Run metadata:

• run_date: 2026-07-25
• job_id: physical_ai_model_to_matter
• task: Task 1 — Physical AI / Embodiment Scan + Task 2 — Ethan Portfolio Engine
• requires_web_current: TRUE
• scope: physical AI / embodiment / robotics / bionics / AI materials / BCI / defense robotics / capital-flow signals
• evidence_rule: current claims below are based on sources fetched in this run.

Executive read

Physical AI is moving from “demo narrative” into three investable infrastructure lanes:

1. compute + simulation + robot foundation models, led by NVIDIA Cosmos / Isaac / GR00T and Google Gemini Robotics;
2. deployment/data flywheels, where Figure, Apptronik, Tesla, Unitree, 1X, Skild, and Physical Intelligence are competing for real-world robot data;
3. defense autonomy, where Anduril and Shield AI are converting autonomy from lab software into production-adjacent military programs. NVIDIA’s July 2026 Cosmos 3 Edge launch is the strongest near-term “model → robot action” infrastructure signal because it explicitly targets local robot control policies on Jetson Thor after post-training on proprietary robot/sensor data. Source: NVIDIA Blog, 2026-07-20 — https://blogs.nvidia.com/blog/siggraph-news-2026/ (blogs.nvidia.com)

The biggest thesis change since earlier Physical AI cycles: capital is no longer only chasing humanoid bodies; it is flowing into the data/simulation/control layer that makes bodies useful. Figure is expanding deployment through BMW and Catalyst Brands; Apptronik has raised nearly $1B Series A capital and opened Robot Park data/training facilities; Skild raised a reported $1.4B Series C at a $14B+ valuation; Physical Intelligence has been reported in talks to raise around $1B at an $11B+ valuation. Sources: Figure news — https://www.figure.ai/news; Apptronik Robot Park — https://apptronik.com/news-collection/welc…REDACTED; TechCrunch / Skild — https://techcrunch.com/2026/01/14/robo…REDACTED/; TechCrunch / Physical Intelligence — https://techcrunch.com/2026/03/27/phys…REDACTED/ (figure.ai)

SIGNAL_LAYER

Physical AI infrastructure: NVIDIA pushes edge world models into robot control

Signal: NVIDIA announced Cosmos 3 Edge as an open world model for local physical AI, with a workflow where developers post-train on proprietary robot/sensor data using DGX Station and deploy on Jetson Thor for real-time manipulation or locomotion policies. This is a direct “model → matter” stack signal: training, simulation, world/action modeling, and edge execution are being bundled. Source: NVIDIA Blog, 2026-07-20 — https://blogs.nvidia.com/blog/siggraph-news-2026/ (blogs.nvidia.com)

Why it matters: This strengthens the thesis that NVIDIA is not only selling GPUs into AI training; it is building the operating substrate for robotics: Train / Simulate / Run. NVIDIA’s investor material explicitly frames physical AI around DGX, Omniverse + Cosmos, and Jetson Thor. Source: NVIDIA July 2026 investor deck — https://s201.q4cdn.com/141608511/files/doc_presentations/2026/07/NDR_July2026_.pdf (s201.q4cdn.com)

Google DeepMind: Gemini Robotics becomes a platform layer, not just research

Signal: Google DeepMind’s Gemini Robotics line now includes Gemini Robotics-ER 1.6, available via Gemini API and Google AI Studio, with improved physical/spatial reasoning and safety-constraint behavior; Google also maintains Gemini Robotics On-Device for local robot execution. Sources: Google blog — https://blog.google/innovation-and-ai/models-and-research/google-deepmind/gemi…REDACTED/; DeepMind model card — https://deepmind.google/models/model-cards/gemi…REDACTED/; Gemini Robotics On-Device — https://deepmind.google/models/gemini-robotics/gemini-robotics-on-device/ (blog.google)

Signal extension: Apptronik says Apollo 2 data from Robot Park supports Gemini Robotics model development through its Google DeepMind partnership. This is important because Google is gaining access to a real humanoid data loop rather than only lab datasets. Source: Apptronik Robot Park, 2026-06-30 — https://apptronik.com/news-collection/welc…REDACTED (apptronik.com)

