The Wealthiest is the Industrialist

This week: vertical integration, Musk pay package, start-up to scale-up, developing automation products from scratch, MagNEO, train-free machine vision, speech-to-reality systems, 3D chips fabricated

Dec 21, 2025

∙ Paid

Shop Talk

Capturing this week’s zeitgeist

The path to extreme financial returns in modern industrial ecosystems is the by-product of vertical integration of the entire stack, a strategy pioneered 150 years ago by JD Rockefeller’s Standard Oil. By starting midstream and imposing rigorous standardization on product quality and delivery metrics, Rockefeller eliminated the efficiency-killing chaos of middlemen and connected disparate business units into a seamless, dominant engine. Today’s challenger and behemoth industrial firms, like BYD, Tesla, SpaceX, Wanhua Chemical, Huawei, Reliance Industries, and more are replicating this blueprint by using standardized digital and physical protocols to unlock stranded value across the chain, effectively owning the market by defining its operational language. However, when different conditions are true the winning strategy may be different.

Get 50% off for 1 year

Assembly Line

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

🕴️ Start-up to Scale-up: Our Path to Maturity in the Rail Industry

KONUX achieved success in the highly regulated rail industry by maintaining a singular focus on digitalization through a blend of IIoT hardware and physics-informed AI. By deploying ruggedized sensors like the KONUX Switch and the CoBrix data platform, they introduced a SaaS-driven predictive maintenance model to an industry historically dominated by hardware incumbents. This strategy shifted rail operations from reactive, time-based maintenance to proactive, condition-based actions, backed by a multidisciplinary team that combined deep data science with trackside engineering expertise.

This specialized approach enabled KONUX to break the pilot loop and become the first startup to win a major public tender with Deutsche Bahn. As of 2025, the company has scaled into a global partner for major operators like Network Rail and ADIF, recording over 500 million train traces. Their success has resulted in measurable outcomes for the industry, including enhanced safety through early failure detection, significant maintenance cost savings, and increased network capacity across international rail corridors.

🧠🦿 Scaling Is Harder Than Building The First Robot

Stefan Dörr-Laukien⁠ is the co-founder and CEO of ⁠NODE Robotics⁠, a Stuttgart-based company building the software layer behind scalable mobile robot fleets. In this episode, Stefan shares how his path into robotics did not start with founding ambitions, but with curiosity for how machines work. From studying mechanical engineering at TUM to hands-on autonomous driving research at BMW and years at Fraunhofer IPA, his motivation was always the same: building systems that work in the real world, not just on paper.

Self-driving Unmanned Forklifts go to the Warehouse /Maeil Business Newspaper/ now more unmanned robot and vision inspection facilities than the number of workers
How to Calculate the True ROI of AMRs in a Fulfillment Center /Locus Robotics/

🖨️🧲 MagNEO: How 3D Printing is Helping Europe Build an Independent Magnet Supply Chain

The MagNEO project is establishing a localized European magnet supply chain by integrating advanced additive manufacturing (AM) with rare-earth-free material science to eliminate dependence on external imports. Utilizing Laser Powder Bed Fusion (LPBF) and Directed Energy Deposition (DED), the initiative enables the production of high-performance AlNiCo and High-Entropy Alloy (HEA) magnets with complex geometries and precise grain alignment that traditional casting cannot achieve. These manufacturing processes are accelerated by machine-learning-assisted modeling for rapid alloy discovery and high-throughput synthesis to validate magnetic properties. Furthermore, the supply chain is secured through circular economy technologies, including the selective recovery of nickel, cobalt, and copper from end-of-life products and the recycling of AM residues into new powders, creating a resilient and sustainable domestic production ecosystem.

