Win or lose: A CEO’s reflections on Artificial Intelligence and Additive Manufacturing

Artificial Intelligence is reshaping industries, and Additive Manufacturing is no exception. AI is changing how businesses operate, innovate, and compete. This shift raises important questions within the AM industry: How will AI impact existing business models, and how should companies adapt to increasingly dynamic environments?

As chairman and CEO of FEHRMANN Tech Group, I have spent seventeen years working at the crossroads of materials science, AI and industrial innovation. With a focus on applying AI-driven insights into practical solutions in materials development and manufacturing, I have been invited to share my perspective on the growing role of AI in Additive Manufacturing. I specifically discuss AI’s opportunities, challenges, and implications for the AM industry, highlighting what I believe top-level decision-makers should be aware of. Further, I consider how senior executives within the industry may need to embrace AI as it becomes increasingly integrated into AM.

Innovation is in AM’s DNA

Countries investing in R&D and scientific development face a common challenge: how to turn their investment in science into tangible benefits for society, such as jobs, safety, comfort, and GDP growth. Additive Manufacturing has been a prime example of successful science-to-business transfer. Despite its disruptive nature, which has challenged conventional manufacturing, AM players have successfully brought the technology to industry readiness, overcoming numerous hurdles along the way.

AM was envisioned to revolutionise sectors such as automotive, aerospace, healthcare, and consumer goods. Its ability to create complex geometries and customised parts opened new horizons for design and production. The technology promised significant advantages: reduced material waste, shorter production times, and the potential for localised manufacturing. However, while many new developments hold such potential, we are all too aware of the limits and constraints that still exist.

To me, the reason for AM’s success lay in the industry’s DNA: visionary entrepreneurs and great engineers who believed in its potential and worked tirelessly to bring it to reality. Think back to the vibrant energy at past conferences, exhibitions, and industry gatherings, when AM was predicted to be a ‘game changer’ for almost everything that had to be produced.

The state of the AM industry today

We all know what followed. Overhype led to a lack of focus on business cases, resulting in unmet expectations and cancelled investments. AM has, so far, failed to establish itself as a significant production technology in industries such as aerospace, automotive, and complex part production for oil and gas. Even the brief resurgence during COVID-19, when resilient supply chains and decentralised production became strategic priorities, did not drive the established players to make meaningful improvements to integrate AM into production floors.

Naturally, the focus shifted to added value, tolerances, and quality standards – key factors for progress. Developments such as multi-laser machines, new materials, and industry-specific quality standards were necessary steps forward. However, it became clear to decision-makers in target industries that AM had lost its reputation as a driver of innovation. Worse still, many players – particularly in the Western world – became preoccupied with protecting their niche and celebrating incremental advancements rather than pursuing truly transformative progress. Meanwhile, conventional technologies, often dismissed as ‘old and boring,’ proved their capabilities with innovations such as gigacasting.

Today, established AM players face multiple challenges. Costs, size limitations, and materials still restrict wider adoption. At the same time, Chinese companies have built one of the most competitive and comprehensive AM markets, flooding the industry with lower-cost alternatives. As if that weren’t enough, the rise of generative AI is reshaping the landscape entirely. Now is the time to reignite the spirit that took AM from a scientific experiment to an industrial application: the drive for disruptive innovation.

The AI age

AI is undeniably influencing nearly every aspect of business and industry. It strongly affects all of our business models – the way we work, produce, develop and lead organisations. Just over two years ago, in November 2022, the release of ChatGPT marked what can be seen as the fifth industrial revolution: the so-called AI revolution. From a societal perspective, this was the beginning of the AI age – or should we call it the CCC (curiosity, creativity, and communication) age? This event signalled the dawn of an era in which AI will continue to challenge human dominance across all sectors of work and life.

