Industrial AI Makes Good Time-Series

This week: Newlab's impact, seeking 'Deep Research' for cheap, mpGEMM for Edge LLMs, human-in-the-loop rapid commissioning, large format hybrid manufacturing platform, paper pallet wraps and straps.

Feb 09, 2025

∙ Paid

Shop Talk

Capturing this week's zeitgeist

Newlab represents a modern reimagining of the great industrial research laboratories, but with a key twist: rather than housing researchers under a single corporate umbrella like Edison's Menlo Park or Bell Labs, it creates an ecosystem where multiple startups can tackle complex physical and technological challenges. Like its predecessors, it focuses on tangible technology development, hosting everything from autonomous vehicle testing to rocket engine trials, but it democratizes access to these resources across hundreds of companies rather than centralizing them within one organization. This approach maintains the emphasis on physical innovation that characterized the golden age of industrial labs while distributing the risk and rewards across a broader network of entrepreneurs.

Many Exponential Industry readers are members of Newlab, so spread the word and you may find a new customer or partner within the community!

Last week's 'Custom Quote' revealed Boom Supersonic’s high-risk high-reward product development strategy. This week we answer ‘will humanoid robots follow AI's rapid growth trajectory?’ Upgrade today ⬇️

Assembly Line

This week's most influential Industry 4.0 media.

DeepSeek R1’s implications: Winners and losers in the generative AI value chain

✍️ Author: Philipp Wegner

The rise of models like DeepSeek R1 signals a potential shift in the generative AI value chain, challenging existing market dynamics and reshaping expectations for profitability and competitive advantage. If more models with similar capabilities emerge, certain players may benefit while others face increasing pressure.

IoT Analytics assesses the key winners and likely losers based on the innovations introduced by DeepSeek R1 and the broader trend toward open, cost-efficient models. This assessment considers the potential long-term impact of such models on the value chain rather than the immediate effects of R1 alone.

Timing is Everything: How SiTime Drives Innovation

👔🗣️ Executive visionaries:

TSMC Founder Morris Chang [Acquired]
A Bootstrapper’s Guide to Re-Industrializing America with Jim Belosic, CEO of SendCutSend [Manufacturing Happy Hour]; Marketplace Beta is Open!
This Is Palantir's CTO Shyam Sankar Vision for Changing How the US Does Defense Spending [Bloomberg]

π0 and π0-FAST: Vision-Language-Action Models for General Robot Control

✍️ Authors: Dana Aubakirova, Pablo Montalvo, Mustafa Shukor, Remi Cadene

We have ported the first robotics foundation models to Hugging Face LeRobot! Both π0 and π0-FAST, developed by Physical Intelligence, are now available in the LeRobot repository, bringing generalist robotic intelligence to the Hugging Face ecosystem.

While large language and vision-language models (LLMs, VLMs) show promise, they lack interaction with the physical world. To bridge this gap, we need models trained on robotic data. Generalist robot models can enhance adaptability, using diverse data to improve generalization and robustness. Instead of training on isolated tasks, pre-training on varied robotic data—similar to LLMs—boosts efficiency and performance.

IoT Renewable Energy Models: Building the Future With Time-Series Data

✍️ Author: Anya Sage

If you’re building monitoring solutions for solar farms, wind turbines, or grid management systems, understanding renewable energy business models—and their foundation in an IIoT-optimized time-series and real-time analytics database—will help you architect better solutions. This article briefly introduces some of these energy models and provides Timescale query examples, simply for demo purposes, to demonstrate the role of time-series data in enabling these models.

dsPRO, a cutting-edge automatic vision machine designed for fasteners

Automated human-in-the-loop commissioning of industrial processes: a glue dispensing case study

✍️ Authors: Jeroen Taets, Jeroen Jordens, Tom Lefebvre

This paper addresses the challenge of tuning process parameters during the commissioning of industrial processes. The process industry is hindered by unique challenges, including the absence of physical models and weakly specified bounds on process parameters. Moreover, multiple experiments can be performed simultaneously in a batch, but only under some extra constraints, such as a constant temperature for every experiment in the batch. These observations motivate the need for practical commissioning tools that can assist human machine operators or technicians. While there are solutions available for automatic tuning, they often lack the simplicity required for an operator-in-the-loop to easily understand and fine-tune the automatic tuning without being a machine learning expert. Therefore, we propose a simple bayesian optimization (BO) method with a single hyper-parameter and a systematic approach that can be easily understood. The effectiveness of the algorithm is demonstrated for glue dispensing, a complex industrial process with diverse properties that make tuning cumbersome. We validated our method in three cases, reducing the commissioning time by an inexperienced human-in-the-loop from days to hours.

