore speech and enable computer control for people with severe motor impairment. URL: https://paradromics.com/news/ (paradromics.com)

Interpretation: BCI is progressing into more serious clinical evidence. Portfolio stance: private watchlist / public medtech-adjacent only. Avoid treating BCI as near-term consumer “mind control” optionality.


Bionics / prosthetics / exoskeletons: useful but still reimbursement- and adoption-constrained

Lifeward announced in February 2026 a definitive agreement to acquire Skelable upper-body exoskeleton technology with AI capabilities for stroke survivors, for $480,000 in Lifeward equity plus nominal cash and milestone terms. URL: https://ir.rewalk.com/news-releases/news-release-details/life…REDACTED (ir.rewalk.com)

Open Bionics’ MedTech Index profile says the Hero Arm is adopted by NHS and equivalent European health systems, and that in March 2026 the company took a £600,000 NatWest loan secured against IP to fund six additional U.S. clinics. URL: https://www.bridgeheadcommunications.com/medtech-index/open-bionics (bridgeheadcommunications.com)

University of Melbourne announced Australia’s first fully integrated bionic arm project in July 2026, combining a lightweight 3D-printed prosthetic hand with a sensorized socket that detects subtle muscle movement. URL: https://www.unimelb.edu.au/newsroom/news/2026/july/aust…REDACTED (unimelb.edu.au)

Interpretation: Bionics/prosthetics remain strategically relevant to dexterity, sensing, soft robotics, and human-machine interface, but direct public-equity opportunity is weaker than robotics infrastructure or defense autonomy.


CAPITAL_FLOW_MAP

Confirmed / high-confidence capital and transaction signals

Signal Capital / transaction Status Portfolio relevance
Agility Robotics / Churchill Capital XI $2.5B pre-money equity value; >$620M expected gross proceeds; ~$200M PIPE; >$300M multi-year Digit v5 contracted orders Definitive business combination announced Potential public humanoid pure-play under expected ticker AGLT; watch de-SPAC quality and revenue conversion. (sec.gov)
Skild AI $1.4B Series C Company-announced Confirms large capital flow into robot-brain layer; private, but supports NVDA/SoftBank/robotics ecosystem thesis. (skild.ai)
Apptronik $935M Series A total at $5B+ valuation Reported by TechCrunch Private humanoid deployment/data flywheel; Google DeepMind partnership relevance. (techcrunch.com)
Walden Robotics $300M seed; $1.1B valuation Company-announced Toyota/NVIDIA/Boeing/Samsung/Prologis participation signals industrial capital appetite. (waldenrobotics.com)
Unitree Robotics >RMB 4.2B IPO fundraising plan; SSE review Exchange / media sources China embodiment scale signal; indirect pressure on Western valuations and component costs. (english.sse.com.cn)
JIATF-401 / Perennial Autonomy $500M IDIQ ceiling U.S. military source Defense autonomy procurement pull; C-UAS as immediate revenue market. (jbsa.mil)
Kratos ~$156M mobile C-UAS contract Company-announced Public defense autonomy exposure via KTOS. (kratosdefense.com)
NSF autonomous labs $20M JHU AIMD-Net; $18.1M CUNY AI-BIOME; $20M UChicago PoLARIS Academic/NSF-related announcements Model→Matter infrastructure funding; long-term science automation thesis. (engineering.jhu.edu)
Synchron $200M Series D BusinessWire/company BCI clinical capital signal; private. (businesswire.com)
Lifeward / Skelable $480K equity-based asset acquisition + milestones Company-announced Small-cap bionics/exoskeleton optionality; high risk. (ir.rewalk.com)

Reported but not confirmed as closed

Signal Reported value Treatment
Anduril Talks at ~$100B valuation Watchlist only until closed; still important as private defense-autonomy valuation marker. (investing.com)
Physical Intelligence Talks to raise ~$1B at $11B+ valuation Watchlist only until closed; confirms investor appetite for robot-brain layer if finalized. (techcrunch.com)