In the early 2010s, a visionary debate between Google’s Sergey Brin and Elon Musk revolved around whether advanced AI should surpass human intelligence (Brin) or whether humans should be safeguarded as the dominant species (Musk). This debate – ultimately influencing the founding of OpenAI – has now reached a pivotal moment in history. As AI continues to expand into new areas, it is not a question of if artificial intelligence will overtake human intelligence, but when.

The motto ‘If you can’t win, cooperate’ seems to offer a guiding principle for coexisting with Artificial Intelligence. As the dominant species on Earth, it is our responsibility to establish a framework for life that upholds our values and laws, providing a solid foundation for peaceful coexistence.

AI for business

Since AI affects everything, it evidently affects business and AM, too. We are witnessing a disruption unlike any other, one that is more fundamental than past innovations such as electricity or the internet. This challenge requires CEOs and top-level decision-makers to adapt fully to this shift.

The first act of generative AI was about understanding the technology, advancing it, and experimenting with its capabilities. This phase lasted for about a year and led to the start of commoditising large language models (LLMs). Whether it’s ChatGPT, Claude, Gemini, Llama, Mistral, or DeepSeek, there’s now a range of sophisticated LLMs to choose from.

At the same time, the necessary infrastructure was being built. Investments in AI reached unprecedented levels globally, with companies like OpenAI (Microsoft), Gemini (Google), and Anthropic (AWS) leading the way. LLMs were deployed on the cloud to meet key organisational needs such as reliability, stability, and data security. These services have made generative AI viable for professional use.

This laid the groundwork for what Andreessen Horovitz and Sequoia Capital defined as Act 2 of genAI – focusing on solving real-world problems. This phase began in 2024, with companies tackling complex tasks across fragmented value chains.

As generative AI became enterprise-ready, many companies began with productivity improvements. Internal chatbots, like Enterprise GPTs, were implemented and filled with expert knowledge and company-specific data to expedite workflows and provide staff with quick answers to queries. The widespread acceptance of AI within organisations led to rapid adoption. Even where companies are hesitant to fully integrate generative AI, employees often use public LLMs like ChatGPT on their own, driving further adoption.

Productivity gains have been a key driver for many generative AI applications. By making knowledge more accessible and providing tools to improve efficiency, AI has streamlined operations across businesses. Early adopters were often marketing and communications departments, which rely on text-based tasks. As new tools continue to emerge, leveraging generative AI to boost productivity is becoming essential.

The challenge now is how companies are organised. Leaders should adopt agile, self-improving structures, empowering employees to drive initiatives and form self-organising teams. A clear framework is needed to maintain resilience and foster productivity. This shift is set to reshape how successful businesses operate.

For context, when ChatGPT was released, I came across an article claiming that only three people were needed to create the next ‘unicorn’ company. While simplified, it highlights how AI enables small teams to drive rapid innovation and achieve success with fewer resources than was conventionally required.

AI for AM

Let’s take a step back. As already discussed, we are living in an era where AI is transforming everything, and the AM industry is no exception. For decades, innovation and change have been central to AM, and there is an exciting opportunity to thrive if it can tap into the current wave of AI-driven transformation.

However, many of the players in our industry have lost this edge. Due to unmet expectations, AM companies have been optimised regarding financial KPIs to secure investments, reducing efforts toward the next big thing. Worse still, companies that had been known for driving innovation have shifted their focus to manifest their position, even stopping internal innovation from new players and scientific developments.

As a major shareholder of several companies and a CEO, I understand that a second-mover strategy avoids risks and investments that may not pay off. However, I fear this approach may not be sustainable in the long term. Without embracing AI’s rapid pace of innovation, companies risk falling behind.

While the framework we need isn’t fully established, I believe AI should be integrated into business performance. Beyond productivity gains, it educates organisations on using AI tools and agile development. This leads to company-wide understanding and acceptance of AI, benefitting those who adopt groundbreaking technologies early.

The EU AI Act’s requirement for organisations to use AI to teach their employees should be seen as a great push to start enabling staff to use AI for the organisation’s sake. Empowering teams instead of micromanaging is crucial. Understanding the technology, its impact, and making the right decisions is more challenging than ever and requires a CEO’s full attention.

