That’s the question that led us to commission our latest global survey: the State of Engineering AI 2025.

We spoke to 300 senior engineering leaders – CTOs, VPs of Engineering, Simulation leaders – across the US and Europe, to understand how prepared engineering organizations really are to adopt and scale AI in their design engineering and simulation workflows.

The results are fascinating, and for me, both a clear validation of the opportunity and a sharp reminder of where the real work lies.

AI Ambition Is Not the Problem – Execution Is

The headline is simple:

• 93% of leaders expect AI to drive productivity gains.
• 30% expect those gains to be “very high”.
• But only 3% say they are achieving that level of impact today.

This is a massive gap (10:1) between current ambition and experience – what we’re calling the Engineering AI expectation v. execution gap.  It’s also not unique to Engineering, many industries go through this phase. But the depth of this gap in Engineering is shaped by some very specific challenges:

Why Engineering Is Different

Unlike fields where large-scale public data and cloud-native workflows are the norm, engineering teams face structural barriers that AI alone cannot magically remove:

• Siloed data: 55% of leaders cite fragmented, inaccessible data as the top barrier to AI progress.
  
• Legacy tools: 42% cite the limitations of traditional desktop CAE tools. Many workflows are not cloud-native or even cloud-connected.
  
• Leadership disconnect: Interestingly, 42% of CTOs perceive significant resistance to AI adoption within their teams, but engineering leaders themselves report this only 25% of the time. In other words: many teams are more AI ready and enthusiastic than leadership assumes.

And finally, engineering data itself is often fundamentally harder to leverage for AI than the text or image data used to train other types of foundation models. This is why I believe the evolution of Physics AI and Engineering AI will take a path that is very much grounded in accelerating the adoption of cloud-native tech stacks across engineering workflows.

What the Leaders Are Doing Differently

The good news is that our survey clearly shows a cohort of engineering leaders who are already achieving transformational results.

These teams share several traits:

• They have modernized their toolstack – favoring cloud-native, open platforms.
• They have invested in ensuring centralized, clean engineering data is captured across workflows – not perfectly, but enough to enable scalable AI.
• They are building and integrating AI agents directly into live workflows – not as bolt-on tools, but as embedded decision-makers at setup, evaluation, and optimization stages.
• They have moved from pilots to production-grade AI use cases that drive real business value (faster design cycles, improved product performance, faster time to market) – rapidly, with confidence, and with clear mandates.
• Critically, they have fostered strong alignment between leadership and engineering teams -AI initiatives are not being led in isolation.
  

Cloud-native users in our survey are 3x more likely to have mature AI programs and 6x more likely to have clean, centralized data – and they are twice as confident they’ll achieve their AI goals in the next 12 months. It’s clear that confidence in AI follows capability with cloud-native CAE tooling, rather than the other way around. 

Where We Go From Here

One of my favorite lessons from the many conversations with engineering leaders had while creating this report is simple:

Don’t aim to “do AI.”
Aim to solve engineering problems better – with AI as a transformational enabler.

Teams that start with a clear, high-impact use case – where collapsing a process from days to seconds changes outcomes – make the fastest progress. Engineering AI is not about replacing engineers. It’s about creating machine-in-the-loop workflows that supercharge engineering creativity and productivity.

The goal is not to bypass human insight, but to multiply it, to deliver unseen levels of engineering innovation. 

My Call to Action for Engineering Leaders

If you are a VP of Engineering, CTO, or simulation leader reading this:

Be aware of the expectation-execution gap – don’t let your organization be part of the “93% hoping, only 3% achieving” statistic.
Look hard at your toolstack, your data readiness, and the leadership alignment needed to move forward. Does your legacy tooling hinder or help AI adoption?
Start with one high-value application and push hard; prove out the impact, then scale with confidence.

And above all: the time to start is now. Engineering AI is no longer a future vision or add-on capability, it is a fundamental enabler and accelerator, and is already transforming how some teams design and innovate today.

