Plugging in AI Agent Factories

This week: Electricity demand and grid impacts of AI data centers (paid), manufacturing month, multi-modal robot learning the robobusiness, quest for a $30,000 EV, hydride 🔋, g-DLP

Oct 05, 2025

∙ Paid

Shop Talk

Capturing this week’s zeitgeist

This past week marked the USA’s celebration of Manufacturing Day on Friday, October 3rd, kicking off a full month dedicated to the innovation and opportunity within the manufacturing sector. Across the country, companies, trade organizations, and educational institutions opened their doors to inspire the next generation. Events range from hands-on foundry demonstrations and facility tours to expert-led symposiums and apprenticeship signing days, all designed to showcase modern, high-tech careers. Highlighting this effort, companies like Mastercam are offering free access to professional CAD/CAM training, providing invaluable resources to students and job seekers alike and demonstrating the industry’s collective commitment to building a robust talent pipeline for the future.

📅 Next up: RoboBusiness, October 15-16, Santa Clara, CA, USA which brings together the brightest minds in robotics to better engineer the future of robotics.

Assembly Line

This week’s Industry 5.0 breakthroughs and frontier technologies of the built world.

🤖🧠🔁 Into the Omniverse: Open-Source Physics Engine and OpenUSD Advance Robot Learning

NVIDIA announced at the Conference on Robot Learning this week groundbreaking advances in open-source physics simulation, open foundation models and development frameworks, including:

Newton Physics Engine: While robots learn faster and safer in simulation, humanoid robots — with complex joints, balance and movements — are pushing today’s physics engines to the limit. Codeveloped by Google DeepMind, Disney Research and NVIDIA, and managed by the Linux Foundation, Newton is an open-source, GPU-accelerated physics engine to advance robot learning. Built on NVIDIA Warp and OpenUSD, Newton enables robots to learn complex tasks more precisely while working seamlessly with robot learning frameworks like MuJoCo Playground and NVIDIA Isaac Lab.

Isaac GR00T N1.6: To perform humanlike tasks in the physical world, humanoids must understand ambiguous instructions and navigate unforeseen scenarios. The latest Isaac GR00T N1.6 open robot foundation model, available soon on Hugging Face, integrates NVIDIA Cosmos Reason, an open reasoning vision language model built for physical AI. Cosmos Reason serves as the robot’s deep-thinking brain and transforms vague instructions into step-by-step action plans using prior knowledge, common sense and physics understanding.

🚙🔋🗣️ Ford CEO Jim Farley on China, tariffs, and the quest for a $30,000 EV

Read the full transcript at The Verge

🔋 Battery with our anode chemistry can deliver up to 4,000 cycles at 100% depth of discharge: Ramesh Narasimhan, EVP and Chief Commercial Officer, Nyobolt /S&P AutoTechInsight/
🔋 Introducing the hydride battery /Chemical & Engineering News/
🔌 Can rapid-charging EVs eliminate ‘range anxiety’? /FT/
🚙🔋🗣️ Why General Motors Boss Mary Barra Is Slamming the Brakes on Lofty EV Ambitions /WSJ/
🚙🗜️ The Giga Casting Process at Giga Berlin - now even better /Tesla/
🛻🏭 Inside our Warsaw Factory /Slate/

🔜 Don’t miss the The Battery Show in Detroit, Michigan this week!

📐🖨️ Why Injection Molding Falls Short at the Micro Scale – And What Engineers Are Doing Instead

To overcome the cost, speed, and design limitations of traditional injection molding at the micro scale, engineers are adopting micro-precision 3D printing as a superior alternative. The key innovation is Projection Micro Stereolithography (PµSL), a proprietary additive manufacturing technology that bypasses the need for expensive and slow-to-produce molds. Its technical approach involves printing complex parts directly from CAD files, enabling the creation of intricate geometries like internal channels and sharp corners with tolerances as tight as ±10 μm. This results in significant business outcomes, including an over 80% reduction in iteration time, allowing for rapid prototyping and low-volume production that empowers industries like medtech and electronics to develop previously unmanufacturable components with greater speed and design freedom.

