Run metadata:

• ran_at: 2026-10-16
• job_id: physical_ai_model_to_matter
• task: Task 1 — Physical AI / Embodiment Scan + Task 2 — Ethan Portfolio Engine
• scope: physical AI / embodiment / robotics / bionics / AI materials / BCI / defense robotics / capital-flow signals
• verification: web-current; claims below are based on sources fetched in this run
• output intent: full report for sol/reports/2026…REDACTED.md; compact index follows for portfolio index

---

EXECUTIVE READ

The signal this cycle is not “humanoids are coming.” The signal is that the stack is separating into robot bodies, robot brains, robot data factories, and defense autonomy platforms. The strongest portfolio-relevant evidence is coming from:

1. production/deployment claims from Figure, Tesla, Boston Dynamics/Hyundai, Agility/NVIDIA;
2. robot foundation-model infrastructure from NVIDIA, Google DeepMind, Figure, Skild, Physical Intelligence, and 1X;
3. defense autonomy procurement and industrialization around Shield AI, Anduril, counter-UAS, and autonomous aircraft;
4. AI-for-materials / autonomous labs as the upstream “model → matter” layer;
5. BCI/bionics moving clinically, but still slower, regulated, and less investable through public equities.

Core thesis update: the near-term monetization path still favors infrastructure and defense autonomy over consumer humanoids. Humanoid demos are improving, but the clearest conversion to revenue/capex/contracts is currently in logistics/manufacturing pilots, autonomy software, robotic safety systems, defense UAS, counter-UAS, and compute/simulation platforms.

---

SIGNAL_LAYER

Humanoid production is moving from prototype theater toward fleet/data economics

Figure: BotQ + BMW + Helix 2.5 = strongest private humanoid deployment signal

Figure reports that BotQ produced over 350 Figure 03 robots and increased throughput from one robot per day to one robot per hour in under 120 days. Figure frames each robot as both a deployable unit and a data-collection engine for Helix. Source URL: https://www.figure.ai/news/ramp…REDACTED (figure.ai)

Figure also reports that Figure 03 arrived at BMW Group Plant Spartanburg in June 2026 after Figure 02 contributed to the assembly of 30,000 cars in 2025; the new BMW workflow involves logistics sequencing, cart pulling, whole-body motion, and manipulation rather than simple pick/place. Source URL: https://www.figure.ai/news/f-03-at-bmw (figure.ai)

Figure’s September 2026 Helix 2.5 update claims zero-shot generalization across 30 homes, with whole-body recovery and long-horizon task behavior; the important signal is not the home market itself, but the shift from site-specific robot programming toward generalized embodied policies. Source URL: https://www.figure.ai/news/heli…REDACTED (figure.ai)

Interpretation: Figure is becoming the clearest private “full-stack humanoid” benchmark: hardware manufacturing, real deployment, robot-brain model, fleet data, and capital intensity. The open question remains paid deployment economics, uptime, safety, and whether internal/company-published metrics translate into customer-scale revenue.

---

Tesla Optimus: mass-production intent visible, but still lower external verification

Tesla’s January 2026 investor materials state that Optimus Gen 3 is intended as the first mass-production design, with supply-chain readiness underway, start of production planned before the end of 2026, and eventual planned capacity of 1 million robots per year. Source URL: https://ir.tesla.com/_flysystem/s3/sec/000162828026003837/tsla-20260128-gen.pdf (ir.tesla.com)

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

Interpretation: Tesla remains a high-beta call option on humanoids because it has manufacturing scale, batteries, motors, AI infrastructure, and internal factory demand. But external deployment proof is still weaker than Figure/BMW-style visible customer workflow evidence. Treat Optimus as strategic optionality inside TSLA, not as a currently proven robotics revenue line.

---

Boston Dynamics / Hyundai: enterprise-ready Atlas, but timeline looks industrial, not consumer

Boston Dynamics says the electric Atlas is now oriented toward real-world industrial applications, beginning with Hyundai as testing ground, after commercial deployment experience with Spot and Stretch. Source URL: https://bostondynamics.com/blog/electric-new-era-for-atlas (bostondynamics.com)

Boston Dynamics’ Atlas evolution page describes Atlas as an enterprise-ready industrial humanoid and says the production-ready Atlas was unveiled with Hyundai in January 2026, with testing at Hyundai’s metaplant in Georgia. Source URL: https://bostondynamics.com/blog/atla…REDACTED/ (bostondynamics.com)

Interpretation: Atlas is moving from research halo to industrial product pathway. Hyundai ownership matters because factory access reduces go-to-market friction. Portfolio implication is more indirect: Hyundai/robotics ecosystem, factory automation, robot components, simulation, and safety.