We created this report not just to benchmark the market, but to help drive the conversation forward. I encourage you to read it, and more importantly, to act on it!

I look forward to hearing what you think.

David

The post The Engineering AI Ambition-Execution Gap: What Our New Global Survey Reveals appeared first on SimScale.

Our team here often says:

We are building SimScale with our customers, not just for them.

Every SimScale employee.. ever

And that’s not just words — it’s how we work. Over the years, our conversations with customers, partners, and now a global community of over 700,000 engineers have deeply shaped the platform we build and the ideas we believe in.

Today, we’re unveiling a clear vision of SimScale’s mission – one that better reflects how we’re helping our customers advance their engineering innovation!

You’ll see us talk about Engineer the Irreplaceable – and I wanted to take a moment to explain what that phrase means to me, why we chose it, and why it captures something very real about what our customers are doing every day.

When I started SimScale with my co-founders, our vision was simple: simulation should not be a bottleneck to innovation.

Too many engineers were stuck waiting!

Waiting for software licenses, waiting for compute capacity, waiting for simulation runs to finish, waiting for IT to support them…

Meanwhile, the pressure to innovate – and innovate faster – was only increasing.

At the time, cloud-native simulation was still a radical idea. Today, it’s becoming the norm.
But what I see again and again from the engineers we work with is that they are building products that truly matter.

Products that aren’t just good. They are irreplaceable:

• The wind turbine blades that need to operate flawlessly in extreme conditions.
• The medical devices that patients and doctors rely on every day.
• The new forms of mobility that will define the next generation of transportation.
• The complex machinery that powers entire industries.

And behind every one of these products is an engineering team that faces huge complexity, shrinking timelines, and relentless market pressure — and yet still pushes the boundaries of what’s possible.

When we developed our new vision, we didn’t invent it in a closed room.
We distilled it from the real words, ideas, and spirit we hear from our customers.

A leader at one of our automotive customers put it perfectly during this process:

The things we build cannot fail. They have to perform — every time.

That’s what it means to Engineer the Irreplaceable.
And for me, this is what SimScale’s mission is about.

We talk a lot about Physics AI and Engineering AI. And yes, the technology is exciting — in many ways, we’re just beginning to see what AI will enable for engineering teams.

But in the end, it’s not about AI. It’s about the products engineers create with it.
It’s about giving teams back time and capability, so they can engineer more freely, explore more boldly, and deliver designs that truly stand the test of time. It’s about amplifying the impact of human engineering expertise.

I think a lot about the phrase “switching time from an enemy to an advantage.”
This is exactly what so many of our customers are trying to do. They’re under incredible pressure, but are also full of ideas. The role of SimScale is to enable them to move faster without compromising quality, without compromising insight, and without compromising what makes their work meaningful.

So why share this today? Because this new SimScale story is not just marketing language.
It’s the result of years of building with you – our customers and community.
And it’s the idea that will guide us as we continue to evolve SimScale.

We are here to help every engineering team – from the world’s biggest enterprises to the most ambitious startups – Engineer the Irreplaceable.

Not every product needs to be beautiful. Not every product needs to be novel.
But the ones that matter – the ones that drive real progress – well they are truly irreplaceable. And so are the teams that build them.

Thank you to every engineer who is part of this journey with us.

David

The post Engineer the Irreplaceable — Why This Idea Matters So Much to Me appeared first on SimScale.

This growth has also been fueled by an equally rapid expansion of model capabilities. The first steps toward multi-modality came quickly and introduced the same text-to-output inference to other content types. We already almost take for granted the ability to generate high-quality images, video, and source code through such tools.

Can AI Generate Engineering Output?

Mechanical engineering teams have adopted these tools as well to accelerate all sorts of work processes. For example, to analyze and summarize RFQs faster or to search faster for technical information. But these use cases are mostly adjacent to the core engineering work and mechanical design. So why is it that we can ask AI to generate very useful text, images, video, and code but not a useful engineering design?