Read more at Boston Micro Fabrication

Zurich-based startup a-metal has closed its seed financing round led by QBIT Capital /startupticker.ch/

🖨️💡 Machine Learning-Driven Grayscale Digital Light Processing for Mechanically Robust 3D-Printed Gradient Materials

✍️ Authors: Jisoo Nam, Boxin Chen, Miso Kim

Grayscale digital light processing (g-DLP) is gaining recognition for its capability to create material property gradients within a single resin system, enabling programmable mechanical responses, enhanced shape accuracy, and improved toughness. However, research on the mechanical robustness of g-DLP is constrained by the limited range of tailorable properties in photocurable resins and insufficient exploration of structural optimization for complex geometries. This study presents a synergistic g-DLP strategy that integrates the synthesis of dynamic bond-controlled polyurethane acrylate (PUA) with a machine learning-based multi-objective optimization, enabling mechanically robust 3D-printed gradient materials. By combining the developed material with this optimization framework, a versatile platform is established for creating mechanically robust g-DLP printed components, applicable in areas ranging from biomimetic artificial cartilage to automotive energy-absorbing structures.

Read more at Advanced Materials and TechXplore

🖨️💡 Stereolithography-based 3D printing of silica with solutions without organic binders /Materials Today/

General Robots

Benjie's Humanoid Olympic Games

I was kinda disappointed by the World Humanoid Robot Games. As fun as real-life rock-em-sock-em robots is, what people really care about is robots doing their chores. This is why robot laundry folding videos are so popular. Those laundry videos are super impressive: we didn’t know how to do that even a few years ago. And it is certainly something that p…

a month ago · 21 likes · 2 comments · Benjie Holson

New Product Introduction

Highlighting new and innovative facilities, processes, products, and services

✈️ Boeing Has Started Working on a 737 MAX Replacement

✍️ Authors: Benjamin Katz, Drew FitzGerald

Boeing has initiated the early stages of development for a new narrow-body aircraft to replace the 737 MAX. This strategic move, aimed at regaining market share from its rival Airbus, involves discussions with Rolls-Royce for a new, more fuel-efficient engine, the design of a new flight deck, and the appointment of a new senior product chief. While facing ongoing challenges, including the certification of new 737 MAX models and delays in its 777X program, Boeing is simultaneously collaborating with NASA on a lighter, more aerodynamic wing design. This long-term project signals the company’s commitment to innovation and its bet on a cutting-edge plane to secure its future, with a potential delivery timeline for the new aircraft in the mid-2030s.

🏭 Launch of Demonstration Experiment for the Joint Development of Microwave-based Low-carbon Lithium Ore Refining Technology

Mitsui & Co. and Microwave Chemical Co. are pioneering a new, low-carbon method for refining lithium ore by utilizing microwave technology, an innovation aimed at significantly reducing the carbon footprint of lithium production. Their technical approach involves a demonstration experiment at a pilot plant in Osaka, where they are using a microwave heating unit to electrify the calcination process, a major source of CO2 emissions in traditional lithium refining. The intended business outcome is to establish a reliable and eco-friendly supply chain for low-carbon lithium, a critical component for EV batteries, with a goal for commercialization by 2030, thereby contributing to a carbon-neutral society and positioning microwave technology as a new global standard in the chemical industry.

DoorDash Unveils Dot, the Delivery Robot Powered by its Autonomous Delivery Platform to Accelerate Local Commerce

DoorDash has introduced “Dot,” an in-house developed autonomous delivery robot, and an “Autonomous Delivery Platform” to enhance local commerce. The key innovation is Dot’s ability to navigate various terrains like bike lanes, roads, and sidewalks at up to 20 mph, a feature designed specifically for the complexities of local delivery. The technical approach involves a sophisticated AI-powered platform that acts as a dispatcher, selecting the most efficient delivery method—be it a Dasher, a Dot robot, or a drone—based on real-time data like speed, cost, and location. This integrated system is designed to produce significant business outcomes, including helping merchants meet rising consumer demand, reducing traffic congestion and emissions, and allowing human Dashers to focus on higher-value deliveries, thereby improving the overall efficiency and reliability of the delivery network for merchants, consumers, and Dashers alike.

Business Transactions

This week’s top funding events, acquisitions, and partnerships across industrial value chains.

3M Might Just Escape Its Toxic Chemical Legacy /Bloomberg/
Woven Capital: Toyota’s VC Arm Reveals $800 Million Fund II /Pulse 2.0/
New industrial innovation: Why Europe needs more than SaaS /Silicon Canals/

🇺🇸🥼 Periodic Labs Raise $300M seed to Automate Science

✍️ Author: Julie Bort

Periodic Labs came out of stealth with a war chest of $300 million as a seed round, backed by a tech industry who’s who: Andreessen Horowitz, DST, Nvidia, Accel, Elad Gil, Jeff Dean, Eric Schmidt, and Jeff Bezos. Periodic Labs was founded by Ekin Dogus Cubuk and Liam Fedus. Cubuk led the materials and chemistry team at Google Brain and DeepMind, where one of his projects was, for instance, an AI tool called GNoME. That tool discovered over 2 million new crystals in 2023, materials that could one day be used to power new generations of technology, researchers say.