---

Agility + NVIDIA Halos: safety layer becomes a monetizable bottleneck

NVIDIA announced Halos for Robotics in June 2026 as a full-stack safety system for robotics and physical AI; NVIDIA says Agility is the first to use it for Digit robots working in factories, warehouses, and logistics operations for customers including Amazon, GXO, Schaeffler, and Toyota Motor Manufacturing Canada. Source URL: https://investor.nvidia.com/news/press-release-details/2026/NVID…REDACTED/default.aspx (investor.nvidia.com)

AP reported in June 2026 that Agility Robotics planned a merger valuing it at $2.5B, positioning it as a public pure-play humanoid route focused on warehouse/industrial Digit deployments. Source URL: https://apnews.com/article/39f2…REDACTED (apnews.com)

Interpretation: The key signal is not only Digit; it is that safety certification and human-detection stacks are becoming core infrastructure. This favors NVIDIA and certified industrial deployment players over purely demo-driven humanoid names.

---

Robot-brain layer is accelerating faster than robot hardware

NVIDIA: GR00T / Cosmos / Isaac = platform play for embodied AI

NVIDIA’s July 2026 technical blog says Isaac GR00T unifies humanoid robot data collection, model training, evaluation, and deployment readiness; it cites ~32K hours of real and ~8K hours of simulated data in the GR00T N1.6 training stack, with ONNX and TensorRT export support. Source URL: https://developer.nvidia.com/blog/deve…REDACTED/ (developer.nvidia.com)

NVIDIA’s newsroom states that Cosmos and GR00T open models/data support robot learning, reasoning, simulation, evaluation, and humanoid full-body control through GR00T N1.6 and Cosmos Reason. Source URL: https://nvidianews.nvidia.com/news/nvid…REDACTED?linkId=100000401177632&ncid=so-link-151530-vt48 (nvidianews.nvidia.com)

NVIDIA also announced an Isaac GR00T reference humanoid robot in June 2026, built on Jetson Thor and intended for research institutions including Ai2, ETH Zurich, Stanford Robotics Center, and UC San Diego. Source URL: https://investor.nvidia.com/news/press-release-details/2026/NVID…REDACTED/default.aspx (investor.nvidia.com)

Interpretation: NVIDIA is not only selling GPUs; it is positioning itself as the operating substrate for physical AI: simulation, synthetic data, robot policies, edge deployment, safety, and reference hardware. This remains the highest-quality public-market exposure to the robot-brain/platform layer.

---

Google DeepMind: Gemini Robotics 2 pushes whole-body intelligence

Google DeepMind’s July 2026 Gemini Robotics 2 post states that the model expands from upper-body/tabletop tasks into whole-body physical AI and is available through Google AI Studio / Gemini Enterprise paths. Source URL: https://deepmind.google/blog/gemi…REDACTED/ (deepmind.google)

Google’s Gemini Robotics ER 2 page says developers can access the model through Gemini API, Google AI Studio, or Gemini Enterprise Agent Platform to build physical AI agents. Source URL: https://blog.google/innovation-and-ai/models-and-research/google-deepmind/gemi…REDACTED/ (blog.google)

Interpretation: Google is moving robotics from lab-only research toward developer/platform distribution. Portfolio implication is indirect but meaningful: GOOGL has a credible robot-brain software layer, but monetization timing is uncertain compared with NVIDIA’s immediate compute/simulation monetization.

---

Physical Intelligence: π0.7 validates cross-embodiment robot foundation model direction

Physical Intelligence’s April 2026 arXiv paper on π0.7 presents a steerable generalist robotic foundation model that can follow diverse language instructions in unseen environments, show zero-shot cross-embodiment generalization, and operate tasks such as laundry folding and espresso-machine operation without task-specific training. Source URL: https://arxiv.org/abs/2604.15483 (arxiv.org)

Interpretation: π0.7 reinforces the thesis that “robot brains” may become portable across hardware. The investable public angle is not Physical Intelligence directly, but compute, data, simulation, edge AI, and potential acquirers/partners.