Accelerating innovation

Let’s take another step back – solving real-world problems. What we could see in the past months was a shift from focusing on productivity gains through AI to tackling more complex tasks like product enhancement and development using AI. It needn’t be an entirely new product, but enhancing the existing portfolio with features and functions that add significant value to the customers. To illustrate briefly, here are a few examples of how AI has solved real-world problems in AM:

Quality assurance of new machines before delivery

New industrial AM machines typically undergo comprehensive quality assurance, including standardised or custom testing. The larger and more expensive the machines, the more complex the testing phase, which often takes several weeks and ties up significant capital and internal resources. Using a software tool based on a generative AI-powered algorithm and machine learning, this testing time has been reduced to mere days by feeding it with curated and experimental datasets.

In-operando quality control of printed parts

Quality control of additively manufactured parts has improved significantly in recent years, and with AI, its power increases. Deviations can now be quickly identified during the build process, stopping the entire build, guiding QA staff on what to test specifically, or even remedied automatically. This has unlocked previously unforeseen savings. By having more evidence on quality measures, final quality assurance now focuses only on areas where expert knowledge or evidence is lacking.

Ultrafast development of advanced materials

Traditionally, product development involved optimising geometries with standard materials, taking years for material development. AI has revolutionised this process, showing that optimal materials can be developed in hours and validated in days. For instance, a new printable material with unmatched levels of electrical conductivity was developed and validated in just two weeks. This was achieved through advanced machine learning, simulating materials from nano- to macrostructure, incorporating FEM for Powder Bed Fusion based on expert-curated data and rapid experimental validation.

A recent MIT study has shed light on AI’s growing impact on materials development, revealing that AI-assisted scientists discovered 44% more materials than their non-AI-assisted counterparts. This advantage translated into a 39% increase in patent filings and a notable 17% boost in product innovation [1].

Ramping up series production fast

AI’s impact is evident in its ability to optimise machine parameter settings, saving time and reducing the need for test builds, which previously blocked machines and delayed production. AI has empowered machine software, bringing significant value by enhancing efficiency and reducing delays.

There are three key requirements for all of these examples: reliability, domain knowledge, and security. Reliability has always been a significant challenge in building AI-powered tools for business. You simply can’t afford to hand over tools to customers or staff that you can’t trust. This is why tools like ChatGPT, which draw from the entire internet, are unsuitable for business. Even if the error rate is just 5%, that’s still too much. As the saying goes, garbage in, garbage out. Input must be curated, and transparency and quality controls must be in place. And let’s face it: you can’t use one LLM for everything.

Expert domain knowledge will continue to be crucial for AI in business. While AI will undoubtedly take over more and more tasks, humans will still be needed. So much knowledge has yet to be published, and AI is only useful if you can trust the outcomes. If you can’t assess the quality of AI’s work, it’s pointless.

Then there’s security, especially data security. This is non-negotiable for any company. Cloud infrastructures have often been met with scepticism, particularly by organisations that rely on their expertise and know-how. It allows non-experts to take on tasks once reserved for specialists, which can pose a real threat if not managed properly. One bad decision and a company that has spent decades building expertise could lose everything. There’s no choice but to embrace AI where possible, but the crucial questions are what tools to use, where to run them, and what data to feed into them.

There’s no free lunch. Take DeepSeek – everyone knew its knowledge would eventually be accessible to Chinese authorities. The same applies to US-based LLM providers. Protecting your data has to be a priority. Day-to-day information can be handled differently from critical business knowledge, but Artificial Intelligence is reshaping industries, and Additive Manufacturing is no exception.

