Digid sensors are believed to be the world’s smallest – and are set to become even smaller: Digid’s technology roadmap forecasts future production of sensors that are just 10nm long.

This nanoscale sensing technology provides the key that will unlock the potential of multiple emerging markets, including physical AI and humanoid robots. In robotics, for instance, Moravec’s paradox – that robots struggle with tasks which humans find easy, such as handling a delicate wine glass – has persisted in part because robot hands lack the billions of nerve endings on a human finger.

Now Digid offers the opportunity for machines to mimic human sensory capabilities: its nanoscale printed sensors can be applied to surfaces such as a robot’s shell in arrays of up to 16 x 16 sensors.

Other applications for Digid sensors include: 

• Force sensing on the blade of a scalpel in robotic surgical equipment
• Temperature sensing on the tip of a temperature probe used in minimally invasive surgery
• Temperature sensors embedded inside battery cells, for safety and performance monitoring
• Biosensors for detecting biological objects such as viruses, or the chemical markers of drugs in the bloodstream.

Dr Konstantin Kloppstech, chief technology officer of Digid, said: "With the start of mass production of Digid sensors, the opportunities to embed sensing on almost any surface or in almost any device have become limitless. Our sensor is so small that it cannot be seen with the naked eye. Now it is up to the imaginations of design engineers to dream of uses for sensing where sensing has never before been possible."

For each design project, Digid supplies a custom sensor and sensor assembly backed by hardware and software integration support. The sensor provides either a voltage or resistance measurement output via an I2C interface. Digid signal processing software converts the raw measurement outputs to useful temperature or force data.

As well as their nanoscale dimensions, Digid sensors also offer negligible self-heating or other distorting effects on sensor readings, and highly accurate and linear measurement outputs.

Minimal digital overhead is required – unlike for other types of sensor, Digid nanoscale sensors require no compensation for drift or other non-linearities

digid.com/contact

The collaboration brings together two complementary offerings with a shared mission to help customers in the United States, and around the globe, lower the cost, complexity, and time it takes to deploy robotics and industrial automation at scale. 

HowToRobot operates what is reported to be the world’s largest automation sourcing platform with more than 20,000 suppliers and expert support helping companies understand what robotics solutions will work for their organisation while also defining project requirements and streamlining deployment. This simplified approach allows customers to cut automation costs by up to 50% and launch robotics solutions four times faster. Holman recently launched a new robotics division, which combines solution design, flexible financing, and lifecycle asset management, enabling customers to launch faster, scale smarter, and maximise the long-term value of their investment with less complexity. 

Through this collaboration, HowToRobot and Holman are able to provide a seamless, end-to-end pathway for automation, making it easier for businesses to move from planning to deployment, manage their automation fleet, and scale solutions with lower risk and greater financial flexibility.

“We’re extremely proud of the investment Holman is making in our company,” said Søren Peters, CEO of HowToRobot. “There is a significant need in the market for making robotics and automation more accessible. The Holman Robotics team certainly shares this strategic vision, and together we look forward to helping customers embrace automation much faster while also eliminating the usual financial and technical roadblocks.”

The investment, which is effective immediately, supports HowToRobot’s continued development of tools and capabilities that make automation more accessible for customers. This includes further development of its AI-driven quotation tool and digital workflows that guide customers from defining their needs to matching with suppliers, obtaining and comparing budgetary and firm quotes, and ultimately selecting a vendor. The funding will also support the company’s transition toward a SaaS-based model and expanded commercial operations in key markets, such as the US and UK.

“For more than a century, Holman has been at the forefront of embracing innovative technologies that have the potential to deliver additional value to our customers,” said Holman’s VP of Robotics Joe Foster. “This investment and collaboration with HowToRobot will further strengthen our groundbreaking robotics offering, allowing Holman to leverage decades of experience managing complex assets to deliver a holistic approach to automation that bundles hardware, data integration, management software, and ongoing support into a single, streamlined solution.”