🎬 Actioning human-in-the-loop systems:

A novel multi-agent reinforcement learning framework for robust exception handling of manufacturing service collaboration based on asymmetric information [Journal of Manufacturing Systems]
Cloud-edge-end collaborative multi-process dynamic optimization for energy-efficient aluminum casting [Journal of Manufacturing Systems]

New Product Introduction

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

UK first as Exeter’s Rapid Fusion set to showcase large format hybrid manufacturing platform

Rapid Fusion, which has enjoyed significant success with its Apollo and Zeus additive manufacturing cells, will showcase ‘Medusa’ at a special open day on 26 February at its Skypark R&D centre in Exeter.

The company has tapped into a £1.2m grant from Innovate UK to develop, test, design and build the industrial 3D printer, which is three times faster than conventional machines, twice as accurate and promises to reduce training and maintenance costs by 30%. The latest AI technology has been integrated into the design, with Medusa boasting a 1.2m3 volume build and able to move at 1200mm per second speeds.

Backed by project partners Rolls-Royce, AI Build and the National Manufacturing Institute Scotland (NMIS), the gantry-style machine combines pallet extruder, filament and CNC machining tool to provide a single-source solution for large moulds and tooling that is typically used by aerospace, automotive, marine and construction companies.

Smurfit Westrock’s paper pallet wrap set to drive down supply chain emissions

Smurfit Westrock has produced a 100% paper-based alternative to the polyethylene stretch wrap which is commonly used to stabilise pallet loads in transportation. The innovative new solution is made from Nertop® Stretch Kraft paper which is fully recyclable. Its capabilities also include high energy absorption and strong performance in humid conditions. The pallet wrap is the latest addition to Smurfit Westrock’s Better Planet Packaging Portfolio which seeks to reduce the environmental impact of packaging.

Encirc, a leading glass supplier which manufactures over three billion containers annually, was the first company in the UK to use the new paper wrap.

Business Transactions

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

Hitachi establishes its fourth Corporate Venture Capital Fund at a size of 400 million USD, which is the largest CVC fund managed by Hitachi to date. The fund will make strategic investments in startups with cutting-edge, digital technologies such as data centers, distributed energy systems, future of work and industrial AI, as well as startups in new frontiers including bio, quantum, nuclear fusion, life science, space and adjacent technologies

Stratasys Announces $120 Million Equity Investment from Fortissimo Capital

Stratasys Ltd. (Nasdaq: SSYS), leader in polymer 3D printing solutions, announced that Fortissimo Capital, a leading Israeli private equity fund investing in technology and industrials, has entered into an agreement to invest $120 million in the Company, acquiring approximately 14% of Stratasys’ issued and outstanding ordinary shares through a direct purchase of 11,650,485 newly issued ordinary shares at $10.30 per share, reflecting a premium of 10.6% over the closing market price on January 31, 2025. Prior to this transaction, Fortissimo held approximately 1.5% of Stratasys’ issued and outstanding ordinary shares. With this transaction, Fortissimo will hold approximately 15.5% of Stratasys’ issued and outstanding ordinary shares. The terms of the agreement also include an 18-month lock-up, as well as customary standstill provisions, subject to certain caveats specified below.

Trace.Space raises $4M to fix supply chain bottleneck holding back EV innovation

Trace.Space, a next-generation Latvian platform designed to tackle the very challenges that are slowing down the automotive industry. The company has just announced a $4 million seed round, led Cherry Ventures, with Outlast Fund and existing investors Nebular, Fiedler Capital, and Change Ventures. This marks Cherry Ventures’ first investment in the Baltics. This funding will fuel engineering, product development and go-to-market efforts as the company continues to attract the best engineers in Europe.

Trace.Space is leveraging modern cloud computing and AI models to enhance collaboration and minimise response times in product development. The company aims to revolutionise requirements management for engineering teams. The platform integrates advanced AI, including GPT algorithms, to assist engineers in documenting and revising requirements up to ten times faster. This acceleration enables the development of higher-quality products with reduced time-to-market.

AMETEK Acquires Kern Microtechnik

AMETEK, Inc. (NYSE: AME) announced the acquisition of Kern Microtechnik, a leading manufacturer of high-precision machining and optical inspection solutions. Kern specializes in the design and manufacture of ultra-precision manufacturing solutions capable of achieving sub-micron level accuracy. Kern’s advanced product portfolio includes high-precision machining solutions and optical tool inspection systems. Their highly engineered solutions support a wide range of applications that demand exceptional precision, including within the medical, semiconductor, research, and space markets.