📸⚙️ Why your next visual quality inspection system will be training-free

Machine vision is undergoing a radical simplification as the industry shifts from discriminative AI models to training-free, generative anomaly detection. Unlike traditional systems that require thousands of labeled images to learn specific flaws, these modern “normality-based” models focus exclusively on the concept of a perfect product, requiring as few as 25 to 50 defect-free images to establish a baseline. By using generative algorithms to model the ideal production state, the software can automatically identify deviations as defects—even those never previously encountered—without manual data labeling or extensive training cycles. This transition effectively moves setup times from weeks to minutes and creates a hardware-agnostic environment that eliminates the need for highly specialized, static lighting or positioning, making advanced quality inspection a “plug-and-play” reality for factory engineers.

🧠🦾 An innovative robotic arc welding system that increases productivity and reduces costs with AI

To meet the client’s needs, X-Machine developed a customized, turnkey welding system integrated with Comau’s S-13 robot enhanced with Artificial Intelligence. The S-13 robot from Comau’s S-Family range was chosen for its high performance and for its compact structure. It has a payload of 13 kg, a maximum reach of 1960 mm and an extremely small footprint (33×33 cm), which make it suitable for operating easily in small and narrow spaces. Thanks to these characteristics, it also guarantees excellent cycle times, great repeatability and high precision of path control. Finally, its hollow wrist configuration, which facilitates and protects the passage of cables while avoiding additional external clutter, is ideal for applications in critical environments, such as arc welding in the steel industry.

In particular, the integration with Artificial Intelligence features has allowed the Comau S-13 robot to achieve even higher levels of productivity and diagnostics. Much more than a simple work unit, X-Machine has designed a complete system automation solution that, through advanced software and the use of AI, also provides the customer with a Digital Twin of the plant, allowing Tuboside to track and optimize the entire production process in real time. Furthermore, by collecting, interpreting and consolidating millions of essential data points, the system helps the customer make effective and concrete operational decisions.

💬🦾🪑 MIT researchers “speak objects into existence” using AI and robotics

Generative AI and robotics are moving us ever closer to the day when we can ask for an object and have it created within a few minutes. In fact, MIT researchers have developed a speech-to-reality system, an AI-driven workflow that allows them to provide input to a robotic arm and “speak objects into existence,” creating things like furniture in as little as five minutes.

With the speech-to-reality system, a robotic arm mounted on a table is able to receive spoken input from a human, such as “I want a simple stool,” and then construct the objects out of modular components. To date, the researchers have used the system to create stools, shelves, chairs, a small table, and even decorative items such as a dog statue.

New Product Introduction

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

💿 First truly 3D chip fabbed at US foundry, features carbon nanotube transistors and RAM on a single die

A Stanford-led research team has achieved a significant breakthrough in semiconductor design by developing a 3D AI chip that delivers a fourfold performance-to-energy gain over traditional 2D architectures. Fabricated at SkyWater Technology, a commercial U.S. foundry, the chip utilizes Monolithic 3D integration to stack logic and memory components, specifically carbon nanotube transistors and resistive RAM (RRAM) vertically on a single die. This dense integration overcomes the memory wall bottleneck by allowing for significantly faster and more efficient data transfer between processing and storage layers, proving that next-generation, high-performance AI hardware can be successfully manufactured using existing domestic foundry processes.

🖨️ Open-sourcing CORE One CAD Files Under the New Open Community License (OCL)

Prusa Research has released the full CAD files (STEP and Fusion 360 assemblies) for their new CORE One and CORE One L 3D printers. This release introduces the Open Community License (OCL), a new licensing framework designed specifically for physical hardware to bridge the gap between traditional software licenses and the realities of commercial manufacturing.

What is Permitted: Users are free to download, modify, and share mods. They can also use the designs for internal business purposes, such as building their own print farms or producing spare parts (Right-to-Repair). The license aims to protect creators from patent trolls by establishing a legal contract that grants rights to the community while making it harder for bad actors to claim ownership of open designs.

