Robotics

Today's Stories

1. 1

   Walden Robotics launches with $1.1B valuation, $300M funding

   Walden Robotics, a full-stack Physical AI company building general-purpose robots that learn and improve through real work, raised $300 million（約480億円） in funding and launched out of stealth at a $1.1 billion（約1800億円） valuation. The round is co-led by Toyota (Toyota Motor Corp, Toyota Invention Partners, and Toyota Ventures) and Deviation Capital, with participation from NVIDIA, Boeing, AE Ventures, Samsung Ventures, Prologis Ventures, CoreWeave Ventures, and others. Walden's robots, which build on research including Diffusion Policy and Large Behavior Models (LBMs), have been performing useful work in production at a Toyota plant in North America since February—moving from first pilot to real work in under two months. Walden represents a major bet that general-purpose robots capable of continuous learning will drive the next wave of industrial automation. The company's ability to move from pilot to production use at scale (at a major automaker, no less) within weeks suggests that robotic systems are approaching practical viability for real manufacturing tasks. The backing of Toyota, NVIDIA, and Boeing signals confidence that in-house data collection, proprietary robotic AI models, and vertical integration in hardware are the winning formula for robot makers.

   Walden launched out of Toyota Research Institute in January 2026 and is already working with customers across multiple industries. The pace of deployment—from February pilot to real production work in under two months—and the roster of industrial and tech investors suggest the company will likely announce additional customer deployments and use cases in the near term.

2. 2

   Humanoid robotics relies on human operators, not AI breakthroughs

   The humanoid robotics industry has raised billions in the past 18 months, but the majority is funding human workers to operate robots remotely — a method called teleoperation — rather than building truly autonomous systems. Teleoperation datasets are over 100,000 times smaller than those used to train today's language and vision models, and the gap cannot close by hiring more operators because the real world constantly changes, requiring new demonstrations faster than humans can provide them. The original pitch for humanoid robots is to replace human labor due to aging populations and labor shortages. But if robots require a permanent stream of human demonstrations to function, they are essentially just a labor system rather than an autonomy solution. This means companies may be building infrastructure that deepens dependency on human workers indefinitely rather than resolving it.

   Reinforcement learning in simulation offers a potential alternative path. Unlike teleoperation, RL systems learn through trial and error across millions of iterations without human operators in the loop, and simulation allows that training to scale directly with compute — more GPUs enable faster iteration — rather than being limited by human labor availability.

3. 3

   Humanoid robots shift AI compute to the edge, away from cloud

   The humanoid robot industry is converging on a 'big brain, small brain' architecture, moving AI compute from the cloud to edge devices and endpoints like hands and feet. This shift distributes processing closer to where robots act, potentially reducing latency and cloud dependency—a structural change in how robot intelligence is deployed.

   DIGITIMES Intelligence predicts Nvidia's CUDA will remain dominant in this new architecture, though the article does not specify a timeline or further details on adoption.

4. 4

   Faraday Future robotics unit ships 220 units in H1, expands into industrial automation

   Faraday Future's AI subsidiary FFAI announced at Automate 2026 in Chicago that June robot shipments are expected to exceed 100 units, bringing total first-half deliveries to more than 220 units and surpassing its original target ahead of schedule. The company introduced the All-New Futurist humanoid robot priced from $89,900 and the FF Faber mobile manipulator series for industrial applications. FFAI is expanding beyond its education-focused robotics business into commercial and industrial automation, targeting warehouse logistics, factory operations, and facility inspection. The Futurist humanoid is the first full-size humanoid in the US to natively support Nvidia Sonic's full-body motion control system, signaling deeper integration with advanced motion control technology.

   FFAI's strategy centers on an open embodied AI ecosystem combining multiple robot form factors with a shared EAI Brain software platform and an EAI Data Factory that collects operational data from deployed robots to improve AI performance over time. The Futurist stands 5 ft 8 in tall, weighs 121 lb, and runs on a dual-battery system providing up to six hours of continuous operation.

5. 5

   Fujitsu partners with Fanuc, Yaskawa, Kawasaki on physical AI using Nvidia tech

   Fujitsu has begun collaboration with Japanese robotics makers Fanuc, Yaskawa Electric and Kawasaki Heavy Industries to develop and deploy physical AI—AI systems that control physical equipment—across manufacturing, logistics and healthcare. The partnership will use Nvidia's physical AI technologies, including Nvidia Cosmos foundation models, Nvidia Omniverse libraries, the Nvidia Isaac robotics platform and the Newton physics engine, to build a collaborative control platform that connects digital systems with robots and other equipment. The initiative addresses Japan's labor shortages, aging workforce and global manufacturing competition by automating tasks in factories (production planning and factory adaptation), logistics and retail (material handling), and healthcare (pharmaceutical transport, patient reception). Fujitsu CEO Takahito Tokita framed the effort as creating "a new social infrastructure in which people and robots work collaboratively across a wide range of industries." The partners plan to develop an open, sovereign platform designed to work across different robots while meeting cybersecurity, operational resilience and data protection standards.

   The companies will now develop a roadmap for technology development and commercialization, with the longer-term goal of expanding physical AI deployment globally and strengthening Japan's position in the robotics industry. Specific use cases include optimizing production planning in manufacturing, automating material handling in logistics and retail by combining planning with real-time sales and inventory data, and automating pharmaceutical and specimen transport plus patient assistance in healthcare.

6. 6

   Toyota and Nvidia expand AI partnership for vehicles, factories, cities

   Toyota and Nvidia broadened their partnership to deploy Nvidia's accelerated computing, AI software and simulation technologies across Toyota's vehicle development, software engineering, factory operations and intelligent transportation systems. The expansion builds on a prior agreement under which Toyota will develop next-generation vehicles with advanced driver-assistance capabilities using the Nvidia DRIVE AGX in-vehicle computing platform and Nvidia DriveOS operating system. The partnership extends AI beyond autonomous driving into manufacturing and urban infrastructure. Toyota is using Nvidia AI models to accelerate safety-critical automotive software engineering through a MISRA-compliant coding assistant based on Nvidia Megatron-LM and Nvidia Nemotron, and deploying digital twins of production environments using Nvidia Omniverse and Isaac Sim to optimize factory workflows before real-world deployment. Woven by Toyota has also developed a multimodal vision-language model using Nvidia H100 Tensor Core GPUs to analyze urban traffic and support city infrastructure decision-making.

   Future Toyota vehicles are designed to offer Level 2++ driver-assistance functionality. The partnership spans vehicle development, software engineering, factory operations and urban mobility technologies through Woven by Toyota.