Humanoids: deployment signals are strongest at Figure and Apptronik; Tesla remains high-upside/high-execution-risk

Figure AI

Signal: BMW announced it is advancing Physical AI in production with the Figure 03 project at Spartanburg, after an earlier Figure 02 pilot in the body shop; BMW says Figure 03 is aimed at a logistics sequencing use case. Source: BMW Group press release, 2026-06-25 — https://www.press.bmwgroup.com/canada/article/detail/T0458794EN/bmw-…REDACTED (press.bmwgroup.com)

Signal: Figure announced a commercial agreement with Catalyst Brands to deploy humanoids into its Reno, Nevada distribution/logistics center, starting with physically demanding supply-chain tasks. Source: Figure, 2026-05-26 — https://www.figure.ai/news/figu…REDACTED (figure.ai)

Signal quality: Better than demo. Still not full proof of scalable unit economics, but BMW + Catalyst are deployment/data-flywheel signals.

Apptronik

Signal: Apptronik closed a $520M Series A-X extension in February 2026, bringing total Series A funding to more than $935M and total capital raised to nearly $1B; participants include Google, Mercedes-Benz, John Deere, QIA, B Capital, and others. Source: Apptronik, 2026-02-11 — https://apptronik.com/news-collection/appt…REDACTED (apptronik.com)

Signal: Apptronik opened its expanded Robot Park facility in Austin and says Apollo 2 robots are active across Robot Park and customer/partner sites, generating data for future humanoid systems with Google DeepMind. Source: Apptronik, 2026-06-30 — https://apptronik.com/news-collection/welc…REDACTED (apptronik.com)

Signal quality: Strong capital + real-world data loop; watch for confirmed paid deployments, uptime, task success, and maintenance cost.

Tesla Optimus

Signal: Tesla’s Q1 2026 shareholder material listed Optimus manufacturing capacity as “construction” in California and Texas, and said preparations for the first large-scale Optimus factory would begin shortly in Q2. Source: Tesla Q1 2026 update PDF — https://ir.tesla.com/_flysystem/s3/sec/000162828026026551/tsla-20260422-gen.pdf (ir.tesla.com)

Signal: Tesla’s Q2 2026 coverage indicates the Model S/X Fremont lines were decommissioned to install first-generation Optimus production lines, with production expected to begin later in 2026. Source: Axios, 2026-07-22 — https://www.axios.com/newsletters/axio…REDACTED; Investing.com Q2 deck recap — https://www.investing.com/news/company-news/tesl…REDACTED (axios.com)

Signal quality: Huge upside if Tesla converts manufacturing competence into robot volume; current evidence remains production-line preparation, not proven high-volume useful work. Treat TSLA robotics as optionality, not confirmed earnings stream.

Unitree

Signal: Unitree received regulatory approval for a Shanghai IPO planned to raise about 4.2B yuan / ~$619M; this could create one of the first listed pure-play humanoid/quadruped robotics exposures from China. Source: Reuters via MarketScreener, 2026-07-03 — https://www.marketscreener.com/news/chin…REDACTED (marketscreener.com)

1X

Signal: 1X launched a World Model Lab in 2026 focused on scaling humanoid intelligence, with an explicit loop from pretraining and robot learning to evaluation, deployment, and on-policy real robot data. Source: 1X, 2026 — https://www.1x.tech/discover/1x-world-model-lab (1x.tech)

Signal: 1X opened a Hayward, California factory and reportedly aims to build home humanoids at scale, with NEO preorder / production claims covered by the LA Times. Source: LA Times, 2026-05-01 — https://www.latimes.com/business/story/2026-05-01/robo…REDACTED (latimes.com)

Signal quality: Strategically important but consumer-home robots are harder than warehouse/factory deployment; do not overweight until there is evidence of customer retention, safety, support, and real task autonomy.