Read more at Prusa and download on Printables

🧠🔁 GEN-0: Embodied Foundation Models That Scale with Physical Interaction

Generalist AI has introduced GEN-0, a pioneering class of embodied foundation models designed to bridge the gap between digital reasoning and physical interaction through a native multimodal architecture. A central breakthrough is the development of Harmonic Reasoning, a novel training approach that enables robots to think and act simultaneously by processing asynchronous, continuous-time streams of sensing and action tokens, bypassing the need for separate reasoning and control architectures. The research identifies the first clear scaling laws for robotics, demonstrating that performance predictably improves with increased data and compute following a power-law relationship similar to large language models. Furthermore, the team observed a critical intelligence threshold around 7 billion parameters; at this scale, models undergo a phase transition that allows them to continuously absorb and generalize from a massive dataset of over 270,000 hours of real-world manipulation data. This cross-embodied framework allows high-fidelity reflexes and physical commonsense to scale across various robotic forms, moving the field toward general-purpose autonomy that learns directly from the chaotic reality of the physical world.

Business Transactions

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

ARM Institute Funds New Robotics & AI Projects to Advance and Grow US Manufacturing Innovation /ARM Institute/ in special topic areas: Multi-Modal Inputs for AI Robotics in Manufacturing, Rapid Re-Tasking and Robot Agility, Multi-Robot, Multi-Human Collaboration, and Adaptive Real-Time Path Planning and Control
Small Business Administration Launches First-Ever Loan Program Dedicated to American Manufacturers /sba.gov/
Over two dozen startup accelerators you can apply to right now! /Rexan Wong on X/
A Private Equity Fund Just Bought Your Company … Now What? /Bain/
Roomba Maker iRobot Files for Bankruptcy and Will Go Private /Bloomberg/

🇬🇧🖨️ Aibuild Secures Strategic Investment from Driventure and 212 NexT to Accelerate ‘Physical AI’ in Manufacturing

Aibuild has secured a strategic investment round led by existing investor IQ Capital, with new participation from Driventure (Ford Otosan’s venture arm) and 212 NexT, alongside continued support from ACT Venture Partners, Force Over Mass Capital, and Atlas Ventures. While the specific funding amount was not disclosed, the capital is earmarked to accelerate the industrialization of “Physical AI” and scale the company’s AiSync platform, a hardware-agnostic software that automates complex toolpath generation for large-scale robotic 3D printing. The company plans to use the funds to expand its U.S. operations via its Silicon Valley headquarters, specifically targeting the aerospace, defense, and energy sectors with secure, air-gapped “dark factory” capabilities. Additionally, the investment will fuel technical advancements in fast physics simulations for defect prediction and the integration of hybrid manufacturing workflows that combine additive and CNC milling processes.

🇺🇸🏭🧠 Axion Raises $37M Series B to Power the Future of American Manufacturing

Axion, the pioneering AI platform for manufacturing product quality, announced a $37 million Series B funding round led by Salesforce Ventures with participation from Bessemer Venture Partners, Schneider Electric Ventures, and other existing investors. Axion’s platform, combining domain-specialized AI algorithms with human-in-the-loop problem solving workflows, serves as the operating layer underpinning America’s manufacturing revival.

Teams use Axion’s proprietary intelligence platform to ensure superior product quality by detecting the earliest warning signals of emerging issues that could impact their customers. By addressing problems swiftly, manufacturers improve uptime, enhance customer experience, and significantly reduce warranty costs. Importantly, Axion shifts quality from reactive to proactive, closing customer feedback loops and unifying data sets to enable manufacturers to ideate and develop the next generation of products, driving innovation across their entire product suite.

🇺🇸🧠📐 Navier Raises $5.6M to Deploy Agent-Driven Engineering, the Next Engineering Unlock After CAD and Simulation

Navier, which builds autonomous engineering teams for hardware design and development, announced $5.6 million in seed funding and its launch from stealth. The round includes participation from GV (Google Ventures), HCVC, and Y Combinator.

Engineering has undergone two major productivity shifts: Computer-Aided Design (CAD) replaced manual drafting in the 1960s, then simulation software enabled virtual testing in the 1990s. Each shift reduced overhead requirements while expanding what engineers and designers could accomplish. Navier introduces the third shift: Agent-Driven Engineering (ADE), in which AI agents automate repetitive workflows between design and engineering disciplines, which still consume a significant portion of an engineer’s time.