---

Skild AI: capital validates cross-hardware robot foundation model demand

Skild AI announced a Series C in January 2026 led by SoftBank, with participation from NVentures/NVIDIA, Macquarie Capital entities, and Bezos Expeditions, valuing the company at over $14B; Skild describes its product as a unified robotics foundation model generalizing across tasks and robot hardware. Source URL: https://skild.ai/blogs/series-c (skild.ai)

Interpretation: Skild is a capital-flow signal that the market is pricing “robot brain” companies like frontier AI labs. This increases the strategic value of robotics datasets, simulation platforms, robot OS/middleware, and edge inference.

---

1X: world-model route to home humanoids

1X announced the 1X World Model Lab in June 2026 to scale humanoid intelligence through embodied world-model pretraining; its stated recipe combines web-scale media, egocentric human video, simulation, teleoperated robot data, and on-policy NEO data. Source URL: https://www.1x.tech/discover/1x-world-model-lab (1x.tech)

1X’s NEO order page says U.S. deliveries start in 2026 and describes NEO as using Redwood AI for learning and repeating tasks. Source URL: https://www.1x.tech/order (1x.tech)

Interpretation: 1X is a serious watchlist name, but consumer-home humanoids remain a harder monetization environment than warehouses, factories, defense, and controlled industrial workflows.

---

CAPITAL_FLOW_MAP

Private capital and valuation signals

Sources: Figure Series C URL https://www.figure.ai/news/series-c (figure.ai); Skild Series C URL https://skild.ai/blogs/series-c (skild.ai); Anduril Axios URL https://www.axios.com/2026/03/04/andu…REDACTED (axios.com); Anduril contract URL https://govtribe.com/award/federal-idv-award/inde…REDACTED (govtribe.com); Shield AI URL https://shield.ai/shie…REDACTED/ (shield.ai); Defense One URL https://www.defenseone.com/business/2026/10/shie…REDACTED/416510/ (defenseone.com); Agility AP URL https://apnews.com/article/39f2…REDACTED (apnews.com); Unitree AP URL https://apnews.com/article/f33f…REDACTED (apnews.com); NSF URL https://www.nsf.gov/news/nsf-…REDACTED (nsf.gov)

---

CONVERSION_LAYER

Highest conversion: defense autonomy

Shield AI says X-BAT production and test sites are being established in Washington and Kansas, with first VTOL flights scheduled for 2026 and mission capability targeted for 2028. Source URL: https://shield.ai/x-bat/ (shield.ai)

Shield AI announced on October 8, 2026 that Shield + U.S. government investment in X-BAT reached $400M. Source URL: https://shield.ai/shie…REDACTED/ (shield.ai)

Defense One reported that Shield AI expects to fly its X-BAT prototype by year-end 2026 and that a Navy $150M award, combined with prior $50M and Shield matching funds, brings total planned development spending to $400M. Source URL: https://www.defenseone.com/business/2026/10/shie…REDACTED/416510/ (defenseone.com)

Shield AI was also selected by the U.S. Navy to compete for up to $800M in ISR services using V-BAT. Source URL: https://shield.ai/shie…REDACTED/ (shield.ai)

Interpretation: Defense autonomy has the clearest path from product → procurement → production → sustainment. X-BAT is not yet a proven production aircraft, but the capital/procurement signal is stronger than most commercial humanoid signals.

---

Medium conversion: industrial humanoids/logistics

Figure/BMW, Agility/Digit, Boston Dynamics/Hyundai, and Tesla internal factory deployment are all converging around controlled industrial environments. The strongest evidence remains customer/factory workflows and production-rate claims rather than general consumer use. Sources: Figure BMW URL https://www.figure.ai/news/f-03-at-bmw (figure.ai); NVIDIA Halos/Agility URL https://investor.nvidia.com/news/press-release-details/2026/NVID…REDACTED/default.aspx (investor.nvidia.com); Boston Dynamics URL https://bostondynamics.com/blog/atla…REDACTED/ (bostondynamics.com); Tesla URL https://ir.tesla.com/_flysystem/s3/sec/000162828026026551/tsla-20260422-gen.pdf (ir.tesla.com)

Interpretation: This is now above demo-level. But the bottleneck is still utilization economics: uptime, safety certification, task coverage, maintenance, and cost per useful hour versus humans or conventional automation.