CONVERSION_LAYER

Deployment → revenue conversion quality

Highest current conversion quality

  • Agility Robotics: disclosed contracted orders and public listing path; still must prove production, deployment, service margin, safety certification, and customer retention. (sec.gov)
  • Figure: reported real factory metrics and new commercial deployments, but remains private and company-sourced; customer-side independent confirmation would improve confidence. (figure.ai)
  • Defense C-UAS: procurement contracts are clearer and nearer-term than humanoid revenue; JIATF-401 and Kratos signals are stronger conversion evidence than most civilian demos. (jbsa.mil)

Medium conversion quality

  • Boston Dynamics/Hyundai: excellent manufacturing pathway and customer alignment, but broader Atlas operations are guided toward 2028 at HMGMA, not immediate mass revenue. (hyundaimotorgroup.com)
  • Google/NVIDIA infrastructure: high strategic importance, but physical-AI revenue is embedded inside broader AI/cloud/GPU businesses and hard to isolate. (investor.nvidia.com)

Low / speculative conversion quality

  • Home humanoids: 1X says U.S. deliveries start in 2026 and broader markets in 2027, but home autonomy remains technically and operationally harder than structured industrial tasks. URL: https://www.1x.tech/discover/neo-home-robot (1x.tech)
  • BCI consumer expansion: Neuralink/Paradromics/Synchron are medical-first; consumer upside is not investable as a near-term public thesis. (neuralink.com)

TIME_MAP

0–6 months: monitor for near-term catalyst

  • Agility de-SPAC progress, ticker transition to expected AGLT, order conversion, and first public-company financial disclosures. (sec.gov)
  • Tesla Q3/Q4 Optimus language: whether “volume production” returns or remains softened. (techcrunch.com)
  • Additional Figure customer-side confirmation from BMW / Catalyst or third-party facility reports. (figure.ai)
  • Defense C-UAS follow-on awards after the $500M JIATF-401 IDIQ and Kratos $156M contract. (jbsa.mil)
  • Unitree IPO final listing details, pricing, float, and any foreign-access pathway. (english.sse.com.cn)

6–18 months: thesis confirmation window

  • Whether humanoid companies report uptime, interventions, cycle-time economics, and gross margins rather than only deployment announcements.
  • Whether NVIDIA/Google robotics models become default middleware across multiple embodiments, not just partner showcases. (investor.nvidia.com)
  • Whether autonomous lab grants produce repeatable closed-loop experiments and industrial customers, especially in materials/polymers/biotech. (engineering.jhu.edu)
  • Whether Paradromics / Neuralink / Synchron publish durable clinical performance metrics and regulatory progress. (neuralink.com)

18–36 months: scale-or-reset window

  • Hyundai/Boston Dynamics Atlas operations at HMGMA by 2028. (hyundaimotorgroup.com)
  • Humanoid cost curves: BOM, service costs, battery replacement, actuators, hand failure rates, field maintenance.
  • Defense autonomy normalization: autonomous interceptors, loitering systems, ISR autonomy, undersea autonomy, and battle-management AI becoming procurement staples.

BOTTLENECK_INDEX

Bottleneck Severity Evidence / rationale Portfolio implication
Dexterous manipulation reliability Very high Figure identified forearm/wrist electronics as a failure point in BMW deployment and redesigned Figure 03 wrist electronics. (figure.ai) Favor companies with field data and component redesign loops; avoid pure demo names.
Cooperative safety Very high Agility explicitly frames Digit v5 around cooperative safety as the unlock for robots working alongside people. (sec.gov) Safety-certified humanoids may command premium enterprise adoption.
Manufacturing scale Very high Tesla’s Optimus scaling language weakened from Q1 to Q2, and Musk described Optimus as Tesla’s hardest product to scale. (ir.tesla.com) Do not price TSLA as if Optimus near-term scale is already proven.
Data flywheel High Apptronik says Apollo fleets collect real-world data across Robot Park and customer sites with Google DeepMind. (apptronik.com) Watch companies with real operation data, not only synthetic simulation.
Sim-to-real gap High NVIDIA is pushing Cosmos/Isaac/Omniverse for physical-AI data generation, simulation, and validation; this indicates the gap remains central. (investor.nvidia.com) Supports infrastructure exposure: NVDA, cloud, sensors, simulation, edge compute.
Regulatory / clinical evidence for BCI High Neuralink and Paradromics are in clinical-study mode; Paradromics’ Connexus is in early feasibility evaluation. (neuralink.com) Treat BCI as medical-device timeline, not consumer-tech timeline.
Reimbursement / adoption for bionics High Open Bionics and Lifeward signals show clinic expansion and technology acquisition, but the capital scale is small and adoption constraints remain. (bridgeheadcommunications.com) Small-cap bionics should be capped unless reimbursement traction is clear.
Defense autonomy ethics / rules of engagement Medium-high JIATF-401 source states autonomous targeting while retaining warfighter decision authority over lethal force. (jbsa.mil) Defense autonomy is investable but headline/regulatory risk is non-trivial.