HowToRobot.com

The contract represents a significant commercial validation of alternative motor winding materials as OEMs seek to reduce exposure to copper supply chain vulnerabilities. With approximately 50% of global copper refining capacity concentrated in a single region, automotive manufacturers are actively exploring technology pathways that offer greater supply diversification.

The UK Government's Critical Minerals Strategy (November 2025) validates these concerns, projecting copper demand to increase from 922,200 tonnes in 2027 to 3,619,000 tonnes by 2035, nearly doubling over the period. The strategy identifies copper supply chain concentration as a strategic vulnerability and commits up to £50 million to support innovation that delivers supply chain diversification.

AEM's technology replaces conventional copper windings with compressed aluminium conductors, leveraging aluminium's established global supply infrastructure and reduced geopolitical concentration risk. The approach maintains power density performance while delivering operational efficiency gains through electrodynamic performance characteristics that contribute to reduced energy consumption per kilometre. 

James Widmer, CEO of AEM, said: "OEMs are evaluating motor technologies not just on performance metrics, but on strategic supply chain resilience. Our aluminium winding architecture addresses critical sourcing vulnerabilities while delivering measurable efficiency improvements and enhanced end-of-life recyclability."

The commercial engagement follows AEM's recent tier 1 supplier agreement for its rare-earth-free SSRD motor technology, indicating growing industry momentum toward materials diversification in electric drivetrain components.

AEM's compressed aluminium winding technology delivers several engineering benefits over conventional copper designs. The material availability advantage is significant: aluminium refining capacity is geographically distributed across multiple regions, reducing single-source supply chain exposure that has become a strategic concern for automotive manufacturers.

Manufacturing flexibility improves through lower material costs and established aluminium processing infrastructure, enabling more diverse supplier qualification. Operational efficiency gains are achieved through enhanced thermal performance characteristics that contribute to reduced energy consumption per kilometre.

End-of-life processing advantages include clean separation of aluminium and steel components, which enables higher-value material recovery without metallurgical contamination. The circular economy benefits are also substantial: recycled aluminium requires 95% less processing energy than primary production, supporting automotive manufacturers' scope 3 emissions reduction targets.

The automotive industry's increasing focus on supply chain resilience reflects multiple strategic concerns. Copper price volatility has introduced significant cost uncertainty into electric motor bill-of-materials, while geopolitical factors affecting mineral processing capacity create potential production continuity risks.

Alternative motor winding materials offer manufacturers strategic optionality in component sourcing, potentially enabling parallel supply chain development and reducing dependency on single-material architectures.

Widmer added: "UK engineering innovation is providing automotive manufacturers with technically validated alternatives to conventional motor designs. This partnership demonstrates that supply chain diversification and performance optimisation are not mutually exclusive objectives."

advancedelectricmachines.com

Nathan Charles, head of customer experience at managed IT and cyber security partner OryxAlign, explores why traditional defences are losing ground as attackers adopt AI-enabled methods, and how engineering and manufacturing organisations can adapt their security strategies to stay resilient.

Legacy tools under strain

Traditional tools built on signature updates or static rules were never designed to match the pace or instability of modern attacks. AI alters code constantly and reshapes its own signals in ways that unsettle tools which depend on stable, recognisable patterns.

Automated scripts test weak points at high frequency and generate rapid chains of intrusion attempts that leave teams sifting through alerts. This creates a landscape where familiar tools still have value yet struggle to provide the confidence that many organisations need when adversaries no longer work at human speed. In engineering and manufacturing environments, this pressure is often intensified by a mix of legacy systems, specialist equipment and networks that were never designed for constant exposure.

Recent analysis from security researchers reports “78 per cent of CISOs now admit AI-powered cyber-threats are having a significant impact on their organisation”, which reinforces the growing limitations of traditional controls.