---

Robot brains: Skild / Physical Intelligence / OpenAI / Meta / Luma

Signal: Skild AI reportedly raised a $1.4B Series C at a $14B+ valuation, with backers including SoftBank, NVIDIA, Macquarie, 1789 Capital, and Bezos Expeditions. Source: Bloomberg Law, 2026-01-14 — https://news.bloomberglaw.com/private-equity/robo…REDACTED (news.bloomberglaw.com)

Signal: Physical Intelligence has been reported in talks to raise about $1B at an $11B+ valuation after previously raising $600M at a $5.6B valuation. Source: TechCrunch, 2026-03-27 — https://techcrunch.com/2026/03/27/phys…REDACTED/ (techcrunch.com)

Signal: OpenAI’s 2026 supply-chain RFP explicitly includes robotics among U.S.-based manufacturing categories, and current OpenAI careers pages show robotics ML systems roles. This is not proof of a deployed OpenAI robot product, but it is evidence of renewed organizational interest in physical AI infrastructure. Sources: OpenAI, 2026-01-15 — https://openai.com/index/stre…REDACTED/; OpenAI careers — https://openai.com/careers/search/?c=c16e…REDACTED (openai.com)

Signal: Meta acquired humanoid robotics startup Assured Robot Intelligence to bolster humanoid AI ambitions, according to TechCrunch. Source: TechCrunch, 2026-05-01 — https://techcrunch.com/2026/05/01/meta…REDACTED/ (techcrunch.com)

Signal: Luma launched an Open Physical AI Lab focused on generalization and world models for physical interaction. Source: Luma, 2026-06-01 — https://lumalabs.ai/news/luma-open-physical-ai-lab (lumalabs.ai)

Defense robotics/autonomy: strongest deployment-to-capital conversion

Signal: Anduril raised $5B at a $61B valuation in May 2026 and is already reported to be seeking a round that could value it around $100B. Sources: TechCrunch, 2026-05-13 — https://techcrunch.com/2026/05/13/andu…REDACTED/; Axios, 2026-07-24 — https://www.axios.com/2026/07/24/andu…REDACTED (techcrunch.com)

Signal: The U.S. Air Force selected Anduril’s Lattice for Mission Autonomy software for the next phase of the Collaborative Combat Aircraft program; Anduril says its YFQ-44A mission autonomy stack first flew on February 24, 2026. Source: Anduril, 2026-06-17 — https://www.anduril.com/news/u-s-…REDACTED (anduril.com)

Signal: Shield AI raised $1.5B Series G at a $12.7B post-money valuation plus $500M preferred financing and announced an acquisition of Aechelon; Shield AI also says its Hivemind autonomy software integrated on Anduril’s Fury/YFQ-44A. Sources: Shield AI funding, 2026-03-26 — https://shield.ai/shie…REDACTED/; Shield AI CCA, 2026-02-13 — https://shield.ai/shie…REDACTED/ (shield.ai)

BCI / bionics / human-machine boundary

Signal: China approved Neuracle’s implantable BCI for hand motor augmentation, described by Nature Biotechnology as the world’s first commercial approval of an invasive BCI product; the system decodes brain signals to control a robotic glove worn by the patient. Source: Nature Biotechnology, 2026 — https://doi.org/10.1038/s41587-026-03101-8 (doi.org)

Signal: Nature Medicine published a 2026 neuroprosthesis study combining intracortical BCI with targeted spinal/cortical neuromodulation to restore clinically relevant hand function and sensation in severe paralysis. Source: Nature Medicine, 2026 — https://www.nature.com/articles/s41591-026-04498-0 (nature.com)

Signal: BrainCo demonstrated a brain-to-robot / BCI platform at WAIC 2026 that uses EEG-based intent decoding to control robotic limbs or humanoid robots, according to TechRadar and PC Gamer coverage. Sources: TechRadar, 2026-07-19 — https://www.techradar.com/ai-platforms-assistants/chin…REDACTED; PC Gamer, 2026-07-21 — https://www.pcgamer.com/hardware/my-s…REDACTED/ (techradar.com)

Signal quality: Medical BCI is crossing from trial/demo to regulated product in China; U.S./EU commercial pathways remain slower. Portfolio exposure is mostly private or medtech-adjacent; public medtech names may benefit indirectly through neuromodulation, surgical navigation, and implantable-device supply chains.