THESIS_STATE

Current thesis grade: Strengthened, with sharper separation between layers

Strengthened

  • Physical AI infrastructure: stronger due to NVIDIA/Google ecosystem moves and robotics foundation model deployment.
  • Industrial humanoids: stronger because Figure and Agility show more concrete commercial signals.
  • Defense autonomy: stronger because procurement is already material.
  • Model → Matter: stronger because autonomous labs are receiving NSF/DARPA-linked funding and moving into lab infrastructure.

Weakened / delayed

  • Home humanoids: still too early; 1X delivery claims are important but not enough to underwrite mass adoption.
  • Tesla Optimus near-term revenue: reduced confidence for 2026 volume assumptions.
  • Bionics public-market opportunity: medically meaningful, but still small, reimbursement-limited, and fragmented.

Core thesis

The investable opportunity is not “one humanoid company wins.” The better thesis is: AI is gaining physical execution pathways through simulation, robot brains, industrial deployment, autonomous defense systems, autonomous labs, and neural/assistive interfaces. Public portfolios should prioritize infrastructure and procurement-backed autonomy while keeping humanoid pure-play exposure controlled until revenue quality appears.


STRATEGIC INTERPRETATION

What changed this cycle

  1. Deployment evidence improved. Figure’s BMW metrics and Agility’s disclosed contracted orders are stronger than ordinary robotics PR because they connect robots to factory/logistics workflows and operational KPIs. (figure.ai)

  2. Humanoid public-market access may expand. Agility’s expected AGLT listing would create a liquid humanoid pure-play; this may pull capital into the sector but also expose weak economics if deployments are costly. (sec.gov)

  3. Defense autonomy is already in procurement mode. C-UAS and autonomous interceptors are moving through sizable contracts now, while humanoids are still proving reliability and ROI. (jbsa.mil)

  4. Robot-brain capital is crowding. Skild, Apptronik, Walden, and reported Physical Intelligence activity show investors are underwriting generalizable embodied intelligence, not just hardware shells. (skild.ai)

  5. Tesla Optimus remains optionality, not confirmation. The Q2 shift in volume-production language raises execution-risk flags for anyone using Optimus as near-term TSLA valuation support. (techcrunch.com)


SOURCES

  1. Figure — F.02 Contributed to the Production of 30,000 Cars at BMW
    URL: https://www.figure.ai/news/production-at-bmw

  2. Figure — F.03 Arrives at BMW
    URL: https://www.figure.ai/news/f-03-at-bmw

  3. Figure — Agreement with Catalyst Brands
    URL: https://www.figure.ai/news/figu…REDACTED

  4. SEC / Agility Robotics + Churchill Capital Corp XI business combination
    URL: https://www.sec.gov/Archives/edgar/data/2074973/000121390026071290/ea029548401ex99-1.htm

  5. Shanghai Stock Exchange — Unitree IPO review
    URL: https://english.sse.com.cn/news/newsrelease/voice/c/c_20260526_10819716.shtml

  6. TechNode — Unitree IPO approved
    URL: https://technode.com/2026/06/02/unit…REDACTED/

  7. NVIDIA — Global robotics leaders take Physical AI to the real world
    URL: https://investor.nvidia.com/news/press-release-details/2026/NVID…REDACTED/

  8. NVIDIA — Open model families for agentic, physical, healthcare AI
    URL: https://investor.nvidia.com/news/press-release-details/2026/NVID…REDACTED/