Rising impact in Britain

Across the UK the impact is already visible. The Cyber Security Breaches Survey 2025 shows that organisations reporting a breach face a mean cost of £3,550 for their most disruptive incident, while a government-commissioned study places the wider economic impact of cyber attacks at around £14.7 billion each year. These figures show that routine incidents still carry weight for organisations across the UK. They also reveal a shift in how attacks unfold.

Automated probing shortens the gap between an initial scan and a serious attempt to breach a system, which forces incidents to gather pace and draws heavily on operational teams. As this tempo increases, older tools struggle to keep their footing and leave practitioners working with less room to anticipate the next stage of an intrusion. For manufacturing and engineering teams, this compression of response time can sit alongside production demands and safety considerations, adding further strain during active incidents.

AI reshapes monitoring

A further challenge appears once AI begins to influence how organisations monitor their environments. Automated tools now scan networks and endpoints for unusual activity, although their outputs often need human context before teams can trust what they see. These systems can present signals that sit close to normal operational patterns, which makes it harder for practitioners to judge whether a change in behaviour deserves closer attention. 

Attackers also use AI to produce misleading indicators that mimic trusted activity or disguise a malicious sequence inside ordinary network traffic, which makes early recognition far harder for automated systems. Without oversight, teams risk either ignoring subtle signals or chasing false leads that drain resources during busy periods.

Building stronger visibility

Security therefore rests on a blend of clear visibility and confident human judgement, supported by processes that help teams act without hesitation. UK engineering and manufacturing organisations benefit from monitoring that builds a steady picture of system behaviour under routine conditions.

Lifecycle planning also supports this picture by keeping endpoints current and reducing the presence of devices that sit outside managed oversight. These adjustments give teams a steadier view of network activity, even as automated tooling produces a heavy flow of alerts. With a clearer picture in front of them, practitioners can step into developing incidents earlier and guide responses with more confidence.

Sharper social threats

Another pressure on security teams comes from the steady rise in social-engineering attempts. Recent global research notes that in 2024 “there was a sharp increase in phishing and social engineering attacks” and that “Generative AI is augmenting cybercriminal capabilities”. These messages often pass through standard filtering and reach staff who may not expect them.

Automated tools can support the screening process, although their outputs need human review to avoid misjudging messages that share traits with legitimate correspondence. As these attempts grow more polished, organisations benefit from awareness training and monitoring practices that keep pace with the evolving character of these attacks.

AI-driven intrusion methods continue to advance, yet organisations can adjust their thinking to meet this change. Traditional tools still hold value, although their protective strength relies on how they sit alongside real-time monitoring and the routine maintenance that keeps systems predictable enough for practitioners to read them with confidence. For engineering-led organisations, maintaining this balance is increasingly part of operational resilience rather than a standalone IT concern.

Further information on approaches that support stronger cyber resilience is available at:

www.oryxalign.com/cyber

Founded in 2011, adi Lean Manufacturing & Design specialises in designing and building bespoke engineering solutions. This ranges from robot cells and full assembly lines to manual workstations and trackside equipment.

The first partnership with Schneider Electric builds on a 10-year relationship and formally recognises the division as a Registered Alliance Partner system integrator within Schneider Electric’s Global Alliance Partner Programme.

Sam Curson, BDM at adi Lean Manufacturing & Design, said: “Achieving partner accreditation reflects our team’s focus on delivering tailored automation systems that cut production costs and unlock new capabilities, as well as improve product quality for our customers.

“Working even more closely with Schneider Electric enhances our technical competency with Schneider products and allows us to meet our client’s needs with more advanced technical solutions.

Samira El Kassimi, robotics and automation business development manager at Schneider Electric, welcomed the strengthened collaboration. She said: “Through the Alliance System Integrator Partner Programme, we work with integrators who share our ambition to deliver more efficient and more sustainable automation. adi Lean Manufacturing & Design brings strong expertise in lean production and turnkey assembly lines.

