My boss asked me this recently, and it's a question we talk about a lot at On Time Edge: do you still need an industrial platform? AI coding tools now make it possible to build a working manufacturing application from scratch in days, using free, open-source databases and app servers. So why pay for a platform when you could build it yourself?
It's the classic build vs. buy question with a new twist. My answer is a reduction to absurdity.
Yes, you can vibe code a solution, describing what you want to an AI tool and letting it write the code, on top of open-source components. You still have to build everything around it: users and roles, logging, backups, database tuning, and connection management. None of that is the application. All of it has to work.
You've spent days or weeks building infrastructure just so you can get started. Then you have to maintain, patch, back up, and secure that infrastructure for as long as the application lives. That ongoing burden is how technical debt builds up in manufacturing.
And where are you going to run it? You need redundant servers separated by distance in case of a fire or other disaster. A few years ago, one of our customers needed 14 servers to cover build, testing, production, and backups. The hardware alone cost over $200,000, before anyone was paid to run it.
Now for the logical argument straight out of Harry Potter. REDUCTIO AD ABSURDUM! I'm not sure I cast that spell correctly, but let's see where it leads.
If writing your own code is better, why stop at the application? Vibe code the database and app server so they are tuned to your exact needs. Then the operating system. Then the firmware. Reinvent everything! At that point you are carving your own wand, and you start by planting the tree.
Nobody does this, because everyone draws a line somewhere. Below the line, you trust proven infrastructure. Above it, you build what makes you different. The only real question is where the line belongs.
An industrial platform moves the line up. It takes care of the boring parts of building an application and lets you focus on the value add: solving the manufacturing problem.
A manufacturer shouldn't need to worry about connection pools, log file management, and man-in-the-middle attacks when the real problem is getting more capacity out of a $2 million machine.
The same goes for manufacturing systems integration. Most manufacturers already have systems they are happy with, and connecting them takes the same plumbing every time: security, logging, error handling, and monitoring. As we've written before, connectivity alone doesn't deliver performance, but you can't get performance without it.
Enterprise industrial platforms often cost six figures per site per year. That investment pays off when it is spread across many sites and many problems. For a single site with a single problem, it is hard to justify.
You are trading some independence for speed and governance, and that trade improves with every problem you add.
I know an industrial platform company that built its CRM on its own platform. They are eating their own dog food, and I think that is genius. It works because the CRM is one of many applications sharing the same foundation.
If a CRM were the only thing we planned to build, an off-the-shelf subscription would win on cost. A CRM is a problem everyone has, and somebody has already solved it.
Manufacturing is harder. Your MES, WMS, QMS, CMMS, and historian have to work together, whether you bought them or built them. That is where an industrial platform shines.
Platforms are no longer one size and one price. Different levels offer different capabilities, and lower-cost options are emerging.
Enterprise industrial platforms carry the full weight of security, infrastructure, and governance. The cost makes sense when it is spread across many sites and many problems.
Lower-cost industrial platforms offer fewer capabilities at a fraction of the price. They still remove the boring parts, and one or two problems can justify them.
My rule of thumb:
Before your team writes a line of code or signs a platform contract, answer these:
So start with the problem you're trying to solve. If it is a problem everyone has, buy the finished product. If it is unique to how you make things, build it.
Either way, those pieces have to share data, and nobody sells the connections between your particular systems and machines. Building and integrating are where an industrial platform earns its keep. Then pick the level.
Once everything runs through one place, you also end up with one place for your manufacturing data, and that is a post for another day.
Weighing whether to build, buy, or adopt an industrial platform? On Time Edge helps manufacturers choose the right platform level and put it to work. Schedule a consultation and we'll talk it through. No owl required.