---

AI materials / autonomous labs: physical AI expands into scientific manufacturing

Signal: NSF-funded cloud-programmable labs are moving AI + robotics into materials discovery. UChicago/Argonne’s PoLARIS received a $20M NSF grant for AI, robotics, informatics, and standards across more than 20 autonomous stations; CUNY received $18.1M for AI-BIOME, a national lab for AI/robotics-enabled bio-inspired materials; Rice received nearly $20M for READINESS, focused on AI-driven autonomous experimentation for semiconductor synthesis. Sources: UChicago, 2026-07-22 — https://www.computerscience.uchicago.edu/news/remo…REDACTED/; CUNY, 2026-07-23 — https://asrc.gc.cuny.edu/headlines/2026/07/cuny…REDACTED/; Rice/EurekAlert, 2026-07-23 — https://www.eurekalert.org/news-releases/1137367 (computerscience.uchicago.edu)

Signal: Orbital Industries raised a $50M Series B to use AI for advanced materials discovery. Source: Fortune, 2026-05-28 — https://fortune.com/2026/05/28/excl…REDACTED/ (fortune.com)

CAPITAL_FLOW_MAP

Mega-rounds / valuation signals

Sources: Anduril TechCrunch/Axios — https://techcrunch.com/2026/05/13/andu…REDACTED/ and https://www.axios.com/2026/07/24/andu…REDACTED; Shield AI — https://shield.ai/shie…REDACTED/; Skild — https://news.bloomberglaw.com/private-equity/robo…REDACTED; Apptronik — https://apptronik.com/news-collection/appt…REDACTED; Physical Intelligence — https://techcrunch.com/2026/03/27/phys…REDACTED/; Unitree — https://www.marketscreener.com/news/chin…REDACTED; Humanoid — https://ca.marketscreener.com/news/robo…REDACTED; Gritt — https://www.gritt.ai/news/gritt-launches-physical-ai; Orbital — https://fortune.com/2026/05/28/excl…REDACTED/ (techcrunch.com)

Capital-flow interpretation

• Defense autonomy is receiving the most deployment-backed capital, because Anduril and Shield AI connect funding to programs, aircraft, autonomy stacks, and government demand.
• Humanoid capital is bifurcating: Tesla/Figure/Apptronik/Unitree/1X are body + deployment plays; Skild/Physical Intelligence/Google/NVIDIA are robot-brain and platform plays.
• Industrial incumbents are repositioning: ABB is divesting Robotics to SoftBank for $5.375B, and FANUC/Fujitsu/NVIDIA collaboration signals legacy industrial robotics is trying to absorb Physical AI rather than be disrupted by it. Sources: ABB — https://www.abb.com/global/en/news/129685/abb-…REDACTED; FANUC — https://www.fanuc.co.jp/en/profile/pr/newsrelease/2026/notice20260716.html (abb.com)

---

CONVERSION_LAYER

Demo → deployment conversion

Highest conversion evidence

1. Defense CCA autonomy: Anduril and Shield AI have named U.S. Air Force CCA autonomy roles and aircraft integration, giving the strongest non-consumer conversion signal. Sources: Anduril / Shield AI — https://www.anduril.com/news/u-s-…REDACTED and https://shield.ai/shie…REDACTED/ (anduril.com)
2. Figure at BMW + Catalyst: real commercial environments, especially logistics/manufacturing, but economics still unproven. Sources: BMW / Figure — https://www.press.bmwgroup.com/canada/article/detail/T0458794EN/bmw-…REDACTED and https://www.figure.ai/news/figu…REDACTED (press.bmwgroup.com)
3. Apptronik Robot Park: strong data flywheel and customer-site activity, but needs quantified paid deployment metrics. Source: Apptronik — https://apptronik.com/news-collection/welc…REDACTED (apptronik.com)
4. Tesla Optimus line conversion: strong manufacturing-intent signal; still needs real output, deployed hours, and task utility. Sources: Tesla Q1 PDF / Axios / Investing.com — https://ir.tesla.com/_flysystem/s3/sec/000162828026026551/tsla-20260422-gen.pdf; https://www.axios.com/newsletters/axio…REDACTED; https://www.investing.com/news/company-news/tesl…REDACTED (ir.tesla.com)

