Top 5 Takeaways from Our Engineering Data Research Report:

Over the course of this series, we’ve demonstrated how the utility of CAD data has outgrown the engineering department, and that providing intuitive access to this resource across the entire company is a significant software problem manufacturers face. Our new research report, How Manufacturers Meet the Growing Demand for Engineering Data, explores how real companies are meeting this challenge.  

We commissioned Tech-Clarity to interview 15 engineering, IT, and digital engineering leaders at manufacturers with more than $1 billion in revenue, across automotive, aerospace, industrial machinery, electronics and high tech, and medtech. We asked how they decide when to build their own applications for 3D CAD and simulation data, and when to buy commercial ones.  

In this piece, we highlight 5 key takeaways from the report and what they mean for manufacturers navigating engineering data challenges and the teams creating solutions to address them. 

This piece is the fifth in our series on our original report “How Manufacturers Meet the Growing Demand for Engineering Data”. The report is out now, and you can access it here. 

Check out the first four posts in this blog series: 

• Why Engineering Data Has Become a Strategic Manufacturing Asset 

• The Hidden Cost of Engineering Data Bottlenecks 

• Why More Manufacturers Are Expanding Access to CAD Data Beyond CAD Applications 

• Why Manufacturers Are Rethinking Build vs. Buy for Engineering Data 

1. Not One Company Said Demand for 3D Data is Shrinking 

Over two-thirds of the leaders Tech-Clarity spoke with said the need to access and use 3D data across their company is growing. The rest said it has held steady, and not a single one reported a decline. 

The reasons they gave lined up closely with the factors we’ve touched on throughout this series, including: 

• Faster Innovation Cycles,  

• More Complex Products,  

• Wider Downstream Use of 3d Data, 

•  Collaboration Across CAD Formats,  

• Simulation Integration 

• Digital Twins.  

• Artificial Intelligence (of course, it's 2026, after all) 

In one example, a medtech product design leader tied the growth directly to time to market: “Using digital twins and CAD to simulate behavior before we build hardware can accelerate design time so we can get more products out quickly.” 

2. Manufacturers Buy by Default, but Almost All of Them Still Build to Some Degree 

Every company interviewed said it prefers to buy a commercial 3D solution when a suitable one exists. That preference shows in the numbers: on average, participants buy 78% of their 3D-related solutions and build the other 22%. 

Only one company said it doesn’t consider building at all. Most sit around 20 to 25% built, a few reach 40%, and a few described building as “very rare.”  

The direction of travel varies too: one participant described moving from roughly 75% built to 75% bought, while another said its share of non-commercial software has grown from about 10% to around 30% in the last three or four years because commercial tools stopped keeping pace with its needs. 

In practice, build and buy are decided project by project rather than by company-wide policy, usually by a cross-functional group spanning engineering, operations, IT, and procurement. 

For application developers, it's hard to think of a clearer endorsement of the idea that, if you build it, they will come. With limited development resources, many organizations are more than happy to license reasonably priced, capable products, so long as they can meet their demands. However, when commercial solutions cannot meet specific requirements, building becomes a more attractive option, as the next takeaway illustrates. 

3. Companies Build to Fill Gaps and Protect Engineering IP 

Unsurprisingly, cost and speed are always factors in the build versus buy decision, but more participants pointed to unmet needs as their main reason to build. While plenty of factors influence the decision and execution, as a rule, companies tend to build when no commercial tool fits their needs. Simulation, testing, and digital twins came up most often as areas where off-the-shelf software falls short. 

The second driver is specific to engineering software. Several companies explained how they chose to build because the application itself would encode proprietary engineering knowledge they’d rather not hand to a vendor, even one willing to add the capability to its product.  

Data privacy, control over architecture, and avoiding vendor lock-in rounded out the list. 

Participants were also candid that custom-built solutions are slower to deploy, likely cost more up front, and need an internal or outsourced team to support them for as long as they’re in use. 

4. The Cost of Scaling is a Factor Teams Often Overlook 

Lower costs were the single most common reason to buy instead of build (cited by around half of participants. Most also associated building with higher costs. 

Despite this, several organizations detailed that commercial software can become surprisingly expensive once the user count and subscription costs climb, catching budgets off guard.  

This issue is particularly poignant when expanding access to 3D engineering data, which often involves providing teams with expensive CAD licenses. The report notes that many companies don’t realize some alternatives allow access to that data without additional seats in engineering applications, and describes building around that approach as less common but “very compelling.” 

We looked at this from the other direction in Why More Manufacturers Are Expanding Access to CAD Data Beyond CAD Applications: the CAD seat is rarely the right vessel for people who only need to view, measure, or pull information from a model. 

5. AI Could Tip the Balance Toward Building 

AI came up as a potential shift in the build-buy balance far more often than Tech-Clarity expected, mostly because of what it’s doing to development itself. AI-assisted coding, or “vibe coding”, lowers the perceived cost of building custom software and expands access to “citizen developers” outside of IT into the process.  

An automotive technology strategy leader said AI-driven engineering “could change the whole ecosystem.” 

While the development process is changing, those AI-built applications still run into the same 3D data problems that make building hard today: multi-CAD assemblies, kernel formats that vendors update regularly, and PMI and metadata that neutral formats like STEP can lose in translation. The report expects AI coding to raise the importance of SDKs for data translation, visualization, and simulation, since coding agents and copilots can pull them into custom applications to bring them up to commercial grade. 

AI’s pull isn’t all in one direction. The  report cites concerns that AI tools could expose sensitive data or learn from proprietary processes, which may push some participants to build more themselves, while others expect to buy more just to keep up with demand for AI capabilities.  

The exact nature of AI effects on software development, the build versus buy debate, and engineering applications as a whole is still murky, but the scale of its impacts will be significant. 

What Comes Next for Engineering Data Applications 

If AI does make building more common, more of these applications will be written by teams without deep CAD expertise, and more of the weight will fall on the components underneath them. To learn more about some of the best building blocks for engineering software on the market, get in touch with Tech Soft 3D toolkits serving CAD data access, visualization, CAE, AI, and more 

The full qualitative report goes further than we have here, with specific examples of what manufacturers have recently chosen to build and to buy, how participants expect the balance to shift, and Tech-Clarity’s recommendations for making the call. 

For those interested in exploring what Tech Soft 3D Toolkits and end-user applications can do for you, contact us.

The full report, How Manufacturers Meet the Growing Demand for Engineering Data, is available now.