---

Lower near-term conversion: consumer humanoids

1X and Figure both point toward home robotics, and Figure’s Helix 2.5 home generalization signal is technically meaningful. But the home is an unstructured safety/regulatory/support nightmare compared with warehouses and factories. Sources: Figure Helix 2.5 URL https://www.figure.ai/news/heli…REDACTED (figure.ai); 1X NEO URL https://www.1x.tech/order (1x.tech)

Interpretation: watch technically; do not underwrite near-term consumer humanoid revenue until shipping volumes, failure rates, support costs, and retention are visible.

---

Emerging conversion: AI materials / autonomous labs

NSF announced $50M for two Materials Innovation Platforms, including an Autonomous Robotic Metallurgist Materials Innovation Platform at Texas A&M that integrates robotics and AI into an autonomous lab for alloy discovery. Source URL: https://www.nsf.gov/news/nsf-…REDACTED (nsf.gov)

Rice announced a nearly $20M NSF award for READINESS, an AI-powered autonomous experimentation platform for semiconductor, electronic, and quantum materials synthesis. Source URL: https://news.rice.edu/news/2026/acce…REDACTED (news.rice.edu)

Texas A&M describes ARM-MIP as a national facility where robots and AI automate repetitive alloy discovery work and release data/software/methods publicly. Source URL: https://experts.tamu.edu/news/2026/08/05/texa…REDACTED/ (experts.tamu.edu)

Interpretation: AI materials are early, but strategically important for robotics because actuation, batteries, heat, motors, semiconductors, lightweight structures, and tactile materials are all physical-AI bottlenecks.

---

Clinical conversion: BCI / bionics

Paradromics announced in September 2026 that its first Connexus BCI participant used the fully implanted intracortical BCI for real-time speech and text communication under an FDA-approved IDE study. Source URL: https://paradromics.com/news/para…REDACTED/ (paradromics.com)

Paradromics also announced in August 2026 FDA approval to expand Connect-One study access to compatible personal devices such as laptops, tablets, and phones. Source URL: https://paradromics.com/news/para…REDACTED/ (paradromics.com)

Synchron’s site states its COMMAND trial completed 12 months of safety follow-up under FDA IDE oversight and that current trials focus on people with paralysis in both arms. Source URL: https://synchronbci.com/ (synchronbci.com)

Shirley Ryan AbilityLab reported in September 2026 a milestone for the first U.S. patient toward a bionic arm anchored to bone, nerves, and muscles, including implanted nerve cuffs intended to provide sensory feedback. Source URL: https://www.sralab.org/articles/press-release/firs…REDACTED (sralab.org)

Interpretation: BCI/bionics are moving from science fiction toward regulated medical-device reality. But portfolio conversion is slower than defense/industrial robotics because reimbursement, trial design, surgical adoption, and FDA pathway dominate.

---

TIME_MAP

0–6 months

• Watch whether Tesla confirms Optimus start-of-production before end-2026, as stated in January materials. URL: https://ir.tesla.com/_flysystem/s3/sec/000162828026003837/tsla-20260128-gen.pdf (ir.tesla.com)
• Watch Shield AI X-BAT first flight / vertical flight milestones before year-end 2026. URL: https://shield.ai/x-bat/ (shield.ai)
• Watch Figure for sustained BotQ throughput beyond one robot/hour and actual customer deployment hours, not just production count. URL: https://www.figure.ai/news/ramp…REDACTED (figure.ai)
• Watch Agility public transaction completion and first reporting-quality metrics if it becomes investable as a pure play. URL: https://apnews.com/article/39f2…REDACTED (apnews.com)