For CEOs, the challenge isn’t just understanding AI’s potential but strategically integrating it to drive efficiency, innovation, and competitive advantage. In this article, Henning Fehrmann, chairman and CEO of FEHRMANN Tech Group, considers AI’s real-world impact on AM. Drawing from personal experience, he offers insights on how AM industry leaders can leverage AI to strengthen their businesses, adapt to market shifts, stay competitive, and safeguard their core data. It’s not surprising that many companies are now building hybrid infrastructures, even if the cloud is still an option. High prices are a concern, but the bigger issue is securing data.

But these decisions aren’t without consequences. You’ll need to invest in your own infrastructure, especially since internal development is no longer enough. With new tools emerging every week, the key is forming strategic partnerships with AI providers offering the best-in-class tools that are responsible and trustworthy. These tools must be infrastructure-agnostic, and you’ll likely need to increase your budget to ensure redundancies.

Henning Fehrmann, chairman and CEO of FEHRMANN Tech Group and its subsidiary, FEHRMANN MaterialsX, is recognised for his expertise in materials science, AI, AM, and augmented reality.

Founded in 1895 in Hamburg, FEHRMANN Tech Group is a family-led company that has evolved into a high-tech business group. The group develops advanced materials and AI-powered tools for materials innovation, integrating digital information into practical applications. Notable innovations include high-sea-safe windows for Onassis’ yacht Christina, the MatGPT® AI platform for rapid materials development, and advanced aluminium solutions for the automotive industry.

FEHRMANN MaterialsX is one of the world’s most active aluminium alloy developers, with a decade of experience in AM materials, including the AlMgty® family of aluminium alloys. These alloys, designed as a next-generation alternative to AlSi10Mg, are optimised for PBF-LB and offer corrosion resistance as well as higher strength and ductility. They are also capable of being processed by casting and extrusion.

Henning serves as chairman of the Innovation Advisory Committee at DESY (Deutsches Elektronen-Synchrotron, Germany’s Electron Synchrotron), spokesman for 3D-Druck Nord, and an advisory board member of the Center for High-Performance Materials and the Federal Institute for Materials Research and Testing. Additionally, he played a role in founding the Artificial Intelligence Center in Hamburg. His contributions have earned him recognition, including Hamburg Person of the Year in Business (2022) and Family Entrepreneur of the Year in the Hamburg Metropolitan Region (2011). Additionally, he is the CEO of YNICORN, an AI fintech start-up that uses automated data analysis to assess business model maturity.

AI in the (near) future

I don’t want to focus too much on the improvements and disruptive developments that you are probably already familiar with. Nobody knows when we will reach Artificial Superintelligence – the direction is still irreversible. However, highlighting a few developments may be helpful:

AI will be used and implemented everywhere. Agents will solve rows of tasks, automating entire processes, including quality checks. Custom LLMs or similar techniques will be used regarding the type of tasks added by specialised tools from experts providing best-in-class solutions for verticals such as metal or plastics Additive Manufacturing combined with industry-specific applications.

Local LLMs may win against the cloud-based competitors due to the aforementioned reasons.
New technologies of generative AI beyond current techniques, such as knowledge graph-based or neurosymbolic AI, will enhance or even disrupt the existing tools.
Finally, I hope that ‘responsible AI’ will become a pleonasm, at least for business.

Concluding thoughts

The current times seem challenging, but they offer unique chances and opportunities. And the risks are not as high as you would imagine. As domain knowledge, experience, and reputation remain crucial to business success, the chances are good that incumbents will strengthen their positions and open new fields of business.

As AI continues to influence Additive Manufacturing, companies will need to adapt to these changes to stay competitive. Integrating AI should be viewed as a practical step for improving operations and addressing current challenges rather than a distant aspiration. By approaching the technology thoughtfully, businesses can maintain relevance in the evolving Additive Manufacturing industry.

Author

Henning Fehrmann
Chairman and CEO
FEHRMANN Tech Group
www.fehrmann.tech/en

References

[1] Toner-Rodgers, A. (December 2024). ‘Artificial Intelligence, Scientific Discovery, and Product Innovation’, https://arxiv.org/abs/2412.17866