“Adding them to our UK Alliance network means more manufacturers can access that capability, backed by Schneider Electric’s global technologies and support.”

adiltd.co.uk

Recent major venture capital (VC) reports from KPMG and PitchBook confirm that AI remains in the lead, accounting for more than half of all investments this year. Data from CB Insights shows that investors’ attention inside the AI market is shifting rapidly toward industrial humanoid robotics. As a result, investors argue the flood of AI capital is pushing robotics toward a speculative zone, with too many startups promising breakthroughs without commercial evidence.

Last quarter, industrial humanoid robotics captured 17 deals – the most of any category. AI was still the primary destination for investors, split into several categories, such as coding AI agents and copilots (14 deals), end-to-end software development AI agents (12), and others.

Rapid growth of the sector has already sparked fears of a bubble from the Chinese leading economic planning industry, which said that the humanoid robotics industry needs to “balance the speed against the risks of bubbles,” Bloomberg reported.

Investors' appetite for humanoid robots is largely driven by AI, because AI gives humanoids a commercial potential that was previously not possible.

According to Daiva Rakauskaitė, the partner and manager of Aneli Capital, a company that manages a €35 million fund for early-stage Central and Eastern European startups, there are strong similarities between today's AI-driven investment boom and the dotcom bubble in the early 2000s, leaving many startups exposed. She expects an AI bubble burst in 2-3 years.

“Many AI startups that can’t yet generate revenue will fail, but we’re reaching a consensus on that in the market. While the same risks persist in humanoid robotics, many investors tend to overlook this,” says Rakauskaitė. “However, it is important to distinguish robotics from humanoid robotics; industrial and logistics robots already generate revenue and can deliver measurable results, while humanoids can’t yet prove their commercial value.”

Currently, companies around the world demonstrate prototypes of robots performing actions from running to boxing, sparking interest from users and investors. However, in the real world, they have few practical commercial applications.

Similar challenges also persist for industrial humanoid robotics. These companies face challenges with inference (ability to make decisions in real time), dexterity (how well the robot can physically handle things), reliability, and cost, which limit the initial use cases to factories and warehouses with predictable sets of tasks, CB Insights report claims.

According to Rakauskaitė, especially now, when investments are driven by hype, VCs should not forget the fundamentals and prioritize revenue-first philosophy, where real money matters more than growth at all costs. 

“Investments in robotics and AI are crucial for the future development of humanity. But investors should remain disciplined and back companies that have realistic goals based on economics, not hype. From day one, startups should aim for early revenue streams through licensing, partnerships and have a clear model of monetization in the near future. The same revenue-first philosophy can be applied to any field,” Rakauskaitė says.

Despite early signs of a bubble in humanoid robotics, she remains confident in the broader robotics sector, where cheaper hardware and rapid advances in AI are accelerating real-world deployment.

According to Rakauskaitė, robotics is an especially promising field for the CEE startups. The region is located close to Germany, the largest industrial robotics market in Europe, which provides a major strategic advancement to scaling.

“The region also has lots of hidden talent. That’s why we dedicated our new fund for this region, aiming to support the talented founders with hands-on guidance and quick decision-making. Many hype-driven investors pull back once the hype fades. Yet to create real innovators, VCs must support them through their full journey. That’s exactly what we are going to do,” Rakauskaitė concludes.

anelicapital.com

Matta’s AI is reported to give factories the ability to see, understand, and improve themselves in real time, understanding any production line within days. It spots defects, traces root causes, and helps teams fix problems before they become costly.

The technology is highly adaptable, capable of working across everything from electronics and automotive to defence and apparel, whether on manual inspection stations, conveyor lines, or robot arms to redefine how products are conceived and created. This flexibility is driving strong demand, with 300+ factories in the pipeline and a new installation every two weeks.