Let’s consider how these types of AI models are trained. Generative AI models have been trained on trillions of tokens, primarily from the internet. Transformer models on huge datasets of public text/code and diffusion models on equally large datasets of text-image pairs. Not only is this training data available in vast quantities, but the data formats are also very straightforward to read and use for model training.

Things look rather different in the engineering realm, the most obvious challenge being that, unlike text or source code, there is little to no public product design engineering data available. Then there is also the question of data quality, in the sense of whether or not a given design is fit-for-purpose, meeting the requirements that it was designed for. Added to that is the fact that the most widely used data formats storing mechanical design information are proprietary, requiring commercial licenses even to read it, let alone manipulate it. In summary, the idea of obtaining and processing millions of engineering designs to train a generative model still looks like a very challenging ask today, but technical progress in this field is happening fast.

Does That Mean That Core Engineering Work Will Remain AI-Free for Now?

Absolutely not. In due course, novel AI approaches might rise to the challenge of handling big chunks of typically manual engineering workflows, possibly including the transformation of a text prompt into a meaningful design, but it is going to take time to get there.

Meanwhile, there are AI engineering workflows that are easier to attain while still very helpful. We can get a long way by using AI to speed up the cycle time for a single design iteration to such an extent that it appears to be instantaneous. We will do this by accelerating all of the steps in the workflow, including CAD generation, model preprocessing and setup, simulation workflows, and the analysis of results.

Once we have all that proven out, an AI agent can then drive the (accelerated) machine, taking design decisions along the way and looping around to discover optimal solutions.

Replacing a human-in-the-loop with a machine-in-the-loop in this way has the advantage of leaving the workflow and toolchain fundamentally unchanged, with the AI system ‘driving’ the tools in the same way that a human does. This means the human can easily understand what is being done and intervene at any point. Most importantly, the human can provide input to direct the AI, for example, where a design needs to balance competing objectives – decisions that require careful consideration and mutual understanding.

Not Just a Case of “Prompt Engineering”

Let’s dig into how we deploy AI to accelerate and augment engineering workflows. Let’s start by looking at how these processes work today. They tend to be centered around the manual engineering work where humans make decisions to advance the iterative design by designing and evaluating the design’s performance, depicted in green below. The CAD system involved can be conceptualized as a computational process going from parameterization to geometry (yellow) and the CAE system going from the simulation setup to the results (blue). 

This is a very simplified conceptual view of the engineering process, but helpful as it differentiates between the unstructured, human workflows in the middle and the purely computational ones left and right. All three can be automated already, to search through a prescribed design space for example. But this automation is very much rate-limited when using so-called traditional physics solvers to evaluate each design. What’s more is that AI can transform this process into something not only automatic, but autonomous.

Introducing Physics AI & Engineering AI

Let’s tackle that first bottleneck of simulation run time (the right-hand block in the diagram above). Depending on the physics and fidelity needed, a computing time of hours to days is not unusual. A growing set of AI methodologies to speed up this solve process is available, from deep learning surrogate models that replace full physics solvers to tools that speed up those ‘traditional’ solvers. Given the availability of a suitable, pre-trained, method, you can reduce the solve time almost to zero. We call these ‘Physics AI’ methods to indicate that, at the core, it’s about predicting physics with AI, and with the big benefit of being able to do that very fast. 

The second, more dispersed bottleneck visible in the process is the human interaction needed to go from a given design to a well defined simulation setup, then to consider the results of that simulation, and lastly to determine which point in the design space to look at next (the middle block in the diagram). These are all steps where an AI agent can assist, facilitate, accelerate, as well as act autonomously – performing complete workflows by operating on the existing tool stack just as a human would. As such, it is performing a series of discrete and logical steps that can be justified or even debated, as you might with a colleague. Since this agent is performing the core engineering work for you, we refer to it as ‘Engineering AI’.

