In Entry 07 we wrote about the Windows Moment — the platform that would make robotics possible at scale. That entry was about software. This one is about hardware. The drones are flying. The humanoids are walking. The defence sector is deploying autonomous systems at a pace that has no historical precedent. The trillion-dollar robot economy is not a 2040 projection. It is a 2030 reality being built today.
When we wrote about the Windows Moment for Robotics, the argument was structural: the value in the robotics revolution would accrue not to the robot manufacturers but to the platform — the operating system that made a thousand manufacturers possible. Microsoft didn't make PCs. Google didn't make phones. The platform captured the value.
That thesis remains correct. But something has shifted since we wrote it. The hardware is arriving faster than the platform debate suggested. Humanoid robots are not prototypes in research labs anymore — they are on factory floors in China and the United States, doing repetitive manufacturing tasks that humans used to do. Drones are not hobbyist toys — they are delivering packages in suburban America, spraying crops across Indian farmland, and fighting wars in Ukraine and the Middle East. The transition from demonstration to deployment happened faster than almost anyone predicted.
Nvidia's Jensen Huang calls Physical AI — AI that interacts with the physical world through robots and autonomous systems — the next great platform shift. He believes the market for physical AI could eventually be worth tens of trillions. Barclays calls the coming years the Decade of the Robot. The numbers are beginning to support both claims.
"The transition from demonstration to deployment happened faster than anyone predicted. The drones are flying. The humanoids are walking. The Decade of the Robot is not a forecast. It has begun."
The low-altitude economy is a term that originated in China but describes a global phenomenon — the commercialisation of airspace below 1,000 metres for delivery, logistics, agriculture, inspection, and urban air mobility. China has moved furthest and fastest. The Chinese government has designated the low-altitude economy as a strategic national priority, projecting it to exceed RMB 1 trillion (~$140 billion) by 2026 and reach $1 trillion equivalent by 2034.
The use cases are more varied than most Western analyses acknowledge. Drone delivery — Amazon Prime Air, Wing, Zipline — is the most visible application but not the most economically significant. Agricultural drones — spraying, seeding, crop monitoring — are already deployed at massive scale across China, India, and Brazil. A single agricultural drone operator in rural India replaces 40 manual sprayers. The economics are not debatable. Infrastructure inspection — power lines, pipelines, bridges, wind turbines — using drones is dramatically cheaper and safer than human inspection. Entire inspection industries are being rebuilt around drone platforms.
eVTOLs — electric vertical take-off and landing vehicles — are the air taxi layer above drones. Joby Aviation, Archer, Lilium's successor companies, and China's EHang are all in advanced testing or early commercial deployment. The first urban air mobility routes — city centre to airport, hospital to hospital — will be commercially operational before 2028. The sky is becoming infrastructure in the same way that roads became infrastructure a century ago.
The humanoid robot is the most ambitious product category in the history of manufacturing. A machine that moves through human environments, manipulates objects designed for human hands, and performs tasks that vary unpredictably — all in real time, without supervision. For decades it was science fiction. In 2026 it is early commercial reality.
Tesla's Optimus is the most widely discussed. But the competitive landscape is richer and moves faster than Optimus alone. Figure AI has deployed humanoids in BMW manufacturing facilities. Apptronik is working with NASA and GE Aerospace. 1X Technologies — backed by OpenAI — is deploying in logistics warehouses. In China, the pace is extraordinary — BYD, CATL, Xiaomi, and more than 20 other Chinese companies have active humanoid programmes. Unitree's G1 humanoid retails at $16,000 — affordable enough that manufacturing companies can run genuine commercial pilots without research budgets.
The $200 billion by 2035 projection from Barclays is not aggressive. It assumes modest penetration of global manufacturing, logistics, and eventually domestic assistance markets. The realistic upside — if the platform thesis from Entry 07 proves correct and a dominant operating system emerges — is substantially larger. The humanoid robot is not a product. It is a general-purpose economic unit. The same robot that assembles cars today can stock shelves tomorrow and assist elderly people the day after.