Doug Brion, co-founder and CEO of Matta, said: “Everyone talks about the glamorous side of manufacturing: generative design, material discovery, digital twins, but few spend time on the factory floor. The hard part isn’t dreaming things up inside a computer; it’s making them work at scale. Manufacturing still runs on human know-how, the kind that lets someone on the line kick a machine just right, or run a finger over a scratch, and say, ‘that’s thirty-four microns wide.’ We’re using AI to capture and scale that tacit knowledge, so engineers can design things that actually work in the real world."

Manufacturing at an inflection point

Manufacturing underpins a third of global economic output yet remains plagued by inefficiencies that waste up to 20 percent of production value and raise emissions. After decades of deindustrialisation, factories are exposed to external geopolitical shocks and must do more with less. Matta provides a practical route to productivity, quality and resilience on today’s shop floor.

At the same time, energy costs are rising, supply chains are fragile, and workforces are ageing. Factories must reshore, decarbonise, and do more with fewer skilled hands. In the UK, vacancies already outnumber qualified engineers, and costs keep climbing. Across Europe and the US, the story is the same.

Matta develops AI that learns the physical rules of production and applies them on the line. Its first product uses unsupervised and self-supervised computer vision to automate quality control and anomaly detection, perform measurements, diagnose root causes, and recommend corrective actions in real time. A central platform lets teams monitor every camera, analyse results and trace parts across the factory for live visibility of issues and bottlenecks.

Matta delivers this as a full plug-and-play system combining hardware, factory integration, AI research, and software. Most deployments are live within hours, with cameras inspecting automatically after a short learning period.

In one polymer manufacturing deployment, Matta achieved over 99% defect-detection accuracy with just ten minutes of data. Recent projects range from inspecting high-speed bottling for defects with a global drinks brand to working with Bowers & Wilkins, where Matta’s AI rapidly measures speaker components to catch issues before assembly.

Beyond detection, Matta partners with OEMs to enable machines to tune themselves. One of these OEMs, Caracol, is integrating Matta’s vision AI for closed-loop control, linking real-time inspection to automatic parameter adjustments on industrial printers and large-format robot additive manufacturing cells.

World-class academic founders

Matta was founded on pioneering research from the University of Cambridge’s Institute for Manufacturing, where co-founders Douglas Brion, who completed a PhD in deep learning-enabled control, and Sebastian Pattinson, Associate Professor of Engineering, first met. Today, Matta is a fast-growing team with experience from MIT, Imperial, BBC R&D, Google X, and Microsoft.

The latest funding will accelerate customer adoption and AI development, expand self-serve deployment, and support Matta’s expansion into key manufacturing regions across Europe and the US, advancing the company’s vision for fully autonomous, end-to-end production.

www.matta.ai

SMEs are the backbone of the UK economy, comprising over 90% of the manufacturing sector. Despite their fundamental role in national economic strength, their experiences and needs are often underrepresented when policy is shaped. The Policy Centre aims to move the conversation about SMEs beyond anecdote, creating a credible, data-driven force to inform government thinking. It wil act as an independent convener for industry, policy organisations, sector collaborators, and government, whose purpose is to ensure that policy solutions address the skills gap and enable productivity and sustainable growth in engineering and manufacturing. 

The Policy Centre will amplify the voice of SMEs by continually gathering evidence and insight from sources such as the SME Advisory Council and the SME Snapshot survey, ensuring their priorities help shape policy. As an independent charity, the Centre will also act as a constructive ally to the Government, offering robust, data-driven intelligence on the sector’s skills challenges and opportunities. In addition, it will strengthen collaboration across the ecosystem by working proactively with membership organisations, trade bodies, sector networks, and large employers, recognising that a unified message carries far greater influence.

Early activity has already focused on key national priorities including: ensuring the Advanced Manufacturing Plan and Industrial Strategy fully reflect the contribution of SMEs; encouraging government to begin exploratory work on the benefits of a skills tax credit to boost employer investment in training; and emphasising the importance of balancing government's commitment to introducing greater flexibility to the Growth and Skills Levy without compromising SME access to levy funds for apprenticeships 

To support this ongoing evidence gathering, the Policy Centre runs the SME Snapshot, a bi-annual survey capturing how SMEs are responding to economic and policy developments.

