CES is a bit like the annual meeting for the major technological innovations of the moment. It took place in Las Vegas from January 7 to 10, 2025. Among the exhibitors, the Luxembourg company Gradel, manufacturer of special automated machines, intended for the nuclear and space markets and the production of “sputtering targets” used in the glass industry.
As the company celebrates its fiftieth anniversary this year, Claude Maack, who joined the company more than 30 years ago and is now its CEO, shares lessons from his American experience.
How did Gradel make it to CES for the first time?
We have been working with the company Data Design Engineering (DDE) for three years now. I met this company of Korean origin at an event in Luxembourg New Space Event, in 2021. We very quickly chose to collaborate on research and development projects, which led DDE to open a branch in Luxembourg in February 2022.
Gradel has developed a technology, Gradel Robotic Additive Manufacturing (GRAM), which allows for reductions in structural parts. Result: weight reduction can be up to 70% compared to a normal metal part. This is made possible by endless impregnation of fibers (carbon, basalt, glass, hemp, etc.) with a resin on a robotic head to then manufacture complex 3D structures of medium and large size.
Gradel has developed a technology, Gradel Robotic Additive Manufacturing (GRAM), which allows for reductions in structural parts. © PHOTO: Gradel
But where does DDE come in? It’s very simple: the Korean company makes it possible to develop algorithms that take over the engineering process from A to Z (from topological optimization, fiber installation, winding path development and automatic programming). taking into account all the rules of design, feasibility and constraints during robotic production across the entire digital process chain).
They will clone the experience of the expertise of our experts. You should know that to develop a component, today we need three to four experts. An engineering project can last from three to four weeks and up to eight months.
To summarize, DDE uses “high performance computing” and does algorithm training. The processes that we (Gradel) are carrying out today will ultimately be done, thanks to these DDE algorithm trainings, by a single general expert who will be able to carry out the entire engineering part alone. The goal is to go ten times faster (or more!), this is a project that DDE has been working on for two years already.
In your opinion, DDE is far ahead of its competitors, 10 years, is that correct?
Yes, at CES 2025, it showed technology that no one has yet and which is developed in Luxembourg. I can confirm to you that throughout the CES show, not all the big tech companies have their technology. DDE presented an Edge device, a technology that combines “Large Multiple Models”. The revolution is that it is operational “offline”.
What does this mean? This generally means that currently, all drones, vehicles and autonomous vehicles are, with some exceptions, connected to a satellite system, local GPS, a 5G system or a Wifi network if it is local. They are all connected to a system and if someone disrupts the satellite signal, everything stops. Vehicles, whether it’s a drone, a car, a small robot, once they are no longer connected, they are lost, they stop.
The solution imagined by Data Design Engineering means that all these vehicles will be able to operate offline, without being connected to any network. The vehicle will then make its own decisions and perform specific tasks for which it has been previously trained. They can know their precise position in an environment. It is governed by a powerful mini-computer weighing just 500 grams in which the system is capable of adapting its constantly evolving decisions.
In itself, to communicate the task to the device, you must connect to send the famous “voice to text” message. As soon as the machine has confirmed the task to be performed, the connection is automatically cut.
You seem to be saying that the technology that DDE is developing is revolutionary?
I don’t know if you realize how important this is. If you think about it purely from a defense perspective. Take drones, boats or planes. All you have to do is connect them today and they will be able to do everything autonomously, remotely, without having to put people on deck, even if the signal is disrupted.
-This technology is tested and operational. We have demonstrated that we can do it. You can see the impact or the interest in certain areas, notably defense or space. Offline technology can work on any device, any application.
It can range from the wheelchair in a hospital to the semiconductor industry. All areas go through it. AI experts thought it was too complicated, that computing power was not yet advanced enough, and that the cloud was necessary. Several companies at CES came to the DDE and Gradel booth, and could hardly believe it: “Are you sure?” Now, it’s possible, DDE gave them a live demonstration.
“You should know that to develop a component, today we need three to four experts. You should know that an engineering project can last from three to four weeks and up to eight months.” © PHOTO: Gradel
How will this offline system, as imagined by DDE, change the cards?
Their technology is built and, as I said, it is adaptable to each application and for different domains. It takes two months of development to put it in place. It’s definitely going to be a turning point in the defense. Today, we try not to send any more soldiers or human beings. This is why a tank weighs between 50 and 70 tonnes, it is to protect the lives inside. If we no longer need to send humans with this offline technology, instead of sending tanks, we will be able to send multiple vehicles at a reduced budget, while increasing efficiency.
Thus, we imagine that drones could be sent several hundred kilometers below the radars and they will be very difficult to detect.
Here I’m talking about defense but there is also space, automobiles, there are no limits. We are starting to get closer to what we saw in science fiction films from around twenty years ago.
To train its algorithms and develop its offline technology, does DDE need data centers?
It’s a paradox. To train its algorithms, DDE actually needs data centers. Once these are trained, there is no longer any need for them, which makes it a much less costly and less polluting solution.
For you, was it obvious to create a partnership between your company Gradel and Data Design Engineering?
The partnership that Gradel has with the company DDE starts from the optimized software that they develop and which interested us. Several major arguments weighed in the balance: very economical computing power, less energy used and less heat to evacuate since this mini-computer is shielded and protected by lightweight GRAM structures.
To this was added light weight which allows for a light structure or equipment. We don’t need a cloud that is error-prone.
As a reminder, Gradel is a company that was created in 1965. We started with special nuclear machines. Then, we produced automated space equipment to handle satellites with ground equipment (MGSE).
Five years ago, we diversified into new technology to provide lightweight structures with higher added value. We do projects with a lot of risk and development. Now we have the possibility of adding a brain to machines and this will open up multiple opportunities for us.