Lastly, let’s turn our attention to the left-hand block – the CAD definition of a design. Once a model has been created and parameterized, generating a new variant based on a new set of parameters is already near-instantaneous. What is very much slower, though, is the process of creating that CAD model in the first place.

There are several exciting technologies emerging in the CAD space that could make the process of CAD generation far faster and more robust. Latent space parameterization, implicit representations, and cloud-native BREP are just three such technologies that could enable vastly faster design iterations, and we are actively working on integrating them into SimScale.

We Are Placing AI Tools in the Hands of Every Engineer

Thanks to its cloud-native architecture with built-in AI infrastructure, SimScale is uniquely able to provide AI features to help you navigate engineering workflows and accelerate performance predictions by leveraging your simulation data in the cloud. As we have explored so far in this blog, unlocking value from AI means touching almost every aspect of the simulation workflow. It requires a deep and immediate connection to models and data which is only practical to do in a cloud-native stack.

Engineering AI and Physics AI are built into SimScale in such a way that it can become second nature to use these tools to supercharge your productivity. SimScale users do not need to deal with any of the typical headaches experienced when attempting to deploy AI tools such as data cleaning/organizing/relocation, model versioning and management, or provisioning of suitable GPU resources for model training and execution. All of these are taken care of by the vertically integrated tool stack and intuitive user experience.

At NVIDIA GTC 25, we announced that we are making it even easier and faster to adopt Physics AI for certain applications by building a set of pre-trained foundation models. The unique aspect of these models is that they are pre-trained on a broad set of designs, providing users with a Physics AI model that they can use out-of-the-box or that they can augment with a small amount of their own proprietary training data. To learn more about foundation models in SimScale, check out this blog.

Unlock AI Value by Selecting an Impactful Application to Start With

Once you have test-driven the capability, the next step is to test-drive the value unlock. Each engineering team we work with has unique legacy data stored, sometimes from decades of engineering work. We frequently see teams expecting to start there, trying to find value in it. The reality is that finding and processing legacy data can be an immensely difficult task, and one that may take a very long time to yield results, even if useful data exists.

We recommend a different approach: Select an engineering process in your organization that – if collapsed to seconds – would create hard value for your organization (revenue or costs) and try tackling that with an AI-powered workflow. 

Remember: The best time to start leveraging AI systems in your engineering team was yesterday. The second best is today – give us a ring!

The post AI Is Sweeping Into Knowledge Work. What About Engineering? appeared first on SimScale.

Why This Launch Matters

Engineering teams are facing unprecedented pressure to innovate faster, improve efficiency, and reduce costs. Yet traditional simulation methods and technologies—slow, manual, and heavily reliant on costly on-premises infrastructure—create bottlenecks that continue to limit innovation. With our new AI Foundation Model, powered by NVIDIA PhysicsNeMo and embedded directly into SimScale’s cloud-native platform, we are removing those barriers for engineers worldwide.

This launch marks another step in our mission to empower engineers to innovate faster. By combining advanced AI models with cloud-native simulation, we’re further unlocking true computational engineering. We are inviting you to imagine a world where:

• Engineers can explore thousands of design options in seconds, not days.
• Design evaluation time and costs are collapsed to near zero.
• Instantaneous design exploration and optimization are brought into reach.

Start Exploring from Physics AI Basecamp

An AI Foundation Model is an ideal way to start adopting AI in an engineering workflow, effectively a “base camp” from which further exploration is immediately possible.

This first foundation model is trained on 1000s of validated computational fluid dynamics (CFD) simulations of centrifugal pump designs. The fact that the model is pre-trained on a comprehensive data set means that engineers can immediately pick it up, apply it to their latest designs, and get valuable performance predictions in seconds.