Ann Watson, CEO of Enginuity, said: “SMEs are the lifeblood of the UK economy yet often fail to be heard by those making policy in key areas at the heart of government – and those honing policy need to listen. Effective government policy depends on meaningful engagement with the people and organisations whose insights and experience are essential to its success. 

“SMEs are huge in number, but that can mean that they can be difficult to identify and engage, and their individual voices lack unification, amplification and clarity. This is where Enginuity’s Policy Centre can really come into its own, creating the epicentre between SMEs, Government and others, ensuring that positive and productive engagement and dialogue take place.” 

Policymakers, partners, and businesses interested in supporting or collaborating with SMEs and the supply chain are encouraged to contact The Policy Centre using the forms on the webpage.

enginuity.org/policy-centre

MACHINE BUILDERS supplying customers in the European Union will already be aware of the Machinery Directive. Its replacement, the Machine Regulation, comes into force in January 2027 and introduces requirements relating to cybersecurity. However, there is another new piece of European legislation that already has implications for machine builders and system integrators. This is Regulation (EU) 2023/2854 on harmonised rules on fair access to and use of data – which can be shortened to the Data Act. Although the Data Act covers a wide range of products, this present article is concerned solely with how the legislation impacts machine builders and system integrators; for simplicity, we will just refer to machine builders.

What is covered?

Regulation (EU) 2023/2854 has been applied since 12 September 2025, though some aspects do not apply until September 2026 or September 2027. It covers ‘connected products’ and, for the avoidance of doubt, this includes products with on-device access, products with wireless connectivity, and products that require a physical connection to be made when needed. ‘Data’ includes data generated by use of the product or related service, metadata necessary to interpret and use the data, and data created when users interact with the product. Even if data is only stored and not processed, then it still falls within the scope if it can be accessed.
Paragraph 14 of the preamble lists various types of connected product, with industrial machinery being one such type. This paragraph also states that prototypes do not fall within the scope of the Data Act, but machine builders should not assume that a one-off special-purpose machine is exempt, even though it could be argued that it is a prototype. Article 31 excludes custom-built data processing, as well as data processing services provided as a non-production version for test/evaluation over a limited time period.

If any data can be accessed by the machine builder, then it is covered by the Data Act. It must therefore be sharable with the end user and, by implication, third parties. On the other hand, information that has been derived from data is excluded from the scope of the Data Act and does not need to be sharable. If data, such as from sensors, is processed but not stored, then it does not need to be sharable. Personal data is covered by other EU legislation, though the Data Act covers personal data that has been anonymised.

Article 7 states that the Data Act does not apply to products manufactured or designed by microenterprises and small enterprises provided they do not have a partner enterprise or linked enterprise and the enterprise is not subcontracted to design or manufacture the product. The same applies to an enterprise that has qualified as a medium-sized enterprise for less than one year, and to connected products for one year after the date on which they were placed on the market by a medium-sized enterprise.

Why is the Data Act needed?

The Data Act recognises the value of data for businesses, consumers and society, largely as a result of the ‘Internet of Things’ (IoT). Furthermore, the European Commission believes that high-quality and interoperable data increases competitiveness and innovation and, therefore, ensures sustainable economic growth. Consequently, the Data Act aims to make it easier for users to share data with third parties or use it themselves, rather than having the data restricted to being stored or processed by, for example, a machine builder. The situation is the same, whether the user has purchased, leased or rented the product.
Standardisation

In common with many EU Regulations, the Data Act contains essential requirements that must be met. In this case, the requirements relate to the form of the data and its usability. Data must always be accessible to a user easily, securely, free of charge, and in a comprehensive, structured, commonly used and machine-readable format.