What to watch next

• Confirmed robot-hours worked in customer environments.
• Gross margin or lease economics per robot.
• Maintenance frequency, teleoperation ratio, and safety interventions.
• Dexterous manipulation performance outside scripted tasks.
• Robot fleet data ownership: who owns learning data, OEM or model provider?
• In defense: program-of-record expansion, production orders, autonomy safety certification, and export controls.

---

TIME_MAP

0–3 months

• NVIDIA Cosmos 3 Edge ecosystem adoption: watch Jetson Thor deployment references and partner announcements. Source: NVIDIA, 2026-07-20 — https://blogs.nvidia.com/blog/siggraph-news-2026/ (blogs.nvidia.com)
• Tesla Optimus production start / Fremont ramp verification. Sources: Axios / Investing.com — https://www.axios.com/newsletters/axio…REDACTED; https://www.investing.com/news/company-news/tesl…REDACTED (axios.com)
• Unitree IPO timing, accessibility, and valuation. Source: MarketScreener/Reuters — https://www.marketscreener.com/news/chin…REDACTED (marketscreener.com)
• Anduril potential new funding around ~$100B valuation. Source: Axios — https://www.axios.com/2026/07/24/andu…REDACTED (axios.com)

3–12 months

• ABB Robotics sale to SoftBank expected to close mid-to-late 2026, which may shift public robotics exposure from ABB toward SoftBank. Source: ABB — https://www.abb.com/global/en/news/129685/abb-…REDACTED (abb.com)
• Apptronik Apollo 2 customer-site deployments and Google DeepMind data loop proof. Source: Apptronik — https://apptronik.com/news-collection/welc…REDACTED (apptronik.com)
• Shield AI / Anduril CCA autonomy flight-demo progression and contract expansion. Sources: Anduril / Shield AI — https://www.anduril.com/news/u-s-…REDACTED; https://shield.ai/shie…REDACTED/ (anduril.com)
• BCI regulatory spread beyond China; U.S./EU approvals remain the key unlock. Sources: Nature Biotechnology / People’s Daily — https://doi.org/10.1038/s41587-026-03101-8; https://en.people.cn/n3/2026/0314/c90000-20436044.html (doi.org)

12–36 months

• Humanoid unit economics: leasing vs sale, support burden, insurance/safety regulation.
• Robot foundation model commoditization risk: open models from NVIDIA, Google ecosystem, Luma, and China-based labs may compress proprietary robot-brain valuations. Sources: NVIDIA Cosmos 3 paper / Luma — https://arxiv.org/abs/2606.02800; https://lumalabs.ai/news/luma-open-physical-ai-lab (arxiv.org)
• Autonomous labs: NSF/DOE-backed cloud labs may create materials-discovery infrastructure, but commercialization timing is likely longer than humanoid/defense timelines. Sources: UChicago / CUNY / Rice — https://www.computerscience.uchicago.edu/news/remo…REDACTED/; https://asrc.gc.cuny.edu/headlines/2026/07/cuny…REDACTED/; https://www.eurekalert.org/news-releases/1137367 (computerscience.uchicago.edu)

---

BOTTLENECK_INDEX

---

THESIS_STATE

Confirmed thesis upgrades

1. Physical AI infrastructure is becoming a real platform market. NVIDIA Cosmos 3 Edge and Google Gemini Robotics show credible platformization. Sources: NVIDIA / DeepMind — https://blogs.nvidia.com/blog/siggraph-news-2026/; https://deepmind.google/models/gemini-robotics/ (blogs.nvidia.com)