We’re continually expanding the dataset used and are expecting the application space to grow over the coming weeks. If your own organization has its own dataset of proprietary designs, you can use a foundation model as a basis from which you can create an enriched, tailored version that produces even more accurate results for your designs. You can also create your own foundation models from scratch using the built-in AI model training infrastructure.

Our first Foundation Model covers the design space for centrifugal pumps and we will be introducing other models to cater for different families of devices that our customers are working on. It is all enabled by integrating NVIDIA PhysicsNeMo into the SimScale platform, using physics-based simulation data with machine learning to deliver results at the speed of AI without compromising on the fidelity engineers demand.

By embedding this model directly into the SimScale platform, we enable engineers to:

• Instantly evaluate pump designs, exploring thousands of design points in seconds rather than days.
• Maintain high accuracy in predictions comparable to traditional CFD methods.
• Seamlessly move between physics simulation and AI by leveraging cloud-native AI models.
• Scale their simulation capabilities globally without added infrastructure costs.

By making accurate Physics AI models much easier to access and get started with, foundation models really have the potential to accelerate AI adoption. If you are curious, find out how to access this model below.

How It Works

Our AI Foundation Model is built on a combination of physics-based simulations, machine learning, and cloud-native deployment:

1. Pre-trained Physics AI: Leveraging NVIDIA PhysicsNeMo, our team trained the AI model using thousands of high-fidelity CFD simulations. This ensures the model captures accurate pressure, flow, and efficiency behaviors across a broad range of pump geometries and operating conditions.
2. Cloud-Native Integration: The model is integrated into SimScale’s intuitive, browser-based platform, enabling engineers to scale simulation use earlier in the design process without any installation or hardware constraints.
3. Instantaneous Optimization: Using optimization coupled with the AI model, engineers can explore thousands of design variants in minutes—identifying optimal configurations far faster than traditional CAE methods.

Unlock Instant Design Optimization

Such easy-to-access physics AI models unlock instantaneous design space exploration. By leveraging Physics AI with SimScale’s cloud-native infrastructure, engineers can now run thousands of design iterations in minutes. This eliminates the traditional “trial-and-error” process, accelerating time-to-market and reducing costs.

Imagine an engineering team tasked with optimizing a new design for efficiency, pressure, and flow. Traditionally, this process requires running separate CFD simulations for each design iteration—often taking days or weeks to explore all viable options. With Physics AI, that same exploration happens in seconds—giving teams the insights they need to make smarter design decisions faster.

Using an entirely cloud-native toolchain consisting of Onshape CAD, SimScale (including the foundation pump model built with NVIDIA PhysicsNeMo), and Google Colab with open-source optimization libraries, we ran an optimization covering study 2700 designs in less than one minute. This methodology really can unlock transformational engineering velocity!

Our Vision: Engineering AI + Physics AI

This launch is more than just a product announcement—it’s a demonstration of what we believe is the future of engineering simulation. At SimScale, we believe that in order to unlock the transformational value that AI can provide to engineering design, you need two branches of AI capability at your fingertips: Engineering AI + Physics AI.

Engineering AI accelerates complete engineering workflows, assisting or automating simulation setup, running design studies and even proposing and implementing design changes autonomously. Engineering AI is both a co-pilot and dependable teammate that helps you meet your engineering challenges and communicates with you using natural language.

Physics AI leverages AI models trained on trusted simulation data to make near-instantaneous performance predictions to drive engineering decisions.

The combination is a powerful new paradigm. SimScale will empower engineers to move beyond traditional design constraints and embrace a new era of real-time simulation, enhanced insight, and accelerated innovation.

Our vision extends far beyond this initial launch. The AI Foundation Model for pump simulation is just the beginning. By investing in AI-first engineering workflows, we aim to democratize access to world-class simulation capabilities for all engineers—from startups to global enterprises.

Our belief is simple: engineering teams should be able to explore more ideas, make better decisions faster, and accelerate their path to market. AI will be the catalyst that makes this possible, and SimScale is leading the charge.