The goal of Periodic Labs is nothing less than to automate scientific discovery, creating AI scientists, the company says. This means building labs where robots conduct physical experiments, collect data, iterate, and try again, learning and improving as they go. The lab’s first goal is to invent new superconductors that it hopes perform better and possibly require less energy than existing superconducting materials. But the well-funded startup also hopes to find other new materials. Another goal is to collect all the physical world data that its AI scientists produce as they mix and heat and otherwise manipulate various powers and raw materials in their search for something new.

🇸🇪🚛 Einride has Raised $100 Million to Accelerate Autonomous Freight and Expand Global Operations

Einride, a technology company that provides digital, electric, and autonomous solutions for road freight, has raised about $100 million in funding from a mix of existing and new investors. The capital will power Einride’s next phase of growth as it scales the deployment of its autonomous freight solutions, deepens technology development, and continues its expansion with customers. Investors include EQT Ventures, one of Einride’s largest shareholders; a global asset management company based on the West Coast of the United States; as well as other investors and shareholders.

Collaborative Fund Leads Phaidra’s $50M+ Series B to Build the AI Factories of the Future

Phaidra, the company building AI agents for AI factories, announced today a Series B funding round of more than $50 million led by Collaborative Fund, with participation from Helena, Index Ventures, NVIDIA, Sony Innovation Fund, and others. The funding will be used to make AI factories — the data center infrastructure purpose-built for AI workloads — radically more resource-efficient via AI agents.

Today, the company collaborates with industry leaders like NVIDIA to design the next generation of AI infrastructure. Phaidra is developing a portfolio of AI agents — from liquid cooling to chiller plants to workload management — that are trained to orchestrate complex systems that exceed the capability of human intuition or hard-coded controls logic, thereby unlocking efficiency gains that were previously unobtainable. With the new capital, Phaidra will accelerate development of its agent ecosystem, deepen its collaboration with NVIDIA, and expand to additional global operators.

🇺🇸🤖 InOrbit.AI Secures Series A Funding to Scale Robot Orchestration Platform, Unlocking a New Era of Software-Driven Physical Workflows

InOrbit.AI, the leading provider of robot orchestration software, today announced the successful closure of its Series A funding round, co-led by L’ATTITUDE Ventures and Globant Ventures, the Corporate Venture Capital Fund of Globant (NYSE: GLOB). The capital infusion will accelerate the development of InOrbit’s innovative robotics software platform and strategically expand its market presence across key industries, including manufacturing, logistics, retail and hospitality.

🇫🇷 TiHive lands €8M to fuel industrial quality control with Terahertz-AI

TiHive, a French industrial deeptech company, has secured €8 million in funding from Karista, Wind, and the EIC Fund to accelerate its international expansion.

Founded in Grenoble in 2017 by Hani Sherry and Carlos Prada, TiHive designs industrial inspection solutions based on proprietary terahertz-on- silicon imaging and artificial intelligence. The company develops real-time, non-destructive inspection systems that bring “see-through” vision to production lines. By combining proprietary terahertz chips with advanced optics and software, it enables manufacturers to inspect every product, at full speed, without interrupting operations.

🇺🇸 Root Access develops tool for engineers of embedded systems, raises funding

New York-based Root Access has built an “AI-native” tool for embedded systems engineers. The company‘s goal is to enable them to modify heavy machinery, robotics, and mission-critical hardware with less time and effort. They announced $2.1 million in pre-seed funding.

Root Access incorporated in 2024 and had more than 100 engineers sign up for its open beta, which it had to close. “We had some really great firmware and embedded engineers from Mitsubishi, HP, Lockheed Martin, and John Deere,” said Eppley. “They were all signing up and saying, ‘Hey, I feel I feel left behind because I see all these AI developer tools, but they’re not for me.’ Everybody knows what a hardware engineer is and what a software engineer is.”

Root Access has signed a strategic partnership with Altium in which its API (application programming interface) pulls schematics directly from Altium into its IDE. It can ingest millions of datasheets from Octopart, which is now under Renesas.