2. Deployment/data flywheels now matter more than demos. Figure’s BMW/Catalyst deployments and Apptronik’s Robot Park indicate the next moat is real-world operational data. Sources: BMW / Figure / Apptronik — https://www.press.bmwgroup.com/canada/article/detail/T0458794EN/bmw-…REDACTED; https://www.figure.ai/news/figu…REDACTED; https://apptronik.com/news-collection/welc…REDACTED (press.bmwgroup.com)

3. Defense autonomy is the most immediate conversion path. Anduril and Shield AI have capital, programs, and government pull. Sources: Anduril / Shield AI — https://www.anduril.com/news/u-s-…REDACTED; https://shield.ai/shie…REDACTED/ (anduril.com)

Thesis unchanged

• Dexterity remains the economic frontier.
• Humanoid form factor is promising but not yet proven as the dominant ROI-maximizing industrial shape.
• Real robot deployment, not viral video quality, determines investability.

Thesis risks

• Overvaluation of private robot-brain companies before revenue proof.
• Humanoid deployment may underperform due to maintenance, safety, teleoperation dependency, or task brittleness.
• Public investors may overpay for broad robotics ETFs with weak pure-play exposure.
• Defense autonomy has ethical/regulatory/export-control risk and procurement-cycle risk.

---

Strategic interpretation

The Physical AI market is separating into body companies, brain companies, infrastructure companies, and deployment owners. The investable center of gravity is currently infrastructure and deployment data, not humanoid aesthetics.

The most portfolio-relevant pattern is that the software layer may both matter enormously and commoditize quickly. NVIDIA and Google are pushing open or broadly accessible physical AI models and tooling; this could compress pure robot-brain startup valuations unless those startups own privileged data, deployment channels, or hardware integration. Sources: NVIDIA Cosmos 3 / Google Gemini Robotics / Luma — https://arxiv.org/abs/2606.02800; https://deepmind.google/models/gemini-robotics/; https://lumalabs.ai/news/luma-open-physical-ai-lab (arxiv.org)

The practical portfolio stance is: own infrastructure first, selectively own optionality second, and avoid confusing robotics-themed exposure with pure physical AI exposure. The near-term public market cannot fully capture Figure/Apptronik/Skild/Anduril/Shield AI because these are private, so exposure must be built through semis, industrial automation, defense, and carefully verified ETFs.

