[Craftsman's Rambling Diary #2] The Future of Factories is Changing—The Potential of Siemens x NVIDIA

Hello everyone. I am a former mold craftsman.
A lot of new things happened at CES 2026 the other day, didn't they?

One is that it became clearly visible that the industrial structure of the automotive industry is changing.
Another is that the potential for physical AI has expanded all at once.
And, what I want to make the theme of this post, the thing I am personally most interested in: the "partnership between Siemens and NVIDIA." If I were to briefly state what will happen with this partnership, I feel it is that "digital twins will become a part of daily life."

Why? I feel it is because Siemens, which has until now advanced along the rationalization path that Western countries are good at, such as full automation, has accepted the concepts of "people looking" and "everyone looking together." There are many people around me who view it that way as well.
In fact, during the nearly seven years I spent at a local European company, I recognized the brilliance of Siemens' systems. When a design change occurs at a large company, the equipment manufacturer's model changes in real-time, or even though it is CAD, CNC data is output when you move the model. At the company where I worked, such a system was already in place...

However, that was something that, for example, a design engineer could use, and it is also a fact that the actual work—staring at a screen and thinking about this and that, or going to the next department to invite relevant engineers to one's desk for discussion—was no different from what I am used to seeing in Japan.

And then comes this partnership.

If you ask what will happen, physical AI will be added to Siemens' systems, and furthermore, through Omniverse, many people will be able to look, verify, and speed up decision-making simultaneously.
I believe that by having people look, it is possible to actually bring together the processes of departments and individuals that were previously independent and operating individually—such as the areas of sensibility and sensuality, or so-called design engineering, the feasibility of shapes, or the lifecycle aspects related to ease of assembly.

I will say it again and again, but I feel that "super-speed decision-making" is the most important thing for companies from now on. I understood that by teaming up with NVIDIA and performing visualization through Omniverse in areas that Siemens was not particularly good at (or so I felt), they can super-speed decision-making, overwhelmingly shorten the lead time to commercialization, and maximize profit speed.

In fact, the announcement stated that "the partnership between Siemens and NVIDIA is an initiative aimed at advanced automation and advanced decision-making in the manufacturing industry by integrating processes from design to manufacturing and operation with digital twins, GPUs, and AI, centered on the 'Industrial Metaverse' and 'Industrial AI OS'."

Specifically, by connecting Siemens Xcelerator and NVIDIA Omniverse, it becomes possible to reproduce real factories and products as high-precision digital twins and perform real-time simulation and optimization.
I understand that while the fact that people are the main players remains the same worldwide, it means that more advanced development and production rationalization can be performed.

What exactly does it entail?

◆ Concept of the partnership

If this goes well, super-speed decision-making will be realized, and there is no doubt that this is exactly what the manufacturing industry needs most.

If it is an AI that truly recognizes the reality of the field, it could become the core engine of industry. If Siemens' knowledge is utilized, the realization itself is not difficult. However, what is important is how systematically the AI creators can understand the processes in the field, the relevance of information, and the process by which results are derived from the combination of various elements; this will be the key, and companies that can design and implement AI with that understanding will likely become the main players in the future.

This is exactly what is called a closed loop in machine tools, and this will dramatically improve accuracy and speed.

◆ Specific initiatives underway

This will allow companies to build full-fidelity digital twins and deploy AI-enabled software-defined systems from the edge to the cloud. Furthermore, reading between the lines, it seems they are involving multi-physics simulators (which can perform various verifications for multiple different materials). I feel that it has truly become an amazing world.

You can take a simulated factory and drop it directly into the real world, allowing for the vertical startup of the entire facility. Of course, a world where factory construction and every piece of equipment within the factory can proceed in parallel in real-time—a world like something out of a sci-fi movie—could become a reality.

◆ Future Outlook

The plan is to deploy fully AI-driven manufacturing hubs as model cases starting in 2026 and beyond, presenting a next-generation smart manufacturing standard architecture that combines the industrial metaverse with 'physical AI'.

Key Points on Technology Stack and Infrastructure
The division of roles sees NVIDIA providing a full-stack AI platform including GPUs, Omniverse, and AI model sets, while Siemens provides industrial domain expertise and automation solutions. This combination makes it easier to build a consistent digital foundation that covers everything from design and simulation to factory control, and from edge to cloud.

Impact on Manufacturing Users
By allowing digital models created at the design stage to be verified directly within a digital twin factory, we can expect a reduction in the number of prototypes and shorter startup times. Furthermore, as AI-driven condition exploration, predictive maintenance, and automated quality inspection advance, accumulating and utilizing on-site tacit knowledge as data is expected to lead to skill leveling and improved productivity.

Recently, in conversations with many clients, I feel that digital twin considerations often start from specific themes such as individual task simulation or process automation. Moving forward, I believe that designing for overall optimization—connecting these across departments to enhance decision-making for the entire business—will become a critical point of discussion.
Summarizing the simulation and automation needs for what we want to see on the shop floor, in design, and in production engineering, and connecting them all at once through a digital twin. It is steady work, but if we can deploy this as a single flow or package, I think Japan could change, or perhaps even take a leading position.

Initially, priority will likely be given to users already utilizing Siemens' systems, but with FANUC also announcing a partnership with NVIDIA, I feel that the time has come for the world as a whole to seriously consider the effectiveness of digital twins.
Digital twins will likely expand not only into manufacturing but also into various other industries and markets.

In that context, as for Japan and our unique corporate culture, I think 'Galapagos-ization' is perfectly fine—in fact, we should just build a Japanese-style digital twin. However, I personally feel that 'the scope for JDSC's presence has expanded' from the perspective of how we can involve related departments, knowledge, and processes beyond just what is right in front of us, and how we can build higher-level digital twins by incorporating agent AI and interconnected systems, rather than just focusing on the immediate tasks.
Since JDSC brings together members with diverse backgrounds, I want to draw a new picture of the manufacturing industry by having everyone clash their ideas together.
(JDSC Ryu Tanaka)

Reference Materials
・NVIDIA Newsroom
https://nvidianews.nvidia.com/news/siemens-and-nvidia-to-enable-industrial-metaverse

・Siemens Official LinkedIn Article
https://www.linkedin.com/pulse/creating-industrial-metaverse-siemens-xcelerator-dirk-didascalou

・NVIDIA Japan Blog
https://blogs.nvidia.co.jp/blog/siemens-and-nvidia-expand-partnership-industrial-ai-operating-system/