6–18 months

• Defense autonomy procurement likely accelerates through X-BAT, V-BAT, counter-UAS, and autonomous ISR. Sources: Shield AI Navy ISR URL https://shield.ai/shie…REDACTED/ (shield.ai); X-BAT funding URL https://shield.ai/shie…REDACTED/ (shield.ai)
• Robot foundation models will compete on cross-embodiment transfer, data volume, simulation realism, and edge deployment. Sources: NVIDIA GR00T URL https://developer.nvidia.com/blog/deve…REDACTED/ (developer.nvidia.com); Physical Intelligence π0.7 URL https://arxiv.org/abs/2604.15483 (arxiv.org)
• AI materials/autonomous labs should start producing better benchmark evidence, not just grants. Sources: NSF URL https://www.nsf.gov/news/nsf-…REDACTED (nsf.gov); Rice URL https://news.rice.edu/news/2026/acce…REDACTED (news.rice.edu)

18–36 months

• Shield AI targets X-BAT mission capability by 2028. URL: https://shield.ai/x-bat/ (shield.ai)
• Boston Dynamics/Hyundai industrial Atlas deployment pathway becomes more testable through factory pilots and workcell integration. URL: https://bostondynamics.com/blog/atla…REDACTED/ (bostondynamics.com)
• BCI may show more real-world patient utility, but broad commercial monetization remains gated by trials, reimbursement, and surgical scaling. Sources: Paradromics URL https://paradromics.com/news/para…REDACTED/ (paradromics.com); Synchron URL https://synchronbci.com/ (synchronbci.com)

---

BOTTLENECK_INDEX

Dexterity and manipulation — HIGH

Walking is increasingly less differentiating; dexterous manipulation, tactile sensing, bimanual coordination, and recovery from errors remain the economic frontier. Figure’s own Figure 03 launch emphasized redesigned hands, palm cameras, sensory suite, and lower-latency vision for Helix. Source URL: https://www.figure.ai/news/intr…REDACTED?id=Figure03 (figure.ai)

Useful robot data — HIGH

Figure explicitly frames larger fleet size as a data-generation engine for Helix. Source URL: https://www.figure.ai/news/ramp…REDACTED (figure.ai)

1X’s World Model Lab explicitly names web-scale media, egocentric human videos, sim, teleoperated robot data, and on-policy NEO data as the scaling recipe. Source URL: https://www.1x.tech/discover/1x-world-model-lab (1x.tech)

Safety/certification — HIGH

NVIDIA’s Halos announcement and Agility integration point to safety as a gating layer for real deployment in warehouses, factories, and logistics. Source URL: https://investor.nvidia.com/news/press-release-details/2026/NVID…REDACTED/default.aspx (investor.nvidia.com)

Manufacturing yield and supply chain — HIGH

Figure says scaling Figure 03 required a new supply chain and vertical integration across actuators, batteries, sensors, structures, and electronics. Source URL: https://www.figure.ai/news/intr…REDACTED?id=Figure03 (figure.ai)

Tesla says Optimus production readiness includes supply-chain readiness and first production line preparations. Source URL: https://ir.tesla.com/_flysystem/s3/sec/000162828026003837/tsla-20260128-gen.pdf (ir.tesla.com)

Energy density / thermal / materials — MEDIUM-HIGH

Autonomous labs and AI materials platforms are receiving public funding because material discovery cycles are too slow for the needs of semiconductors, alloys, extreme-condition materials, and next-generation physical systems. Sources: NSF URL https://www.nsf.gov/news/nsf-…REDACTED (nsf.gov); Rice URL https://news.rice.edu/news/2026/acce…REDACTED (news.rice.edu)

Medical regulation / reimbursement — HIGH for BCI/bionics

Paradromics and Synchron are progressing clinically, but both remain within FDA/IDE and medical-device adoption constraints. Sources: Paradromics URL https://paradromics.com/news/para…REDACTED/ (paradromics.com); Synchron URL https://synchronbci.com/ (synchronbci.com)

THESIS_STATE

Confirmed

• Physical AI is splitting into layers: model/platform, robot body, data flywheel, safety/certification, deployment operations, and domain-specific use cases.
• Defense autonomy has the strongest near-term conversion to funded programs.
• Industrial/logistics humanoids are more credible than consumer home humanoids for near-term monetization.
• NVIDIA is the strongest public-market infrastructure proxy.
• Robot foundation models are becoming a major capital category.
• AI materials/autonomous labs are upstream to the physical-AI supply chain.