Join Us in Shaping the Future

I encourage you to experience this transformative technology firsthand. Our first AI Foundation Model is now available to SimScale customers worldwide.

We’re also hosting a live webinar with NVIDIA to showcase this technology in action. We’ll explore real-world use cases, demonstrate the model’s capabilities, and answer your questions directly. Register now to see how AI-powered simulation can revolutionize your engineering workflows.

The post Engineering the Future: Launching the first AI Foundation Model for Pump Simulation with NVIDIA appeared first on SimScale.

Server Room Cooling Analysis of Server Room Cooling Process

First, we upload the CAD model of the server room. The air is circulated throughout the room, entering through the inlet (lower box on the wall) and leaving it through the outlet (higher box on the wall). Hot air is streaming from the cooling outlets of the server racks into the room. The image below shows the SimScale pre-processing viewer displaying the CAD model which was uploaded as a STEP model.

The next step is to create the computational grid. In this case, we choose an automatic hex-dominant mesh for internal aerodynamics with a boundary layer refinement. A cut view is shown in the figure below. The second figure zooms into the refined cell layers at the wall to resolve the viscous boundary layer.

The final step before we can run the simulation is to define boundary conditions and solver settings. We are using a natural convection solver and adapting the numerical settings to stabilize the simulation run. This simulation took around one hour to complete on an 8-core machine.

Server Room Cooling Conclusion

The results show that the cooling concept is working efficiently. The outgoing flow of the inlet is blocked by the server racks, and the air circulates through the gaps and in between the walls and the racks.

The hot air from the servers is carried upward by the flow and convection effects, and thereby does not increase the temperature of other servers. One can now run several “what if” scenarios to increase the efficiency of the cooling process, or to adjust the detailed layout of the servers within the room.

Do you want to learn more about HVAC simulation? Here you can find more information about how CAE can help you improve your HVAC system.

Also, this article might interest you: 5 Simulation Projects for Heating, Ventilation, and Air Conditioning.

Download this case study for free to learn how the SimScale CFD platform was used to investigate a ducting system and optimize its performance.

Download Case Study

The post Server Room Cooling — HVAC Simulation appeared first on SimScale.

The Challenge

The installation situation of apparatus engineering devices can be critical for their performance. This project’s goal was to create a homogeneous flow field in front of the apparatus in order to enhance its performance.

Duct Design Inlet Duct Design Simulation

Parallel to the duct design, the flow simulation capabilities of SimScale have been used to analyze the impact on the flow pattern for each design. It was possible to use the insights into the flow behavior gained from using SimScale to make early design decisions, without the overhead of testing and prototyping.

During the project, multiple simulations were carried out for multiple designs. One of the simulations took around 30 minutes on an 8-core machine. For a well-trained user, the simulation setup takes less than 10 minutes.

The image below shows the integrated post-processing environment of SimScale with a streamline visualization of the velocity field.

Duct Design Simulation Results

The results speak for themselves. The image below shows a color map visualization of the velocity field in the middle plane of two designs. The design on the left is the original version while in the design on the right, turning vanes have been added to the corner. The flow field near the outlet (on the left) is very inhomogeneous for the design without vanes. The reason for this flow behavior is the large recirculation region behind the corner of the duct. The design on the right shows a much more beneficial behavior: the airflow leaves the duct uniformly, which was this design project’s objective.

The streamline visualization of the velocity field below illustrates the reason for the resulting flow behavior more clearly. In the design without turning vanes, a large recirculation region appears behind the corner; in this design, the efficient flow channel is reduced to almost half of the channel width. This shows that investing a small amount of time in simulation while designing an inlet duct can significantly improve the performance of the overall system.

Sign up for a free community accountand perform your own simulation with SimScale.

Download this case study for free to learn how the SimScale CFD platform was used to investigate a ducting system and optimize its performance.

Download Case Study

The post Inlet Duct Design Optimization with CFD Analysis appeared first on SimScale.