The Ukraine conflict changed everything. For the first time in a major war between technologically capable adversaries, autonomous drone systems have played a decisive tactical role — not as support tools but as primary weapons. Ukrainian FPV drones — costing $400 each — have destroyed Russian tanks worth $3 million. The economics of autonomous warfare have inverted. The cheap autonomous system now defeats the expensive manned platform.
This reality is producing the fastest defence procurement shift since the introduction of precision-guided munitions in the 1970s. Every major military on earth is restructuring its procurement priorities around autonomous systems. The United States Replicator Initiative aims to field thousands of attritable autonomous systems within two years. The UK's Project MORPHEUS is integrating autonomous ground vehicles into infantry brigades. Israel's autonomous border patrol systems have been operational for years. China's drone swarm programmes are the most advanced in the world by volume.
The defence robotics market is growing at rates that conventional defence procurement never achieves — because the procurement cycle for a $400 drone is weeks, not decades. AeroVironment, Anduril, Shield AI, Joby (military variant), and dozens of smaller companies are building a new defence industrial base — one defined by software, autonomy, and attritable hardware rather than the exquisite manned platforms that dominated the previous century of warfare.
We state this observation without celebration. War is the greatest failure of human civilisation. But the economics of autonomous defence systems are changing the strategic calculus of every government on earth — and creating one of the fastest-growing investment categories of the decade. The honest investor must understand it.
Physical AI is Jensen Huang's term for the convergence of large language models, computer vision, sensor fusion, and real-time decision making into systems that interact with the physical world. It is the software platform that makes drones intelligent, humanoids capable, and autonomous vehicles safe. Without physical AI, a drone is a flying camera. With it, a drone is an autonomous agent that can plan routes, avoid obstacles, identify targets, and adapt to unexpected conditions without human instruction.
NVIDIA's Isaac platform is the leading physical AI infrastructure. AWS RoboMaker, Google's RT-2, and Microsoft's autonomous systems division are building competing platforms. But NVIDIA's hardware advantage — the same GPU dominance that powers language model training — translates directly to physical AI training and inference. The company that built the picks and shovels of the language AI revolution is building the picks and shovels of the physical AI revolution. The addressable market, in Jensen Huang's words, could eventually be worth tens of trillions.
The defence robotics market deserves its own honest treatment because it is growing faster than any other segment — and because most investment analysis avoids the ethical complexity by either ignoring it or treating it as purely commercial. The NGE position is this: we observe reality as it is, we state it plainly, and we note the ethical weight without pretending it does not exist.
The hardware is moving faster than the platforms. In Entry 07 we said the platform would capture the value. That remains true long-term. But near-term, the hardware companies — DJI, Unitree, Figure AI, AeroVironment — are generating real revenue from real deployment. Do not wait for the platform to emerge before taking the hardware opportunity seriously.
China's speed advantage in drones is structural. DJI controls approximately 70% of the global consumer and commercial drone market. The supply chain, the manufacturing expertise, the regulatory support — all concentrated in China. Western governments are attempting to restrict DJI for national security reasons. That restriction creates opportunity for Western alternatives (Skydio, Autel) but does not eliminate DJI's structural advantages in non-restricted markets.
Defence robotics is the fastest-growing but most ethically complex segment. Autonomous weapons systems raise questions about accountability, proportionality, and the laws of armed conflict that international law has not yet answered. These questions will eventually produce regulation. What form that regulation takes — and when — will materially affect the defence robotics investment thesis. This uncertainty must be priced.
The $1Q connection is Force 04 — AI Supercycle — and Force 11 — Robotics. But the deeper connection is this: the $1Q economy requires productivity growth at a rate that human labour alone cannot supply. Robots — in factories, in logistics, in agriculture, in construction — are the productivity multiplier that makes the quadrillion mathematically possible. The Decade of the Robot is not a technology story. It is an economic necessity.
Part of an ongoing journal — observations recorded when something in the world economy is worth saying. No schedule. No noise. Not investment advice. See also: Entry 07 — The Windows Moment for Robotics.