Strengthened

• Model → Matter thesis strengthened by AI-materials funding and autonomous labs.
• Robot-brain thesis strengthened by GR00T, Gemini Robotics 2, π0.7, Skild, Figure Helix, and 1X world models.
• Defense autonomy thesis strengthened by Shield AI and Anduril procurement/capital signals.
• Humanoid manufacturing thesis strengthened by Figure BotQ and Tesla production-line preparation.

Not yet proven

• Humanoid gross margins.
• Humanoid customer retention and uptime at scale.
• Home robot support economics.
• Whether robot foundation models become winner-take-most platforms or commoditized open models.
• Whether Tesla converts Optimus plans into credible external deployment.
• Whether BCI becomes investable through public equities within a reasonable time horizon.

Thesis adjustment

Previous posture: “humanoids are an emerging frontier; infrastructure is safer.”
Current posture: “physical AI is now a layered industrial buildout; overweight infrastructure and defense autonomy, selectively track humanoid deployment leaders, avoid demo-only names.”

---

STRATEGIC INTERPRETATION

The best investable layer is still infrastructure

The broadest capture layer remains compute, simulation, edge deployment, safety, and developer tooling. NVIDIA has the cleanest alignment because physical AI demand touches data centers, simulation, robot training, VLA models, edge inference, and safety. Sources: NVIDIA GR00T URL https://developer.nvidia.com/blog/deve…REDACTED/ (developer.nvidia.com); NVIDIA Halos URL https://investor.nvidia.com/news/press-release-details/2026/NVID…REDACTED/default.aspx (investor.nvidia.com)

Defense autonomy is the fastest conversion channel

Commercial humanoids need to beat labor economics; defense autonomy needs to solve urgent capability gaps. Shield AI and Anduril show how autonomy converts into contracts, development money, and industrial facilities faster than consumer robotics. Sources: Shield X-BAT URL https://shield.ai/shie…REDACTED/ (shield.ai); Anduril contract URL https://govtribe.com/award/federal-idv-award/inde…REDACTED (govtribe.com)

Humanoid winners will likely be data/manufacturing operators, not demo leaders

Figure’s BotQ + BMW + Helix loop is the pattern to watch: manufacturing capacity, deployment, real-world data, model improvement, more deployment. Source URL: https://www.figure.ai/news/ramp…REDACTED (figure.ai)

AI materials is early but structurally important

Physical AI cannot scale on software alone. Better batteries, actuators, semiconductors, heat dissipation, tactile skins, alloys, and lightweight structures matter. NSF/Rice/Texas A&M signals show institutional capital moving into autonomous materials discovery. Sources: NSF URL https://www.nsf.gov/news/nsf-…REDACTED (nsf.gov); Rice URL https://news.rice.edu/news/2026/acce…REDACTED (news.rice.edu)

BCI/bionics remain long-cycle strategic optionality

The clinical breakthroughs are real, but investable public-market routes remain unclear. The field belongs on the watchlist, not as a core near-term portfolio driver unless Eth has specific medtech/neurotech exposure available. Sources: Paradromics URL https://paradromics.com/news/para…REDACTED/ (paradromics.com); Shirley Ryan URL https://www.sralab.org/articles/press-release/firs…REDACTED (sralab.org)