  9. NVIDIA Blog — GTC 2026 virtual worlds / physical AI
    URL: https://blogs.nvidia.com/blog/gtc-…REDACTED/

  10. Google DeepMind — Gemini Robotics-ER 1.6
    URL: https://deepmind.google/blog/gemi…REDACTED/

  11. Google DeepMind — Gemini Robotics-ER 1.6 model card
    URL: https://deepmind.google/models/model-cards/gemi…REDACTED/

  12. Apptronik press releases
    URL: https://apptronik.com/company/press-releases

  13. Skild AI — Series C announcement
    URL: https://www.skild.ai/blogs/series-c

  14. TechCrunch — Apptronik raised $935M at $5B+ valuation
    URL: https://techcrunch.com/2026/02/11/huma…REDACTED/

  15. TechCrunch — Physical Intelligence reported funding talks
    URL: https://techcrunch.com/2026/03/27/phys…REDACTED/

  16. Walden Robotics — $300M launch from stealth
    URL: https://www.waldenrobotics.com/news/wald…REDACTED

  17. Tesla Q1 2026 update PDF
    URL: https://ir.tesla.com/_flysystem/s3/sec/000162828026026551/tsla-20260422-gen.pdf

  18. TechCrunch — Tesla Q2 2026 / Optimus volume-production language
    URL: https://techcrunch.com/2026/07/22/tesl…REDACTED/

  19. Boston Dynamics — Atlas product version
    URL: https://bostondynamics.com/blog/bost…REDACTED/

  20. Hyundai Motor Group — CES 2026 AI robotics strategy
    URL: https://www.hyundaimotorgroup.com/en/story/hyun…REDACTED

  21. Boston Dynamics — Enterprise robotics / Atlas deployment
    URL: https://bostondynamics.com/blog/ente…REDACTED/

  22. JIATF-401 — $500M counter-UAS contract
    URL: https://www.jbsa.mil/News/News/Article/4496078/join…REDACTED/

  23. Kratos — ~$156M mobile C-UAS contract
    URL: https://www.kratosdefense.com/newsroom/krat…REDACTED

  24. Anduril — CBP 200 additional Extended Range Sentry Towers
    URL: https://www.anduril.com/news/andu…REDACTED

  25. Anduril — British Army Land Autonomous Collaborative Platforms
    URL: https://www.anduril.com/news/brit…REDACTED

  26. Investing.com / Reuters — Anduril reported $100B valuation talks
    URL: https://www.investing.com/news/economy-news/excl…REDACTED

  27. Nature Communications Materials — autonomous materials labs with multi-agent AI
    URL: https://www.nature.com/articles/s43246-026-01219-5

  28. Johns Hopkins — $20M NSF AI-powered materials lab network
    URL: https://engineering.jhu.edu/news/buil…REDACTED/

  29. EurekAlert / CUNY — $18.1M NSF AI-powered materials lab
    URL: https://www.eurekalert.org/news-releases/1137370

  30. UChicago / Argonne — $20M NSF robot-operated polymer lab
    URL: https://www.computerscience.uchicago.edu/news/remo…REDACTED/

  31. Medra — AI Experimentalist and DARPA collaboration
    URL: https://www.medra.ai/darpa-collaboration

  32. Neuralink updates
    URL: https://neuralink.com/updates/

  33. Neuralink device-control trials
    URL: https://neuralink.com/trials/device-control/

  34. BusinessWire — Synchron $200M Series D
    URL: https://www.businesswire.com/news/home/20251106150841/en/Synchron-Raises-%2420…REDACTED

  35. Paradromics news
    URL: https://paradromics.com/news/

  36. Lifeward — Skelable AI upper-body exoskeleton acquisition
    URL: https://ir.rewalk.com/news-releases/news-release-details/life…REDACTED

  37. Open Bionics MedTech Index
    URL: https://www.bridgeheadcommunications.com/medtech-index/open-bionics

  38. University of Melbourne — integrated bionic arm
    URL: https://www.unimelb.edu.au/newsroom/news/2026/july/aust…REDACTED

  39. 1X — NEO Home Robot
    URL: https://www.1x.tech/discover/neo-home-robot

Source in the house: eth-memory/sol/reports/2026-08-14-physical-ai-model-to-matter.md☉ Sol