🤖The Platform Wars Begin - Physical AI Signal Report - February 2026

Executive Snapshot — February in 5 Minutes

Five signals that mattered:

The Big Picture: What Changed This Month

February answered a question that January only asked: who owns the software layer of Physical AI? NVIDIA wants to be the infrastructure and simulation backbone. Google — via both DeepMind/Atlas and now Intrinsic — wants to be the brain and development platform. The integration of Intrinsic reflects a broader industry shift toward embedding AI directly into physical systems such as factory robots, warehouse automation platforms, and other industrial equipment. These aren't product launches anymore. They're platform land grabs — and the winners will determine how the next decade of automation gets built.

Top Announcements: The February Shortlist

🔹 Google absorbs Intrinsic — Alphabet-owned Intrinsic, which builds AI models and software designed to make industrial robots more accessible, is joining Google. Intrinsic will remain a distinct entity but will work closely with Google DeepMind and tap into Gemini AI models and cloud services. Android for robots just got a Google-sized budget.

🔹 NVIDIA GTC Washington — AI as national infrastructure — Jensen Huang pitched not just faster chips or smarter robots, but an operating plan for how intelligence gets manufactured — in factories, on roads, and at grid scale. Foxconn is already using Omniverse to model and operate a 242,000 sq ft Houston plant assembling NVIDIA AI systems.

🔹 Agility Robotics × Toyota Canada — first scaled RaaS humanoid deal — In February 2026, Agility deployed more than 7 Digit units at Toyota Motor Manufacturing Canada in a Robot as a Service deal. Toyota's Canadian operation is one of the company's largest outside Japan. This is the commercial deployment benchmark everyone else will be measured against.

🔹 Figure AI's Helix 02 breakthrough — Figure's Helix 02 neural network controls the entire humanoid from pixels for long-horizon tasks like autonomous dishwasher unloading across rooms — no resets, integrating walking, manipulation, and balance seamlessly.

🔹 Boston Dynamics leadership transition — CEO departure mid-production ramp at a company scaling Atlas for Hyundai and Google DeepMind fleets. Continuity of execution is the key question for Q1.

🔹 NVIDIA + US manufacturing partners — Toyota is using digital twins of its Georgetown, Kentucky facility; TSMC is using Omniverse to accelerate fab design at its Phoenix, Arizona facility; Wistron is implementing digital testing at its Fort Worth, Texas facility.

This newsletter? It’s powered by the platform built for growth, monetization, and jaw-droppingly good reader experiences.

From sponsorships that actually pay you fairly to referral programs that grow your list on autopilot, beehiiv gives publishers, creators, and writers the tools to grow their newsletter like never before. And yeah, it is just that easy.

Power your content with beehiiv

Embodied AI & Robotics: The Deployment Story Is Now Here

Humanoids: RaaS is the unlock

The Agility-Toyota deal is the clearest signal yet that humanoids are entering the commercial phase — not as pilot curiosities but as service contracts with measurable unit economics. The useful metric for humanoid robots in 2026 is not whether they can do backflips — it is whether they can pick up a component, carry it across a factory floor, and place it correctly, thousands of times per day, without halting production. Agility is also working toward ISO functional safety certification for Digit, which would make it the first humanoid cleared to work alongside people without physical barriers.

The Platform Wars: Google vs NVIDIA

Just as Android runs across phones and tablets from many manufacturers, Intrinsic aims to do the same for robotic systems — partnering with FANUC, Universal Robots, and KUKA. Meanwhile, NVIDIA's GR00T, Cosmos, and Isaac stack is positioned as the training and simulation backbone that every robot maker runs on. Both strategies are credible. The question is whether the industry ends up with one platform or two.

Drone Delivery: The Quiet Commercial Winner

Wing is expanding its Walmart drone delivery operations, adding service to 150 more US stores and extending coverage from Los Angeles to Miami. By end of 2026, roughly 40 million Americans could have access to the service. In Dallas-Fort Worth — Wing's most mature market — some customers order multiple times per week. This is the most commercially mature Physical AI category right now, full stop.

AI in Manufacturing: Intelligence Moves Closer to the Machine

NVIDIA introduced Mega, a simulation framework that lets engineers test fleets of robots and factory controllers in virtual sync, and Halos, an overhead vision-AI system that automatically slows or stops machines when people cross their paths. The company has also created an ANSI-accredited Halos Inspection Lab to certify functional safety for physical AI — a sign that the industry is starting to build the governance infrastructure that enterprise buyers require before committing capex.

The deeper trend: manufacturers are moving from AI as analytics toward AI as operational control. The companies winning the next phase won't sell insights — they'll run the loop.

Winners, Losers & Watchlist

✅ Winners

⚠️ At Risk

👀 Watchlist for March

What to Watch in March 2026

Closing Note

February clarified the competitive map: Physical AI isn't a startup story anymore — it's a platform war between the largest technology companies on earth. NVIDIA owns the infrastructure and simulation layer. Google is making its move on the software and intelligence layer. And on the factory floor, the first real commercial deployments are setting the benchmarks that will define the entire decade. The builders who understand this convergence — and position themselves at the intersection — are the ones who will define what comes next.

Subscribe to the Physical AI Builders Newsletter → SVE.io Join the community → discord.gg/YAvNRdav

— Physical AI Builders Team

Further Reading