---

SOURCES

1. Figure — “Ramping Figure 03 Production” — https://www.figure.ai/news/ramp…REDACTED (figure.ai)
2. Figure — “F.03 Arrives at BMW” — https://www.figure.ai/news/f-03-at-bmw (figure.ai)
3. Figure — “Helix 2.5: Zero-Shot 30-Home Generalization” — https://www.figure.ai/news/heli…REDACTED (figure.ai)
4. Figure — “Introducing Figure 03” — https://www.figure.ai/news/intr…REDACTED?id=Figure03 (figure.ai)
5. Figure — “Figure Exceeds $1B in Series C Funding at $39B Post-Money Valuation” — https://www.figure.ai/news/series-c (figure.ai)
6. Tesla investor material, Jan. 2026 — https://ir.tesla.com/_flysystem/s3/sec/000162828026003837/tsla-20260128-gen.pdf (ir.tesla.com)
7. Tesla investor material, Apr. 2026 — https://ir.tesla.com/_flysystem/s3/sec/000162828026026551/tsla-20260422-gen.pdf (ir.tesla.com)
8. Boston Dynamics — “An Electric New Era for Atlas” — https://bostondynamics.com/blog/electric-new-era-for-atlas (bostondynamics.com)
9. Boston Dynamics — “Atlas’ Evolution From Research Robot to Industrial Humanoid” — https://bostondynamics.com/blog/atla…REDACTED/ (bostondynamics.com)
10. NVIDIA — “Develop Humanoid Robot Policies End-to-End with NVIDIA Isaac GR00T” — https://developer.nvidia.com/blog/deve…REDACTED/ (developer.nvidia.com)
11. NVIDIA — “NVIDIA Releases New Physical AI Models…” — https://nvidianews.nvidia.com/news/nvid…REDACTED?linkId=100000401177632&ncid=so-link-151530-vt48 (nvidianews.nvidia.com)
12. NVIDIA — “Halos for Robotics” — https://investor.nvidia.com/news/press-release-details/2026/NVID…REDACTED/default.aspx (investor.nvidia.com)
13. NVIDIA — “Isaac GR00T Reference Humanoid Robot” — https://investor.nvidia.com/news/press-release-details/2026/NVID…REDACTED/default.aspx (investor.nvidia.com)
14. Google DeepMind — “Gemini Robotics 2 brings whole body intelligence to robots” — https://deepmind.google/blog/gemi…REDACTED/ (deepmind.google)
15. Google — “Gemini Robotics ER 2” — https://blog.google/innovation-and-ai/models-and-research/google-deepmind/gemi…REDACTED/ (blog.google)
16. Physical Intelligence — π0.7 arXiv paper — https://arxiv.org/abs/2604.15483 (arxiv.org)
17. Skild AI — “Announcing Series C” — https://skild.ai/blogs/series-c (skild.ai)
18. 1X — “The 1X World Model Lab” — https://www.1x.tech/discover/1x-world-model-lab (1x.tech)
19. 1X — “Order NEO” — https://www.1x.tech/order (1x.tech)
20. Shield AI — “X-BAT” — https://shield.ai/x-bat/ (shield.ai)
21. Shield AI — “Shield AI and U.S. government investment in X-BAT reaches $400 million” — https://shield.ai/shie…REDACTED/ (shield.ai)
22. Defense One — “Shield AI expects to fly its unmanned fighter jet within months” — https://www.defenseone.com/business/2026/10/shie…REDACTED/416510/ (defenseone.com)
23. Shield AI — “Selected by U.S. Navy to compete for $800M in ISR services with V-BAT” — https://shield.ai/shie…REDACTED/ (shield.ai)
24. Shield AI — “GE Aerospace light off X-BAT engine…” — https://shield.ai/shie…REDACTED/ (shield.ai)
25. Shield AI — “V-BAT operational trust” — https://shield.ai/how-…REDACTED/ (shield.ai)
26. Anduril / GovTribe — Anduril IDIQ contract — https://govtribe.com/award/federal-idv-award/inde…REDACTED (govtribe.com)
27. Axios — Anduril raising $4B at $60B valuation — https://www.axios.com/2026/03/04/andu…REDACTED (axios.com)
28. AP — Agility Robotics planned $2.5B transaction — https://apnews.com/article/39f2…REDACTED (apnews.com)
29. AP — Unitree Shanghai debut / humanoid shipment context — https://apnews.com/article/f33f…REDACTED (apnews.com)
30. NSF — $50M Materials Innovation Platforms — https://www.nsf.gov/news/nsf-…REDACTED (nsf.gov)
31. Rice University — nearly $20M AI-powered materials lab — https://news.rice.edu/news/2026/acce…REDACTED (news.rice.edu)
32. Texas A&M — self-driving laboratory for metals — https://experts.tamu.edu/news/2026/08/05/texa…REDACTED/ (experts.tamu.edu)
33. Paradromics — real-time speech Connexus BCI — https://paradromics.com/news/para…REDACTED/ (paradromics.com)
34. Paradromics — FDA approval for personal devices in Connect-One — https://paradromics.com/news/para…REDACTED/ (paradromics.com)
35. Synchron — BCI home / trial status — https://synchronbci.com/ (synchronbci.com)
36. Shirley Ryan AbilityLab — bionic arm anchored to bone/nerves/muscles — https://www.sralab.org/articles/press-release/firs…REDACTED (sralab.org)