Clauses in the Data Act provide for harmonised standards that, if complied with in full, would provide a presumption of conformity with the essential requirements. In the absence of such standards, ‘common specifications’ can provide a presumption of conformity. At the time of writing, no harmonised standards or common specifications have been published but these may follow in due course.

Contractual arrangements

When a machine is placed on the market in the EU, whether for sale, lease or rent, information about sharable data must be provided before a contract is concluded. This includes the data functions available, how they can be accessed, the type and volume and format of the data, whether data is generated continuously and/or in real time, and the nature, location and retention period of data.

A contract must cover the basis for a manufacturer’s use of product data, and the terms could exclude or limit the user from accessing all or some of the data. Some data might be classified as trade secrets, in which case the data holder can require data users to treat it as trade secrets.

Within the Data Act, there are clauses to prevent product suppliers from imposing unfair contractual terms on customers. The EC has published non-binding model contractual terms in a document ‘Final Report of the Expert Group on B2B data sharing and cloud computing contracts.’ Nevertheless, Article 1, Clause 6 of the Data Act states that it does not apply when voluntary agreements are in place for exchanging data.

Compensation

If a data holder (such as a machine builder) is requested by the user to make data available to a third party, then the data holder can require the third party, not the user, to pay reasonable compensation for the cost of providing the data, but not for the data itself.

The EC has foreseen that levels of compensation might be contentious, so the Data Act sets out arrangements for resolving disputes and lays the foundations for dispute settlement bodies that can decide whether compensation is reasonable.

Initially, providers of data processing services can charge users for switching between different providers. However, these switching charges will be abolished after three years.

Sharing data with authorities

So far, we have focused on situations where, typically, a user wishes to share data with a third party of their choosing. In addition, the Data Act covers the requirement for data holders to make data available to public sector bodies, the Commission, the European Central Bank and Union bodies when there is an exceptional need, such as in the event of a public emergency. Data holders are entitled to compensation for making the data available. Micro and small businesses are exempt from the requirement to share data with authorities.

Legal representation

For machine builders based outside the EU, a key point to note is Article 37, Clause 11: ‘Any entity falling within the scope of this Regulation that makes connected products available or offers services in the Union, and which is not established in the Union, shall designate a legal representative in one of the Member States.’ Clause 12 explains what the legal representative is mandated to do, which is essentially to act on behalf of an entity to cooperate with the relevant authorities and, upon request, demonstrate how connected products and related services are in compliance with the Data Act.

Hold Tech Files is based in the Republic of Ireland and is therefore established in the EU. Hold Tech Files performs numerous roles for non-EU machine builders in accordance with various EU legislation, including acting as a legal representative in line with the requirements of the Data Act.

Summary

From the perspective of a non-EU machine builder exporting to the EU, complying with the Data Act requires the following unless there are relevant exemptions:

- Certain information about the data and its usability must be made available before a sale, lease or rental contract is concluded

- There must be an agreement with the data user regarding which data is sharable, the characteristics of that data and how it would be shared, and this agreement must be fair to both parties

- If harmonised standards or common specifications have been published, then the data should comply with these unless it can be shown to meet the essential requirements (stated in Article 33) another way

- Data and metadata must be suitable for sharing with the data user or, upon request from the user, a third party

- Upon request from the user, the data holder must be ready to share the data with the user or a third party

- A method should be established for calculating the reasonable level of compensation that can be claimed for transferring data to a third party or the authorities

-  Before placing the product on the market, a machine builder outside the EU must appoint a legal representative who is established in the EU.

Derek Coulson is director of Hold Tech Files

www.holdtechfiles.eu

WHEN A robot stops, everything around it tends to stop too. The battery pack isn’t usually the first thing people think about, yet it’s often the part that determines whether the machine performs as it should. When it’s designed around the real duty cycle and tested properly, it works quietly in the background, keeping operations running as planned. When it isn’t, problems appear quickly - shorter run times, heat build-up or premature wear that can halt operations altogether. Consistent performance and reliable operation are the result of design decisions made early, based on how the robot will actually operate day to day.

