Read Between the Machines

This week: verbal flubs, natural humanoid walk, podcast in a clean room, speciality chemicals renaissance, CNC AI agents, wiring harness automation, FLNG breakthrough, and 3d printed train station 🚅

Mar 30, 2025

∙ Paid

Shop Talk

Capturing this week's zeitgeist

In manufacturing environments where precision is paramount, our words don't always follow suit. Just as Ford executive Mike O'Brien meticulously documented his colleagues' linguistic missteps over decades, manufacturing floors worldwide are hotbeds for verbal gymnastics and colorful language.

This newsletter is no different, with intentional polysemous and antithesis titles. A few of my recent favorites:

Supersonic Fiber Placement - simultaneously celebrating supersonic flight and the manufacturing process to make the aircraft.
Mixed Screw Assembly Floor Plan – ambiguously describing mixed-model lines, a bed for the CEO on the factory floor, and a $45 screw that caught outrage.
Production Line Speed is Concrete - noting the irony of lines moving at the pace of setting cement while waiting on concrete to build AI factories.

Like O'Brien's legendary collection of verbal flubs, these titles play with manufacturing terminology in ways that make the technical delightfully accessible. While his spreadsheet documented linguistic accidents, our wordplay is deliberate – turning industry jargon into an art form that rewards the attentive reader with an unexpected smile. I hope you enjoy the puns as much as I do writing them! After all, it’s an industry with factory shamans who know how to ‘read between the machines’ to keep the factory humming.

For this year’s flash meetup at Hannover Messe, we will meet at the Crosser booth, Hall 17, Booth 15/1, right next to Microsoft on Tuesday, April 1, 2025 at 9:30 CEST 🥳

Add to Google Calendar

Add to Outlook

Add to Outlook mobile

Assembly Line

This week's most influential Industry 4.0 media.

Natural Humanoid Walk Using Reinforcement Learning

In our simulator, thousands of Figure 02 robots are simulated in parallel, each with unique physical parameters. These robots are then exposed to a wide range of scenarios they might encounter, and a single neural network policy learns to operate them all. This includes encountering various terrains, changes in actuator dynamics, and responses to trips, slips, and shoves.

To bridge the “sim-to-real gap” we use a combination of domain randomization in simulation and a kHz-rate torque feedback control on the robot. Domain randomization bridges the sim-to-real gap by randomizing the physical properties of each robot, simulating a breadth of systems the policy may have to run on. This helps the policy to generalize zero-shot to a physical robot without any additional fine-tuning.

We additionally run the policy output through kHz-rate closed-loop torque control to compensate for errors in actuator modeling. The policy is robust to robot-to-robot variations, changes in surface friction and external pushes, producing repeatable human-like walking across the entire fleet of Figure 02 robots. This is highly encouraging, as it indicates our technology can scale effectively across the entire fleet, without any additional engineering effort, supporting broader commercial operations.

We Made This Podcast Inside MIT's Massive Nanotech Clean Room

The Future of Specialty Chemicals

This new era will be defined by how well manufacturers adapt to evolving customer needs. For decades, success was measured by output and price. But today’s customers need reliability, responsiveness, and sustainability—without sacrificing performance or cost. Modern chemical production should require fewer inputs, generates less waste, and lowers costs—not just because of regulatory pressure, but because it makes business sense.

Sandvik: CNC AI Agents

Sandvik launches Artificial Intelligence in CAM software together with Microsoft /Sandvik/
CAM as a Service with dynamic toolpath generation ability for process optimization in STEP-NC compliant CNC machining /Journal of Manufacturing Systems/
Where is AI Being Used to Improve AMRs? /EETimes/
MakinaRocks Signs MOU with Siemens DI to Drive Edge AI Innovation in Manufacturing /MakinaRocks/

Chat with MES: LLM-driven user interface for manipulating garment manufacturing system through natural language

✍️ Author: Zhaolin Yuan, Ming Li, Chang Liu, Fangyuan Han, Haolun Huang, Hong-Ning Dai

This paper presents Chat with MES (CWM), an AI agent system, which integrates LLMs into the Manufacturing Execution System (MES), serving as the “ears, mouth, and the brain”. This system promotes a paradigm shift in MES interactions from Graphical User Interface (GUI) to natural language interface”, offering a more natural and efficient way for workers to manipulate the manufacturing system. Compared with the traditional GUI, both the maintenance costs for developers and the learning costs and the complexity of use for workers are significantly reduced. This paper also contributes two technical improvements to address the challenges of using LLM-Agent in serious manufacturing scenarios. The first one is Request Rewriting, designed to rephrase or automatically follow up on non-standardized and ambiguous requests from users. The second innovation is the Multi-Step Dynamic Operations Generation, which is a pre-execution planning technique similar to Chain-of-Thought (COT), used to enhance the success rate of handling complex tasks involving multiple operations. A case study conducted on a simulated garment MES with 55 manually designed requests demonstrates the high execution accuracy of CWM (80%) and the improvement achieved through query rewriting (9.1%) and Multi-Step Dynamic operations generation (18.2%). The source code of CWM, along with the simulated MES and benchmark requests, is publicly accessible.

Read more at Journal of Manufacturing Systems

New Product Introduction

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

Q5D to debut the world’s first wiring harness automation robot to take 200 USD off the cost of a car

The SQ25W is designed for automating manufacture of larger wire harnesses and with a build volume of 2.5m x 1.5m x 0.15m. This new model is large enough to accommodate a complete vehicle headliner and assemble the wiring directly on to it. Headliners face severe space constraints in accommodating wiring for internal lighting, infotainment, noise suppression technologies, microphones, and multiple sensors, including a growing number of cameras and safety devices. The robot cell adds insulated wires directly to product assemblies, securing them in place and terminating the wires to connectors. The completed assembly is electrically a drop-in replacement for existing designs. The integration of wiring onto the component saves wire assembly labour, and eliminates the need for clips, tie wraps, and NVH (noise and vibration) related components.