---

Sources

1. NVIDIA — SIGGRAPH 2026 / Cosmos 3 Edge — https://blogs.nvidia.com/blog/siggraph-news-2026/
2. NVIDIA — July 2026 investor deck — https://s201.q4cdn.com/141608511/files/doc_presentations/2026/07/NDR_July2026_.pdf
3. NVIDIA — Cosmos 3 arXiv — https://arxiv.org/abs/2606.02800
4. NVIDIA — GR00T N1.6 — https://research.nvidia.com/labs/gear/gr00t-n1_6/
5. Google DeepMind — Gemini Robotics — https://deepmind.google/models/gemini-robotics/
6. Google DeepMind — Gemini Robotics On-Device — https://deepmind.google/models/gemini-robotics/gemini-robotics-on-device/
7. Google — Gemini Robotics-ER 1.6 — https://blog.google/innovation-and-ai/models-and-research/google-deepmind/gemi…REDACTED/
8. DeepMind — Gemini Robotics-ER 1.6 model card — https://deepmind.google/models/model-cards/gemi…REDACTED/
9. Figure — News index — https://www.figure.ai/news
10. Figure — Catalyst Brands agreement — https://www.figure.ai/news/figu…REDACTED
11. BMW — Figure 03 Spartanburg project — https://www.press.bmwgroup.com/canada/article/detail/T0458794EN/bmw-…REDACTED
12. Apptronik — $935M Series A — https://apptronik.com/news-collection/appt…REDACTED
13. Apptronik — Robot Park — https://apptronik.com/news-collection/welc…REDACTED
14. Tesla — Q1 2026 update PDF — https://ir.tesla.com/_flysystem/s3/sec/000162828026026551/tsla-20260422-gen.pdf
15. Axios — Tesla AI/robotics spending and Optimus production context — https://www.axios.com/2026/07/22/tesl…REDACTED
16. Axios — Tesla Fremont / Optimus note — https://www.axios.com/newsletters/axio…REDACTED
17. Investing.com — Tesla Q2 2026 deck recap — https://www.investing.com/news/company-news/tesl…REDACTED
18. Unitree IPO — Reuters via MarketScreener — https://www.marketscreener.com/news/chin…REDACTED
19. 1X — World Model Lab — https://www.1x.tech/discover/1x-world-model-lab
20. LA Times — 1X Hayward factory / home robots — https://www.latimes.com/business/story/2026-05-01/robo…REDACTED
21. Skild AI — Bloomberg Law — https://news.bloomberglaw.com/private-equity/robo…REDACTED
22. Skild AI — TechCrunch — https://techcrunch.com/2026/01/14/robo…REDACTED/
23. Physical Intelligence — TechCrunch — https://techcrunch.com/2026/03/27/phys…REDACTED/
24. OpenAI — U.S. AI supply-chain RFP including robotics — https://openai.com/index/stre…REDACTED/
25. OpenAI careers — robotics ML role — https://openai.com/careers/search/?c=c16e…REDACTED
26. Meta / ARI robotics acquisition — TechCrunch — https://techcrunch.com/2026/05/01/meta…REDACTED/
27. Luma — Open Physical AI Lab — https://lumalabs.ai/news/luma-open-physical-ai-lab
28. Anduril — $5B raise — TechCrunch — https://techcrunch.com/2026/05/13/andu…REDACTED/
29. Anduril — reported ~$100B round — Axios — https://www.axios.com/2026/07/24/andu…REDACTED
30. Anduril — USAF CCA mission autonomy — https://www.anduril.com/news/u-s-…REDACTED
31. Shield AI — CCA mission autonomy — https://shield.ai/shie…REDACTED/
32. Shield AI — Series G / Aechelon — https://shield.ai/shie…REDACTED/
33. Nature Biotechnology — China BCI approval — https://doi.org/10.1038/s41587-026-03101-8
34. Nature Medicine — neuroprosthesis for hand movement/sensation — https://www.nature.com/articles/s41591-026-04498-0
35. People’s Daily — China invasive BCI market approval — https://en.people.cn/n3/2026/0314/c90000-20436044.html
36. TechRadar — BrainCo brain-to-robot interface — https://www.techradar.com/ai-platforms-assistants/chin…REDACTED
37. PC Gamer — WAIC 2026 BCI robotic control demo coverage — https://www.pcgamer.com/hardware/my-s…REDACTED/
38. UChicago — PoLARIS NSF autonomous materials lab — https://www.computerscience.uchicago.edu/news/remo…REDACTED/
39. CUNY — AI-BIOME NSF materials lab — https://asrc.gc.cuny.edu/headlines/2026/07/cuny…REDACTED/
40. Rice / EurekAlert — READINESS autonomous semiconductor materials lab — https://www.eurekalert.org/news-releases/1137367
41. Fortune — Orbital Industries $50M Series B — https://fortune.com/2026/05/28/excl…REDACTED/
42. Gritt — $32.4M physical AI infrastructure robotics launch — https://www.gritt.ai/news/gritt-launches-physical-ai
43. ABB — Robotics division sale to SoftBank — https://www.abb.com/global/en/news/129685/abb-…REDACTED
44. ABB — NVIDIA industrial physical AI partnership — https://www.abb.com/global/en/news/134030/prsr…REDACTED
45. FANUC — Fujitsu physical AI collaboration — https://www.fanuc.co.jp/en/profile/pr/newsrelease/2026/notice20260716.html
46. KOID — KraneShares Global Humanoid Robotics and Physical AI Index ETF — https://kraneshares.com/etf/koid/