The first step is to define how the robot will really work in service: how long it runs for, how often it charges and the environment it operates in. A unit working in a clean warehouse faces very different demands to one that spends its days on a loading dock or outdoors. Those early decisions shape everything that follows – from electrical design to how the pack is assembled and validated. Leaving these details open too long is one of the most common causes of redesigns, delivery slips and added cost.

Avoid the freeze

Without clear checkpoints, assumptions multiply and time disappears. To avoid drift and late redesigns, development needs to follow defined stages. The first, Scope Freeze, is where the technical requirements, compliance planning and project timelines are agreed so that everyone is working to the same expectations. The second is working towards Design Freeze, when the detailed design has been reviewed and validated – drawings, test plans and documentation are finalised, so the pack is ready for a prototype build. The final stage is to validate costs, including the bill of materials, 3rd party supplier quotes, tooling, and commercial plan in order to reach Cost Freeze before moving to production. Taking this step-by-step approach keeps teams aligned, prevents scope creep and gives customers confidence that progress is controlled and measurable.

Once the application of the end-product is fully understood, chemistry choice becomes the next key decision. Lithium-iron-phosphate (LFP) cells are stable and long-lasting, making them a reliable option for robots that operate continuously. Nickel-manganese-cobalt (NMC) offers higher energy density where space is limited, which suits compact autonomous mobile robots (AMRs) and drones. Lithium-titanate (LTO) performs well for fleets that need very fast charging or work in colder environments. There’s no universal answer; the right chemistry is the one that balances energy, cost and weight in line with the duty cycle. An experienced battery design and manufacturing partner will help an OEM’s project team to weigh up the pros and cons for different chemistries and will guide the right decision based on each product’s specific requirements.

It’s also worth deciding early how much usable capacity the pack should retain after a shift. Too little and the robot may fail to complete its route; too much adds unnecessary cost and weight. A good rule of thumb is to design for around 80% of original capacity at end-of-life, as this aligns with common industry definitions of usable battery life and keeps performance consistent over time. Having these realistic conversations early prevents costly field issues and builds shared understanding across the project team.

One-off engineering

Developing a custom pack always involves an element of one-off engineering. Drawings, simulations, prototypes, test engineering and certification, together known as non-recurring engineering (NRE), are what turn an idea into a manufacturable product. Planning for this work up front shortens development overall and helps avoid costly redesigns. It also means cost, safety and performance targets can be validated together, rather than discovered late in the project. It’s what turns a one-off prototype into a product that can be built consistently and safely at scale.

Regulation is now an equally important consideration. From 18 February 2027, industrial, EV and light-transport batteries above 2 kWh must carry a digital passport accessible via a QR code. Larger robot packs will fall within scope first, while smaller systems may follow as the rules develop. Building traceability into the design - linking cell batches, process data, firmware versions and serial numbers - means the information needed for the passport already exists, rather than having to be recreated later.

Although it may sound bureaucratic, the benefit goes beyond compliance. A clear data record allows maintenance teams to trace faults quickly, gives customers the evidence they need during audits and simplifies recycling at the end of service life. It also protects OEMs if safety or performance questions arise years down the line, because the details are already documented.

Battery design rarely gets the spotlight, but it’s what underpins reliability. Getting the fundamentals agreed early, planning the engineering work properly and thinking ahead to new regulations makes the difference between a programme that runs smoothly and one that constantly needs attention. Reliability isn’t luck, it’s the result of disciplined design. With that groundwork in place, the battery becomes the most dependable part of the robot - the one nobody has to think about at all.

Will your next robot battery be built that way from the start?

Mark Rutherford is CEO of Alexand Battery Technologies

www.alexandertechnologies.com