Samsung Heavy makes breakthrough in FLNG liquefaction technology

✍️ Author: Hyung-Kyu Kim, Woo-Sub Kim

Samsung Heavy Industries Co., a major Korean shipbuilder, has made a breakthrough in its liquefaction technology for floating liquefied natural gas (FLNG) production facilities, challenging long-standing dominance by US rivals. FLNG construction usually costs between 2 trillion won and 4 trillion won ($1.5 billion-$3 billion) per unit, with liquefaction equipment accounting for as much as 35% of that figure.

The technology, known as SENSE IV, is often referred to as the “heart” of the FLNG unit, a massive offshore structure that extracts, liquefies and transfers natural gas from subsea fields. SENSE IV can liquefy up to 2 million tons of natural gas annually and reduce energy consumption by up to 14% per ton compared to existing liquefaction equipment.

Read more at The Korea Economic Daily

Litmus Targets the Energy Sector Through Partner Agreements With SLB /Newswire/

JR West and Serendix to 3D print Hatsushima train station in Japan

✍️ Author: Davide Sher

The station’s primary components will be fabricated using 3D printing technology implemented by Serendix. The Japanese 3D construction printing service provider owns several concrete 3D printers and has recently added a large system from COBOD. The printed sections will undergo reinforcement with steel bars and concrete filling before on-site assembly. Cranes will lift and connect the prefabricated parts, allowing for a rapid build process. The entire structural assembly is expected to take just six hours, from the night’s last train to the first train in the morning. This method significantly reduces on-site labor and construction time compared to traditional steel and concrete methods.

Business Transactions

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

Former Cruise CEO Vogt's robotics startup, The Bot Company, valued at $2 billion in new funding /Reuters/

🇨🇦 Nanoprecise Closes US$38M Series C

Nanoprecise Sci Corp is breaking the reactive cycle that has been a fixture of industrial maintenance for too long. On the heels of triple-digit growth in 2024, Nanoprecise is moving boldly to scale its world leading Energy-Centered Maintenance (ECM) approach and enhance its AI-driven diagnostics that deliver actionable explanations, rather than just alert notifications.

The oversubscribed equity financing led by Yaletown Partners and co-led by BDC's Industrial Innovation Venture Fund , with participation from Export Development Canada (EDC), BMO Capital Partners, alongside a Credit facility from CIBC Innovation Banking, provides Nanoprecise the resources to rapidly expand its market presence and accelerate its strategic initiatives.

Most predictive maintenance programs count saving downtime as the only major ROI, which is true for roughly 10 to 20% of production critical machinery. However, 80 to 90% of the rotating equipment in any process industry has some form of redundancy, so solely saving downtime does not capture full value. On the remaining 80% of a plant's equipment, Nanoprecise's trademarked ECM approach provides maintenance insights alongside energy consumption patterns which, if acted upon, provide energy savings that justify using ECM not just for saving downtime on critical machinery, but also on the "balance of plant" equipment.

🇺🇸 CADDi raises $38M Series C Extension led by Atomico to double down on rapid growth of their AI data platform for manufacturing companies

CADDi, the AI data platform for empowering manufacturing companies globally by transforming legacy drawings and supply chain data into fuel for digital transformation, today announces the closing of a $38M Series C Extension led by Atomico, a leading European growth investment firm. The round saw further doubling down of existing growth investors like Global Brain and Minerva Growth.

CADDi's AI data platform is on a mission to democratise supply chain data, starting with technical drawings; the company’s software platform turbocharges the legacy tech stack by enriching product level data and structuring it such that it can be used for making better decisions. The foundation of this manufacturing intelligence layer is an AI-enabled search engine that tackles a plethora of use cases. One example is to optimise parts pricing through a similarity search of historical technical drawings to find pricing inconsistencies. For an average customer, one project typically involves over 10,000 individual technical drawings. Similarity search means employees save time and the company saves money by optimising negotiating positions in the procurement process, as well as ability to find alternative suppliers on a part level linked to the actual technical drawings.

🇺🇸 Arcade, First-Ever AI Physical Product Marketplace, Raises $25M Series A and Expands Into Home Goods

Arcade, a generative AI marketplace where what can be dreamed can be made, announced a $25 million Series A fundraise, bringing its total fundraise to $42 million. The company also announced its launch of home goods, marking a major expansion from its beta launch in jewelry a few months ago. Arcade is at the forefront of a major AI-driven shift in commerce, where with just a few words or an image, consumers can instantly receive unique AI-generated design suggestions or design something themselves.

The funding round and category expansion come just months after Arcade's beta launch in September 2024. The platform has seen rapid adoption among consumers and makers, and in just three months, users on Arcade have created the largest jewelry assortment in the world - with 650,000 jewelry designs created.

🇺🇸 Contoro Robotics Secures $12M Series A to Bring AI-Powered Trailer Unloading to Scale

Warehouse automation is hitting a major inflection point, and Contoro Robotics is at the forefront. The company announced a $12M Series A to scale its AI-powered trailer and container unloading robots—solving one of the most grueling and labor-intensive challenges in logistics. The round brings in new investors Doosan, Coupang, Amazon Industrial Innovation Fund, and IMM, alongside continued backing from SV Investment, KB Investment, Kakao Ventures, and Future Play, bringing Contoro’s total funding to $22M.