Hey there,
It’s Robin, your CFD friend & this week I’m asking – “what if doing CFD was free?” – a little thought experiment to noodle on while cutting the grass this afternoon.
It all stems from the confluence of three things; an article on freemium for CAD, the book Free by Chris Anderson & an email exchange with a couple of your fellow subscribers.
The book was the main driver. It’s an exploration of how markets, especially digital ones, have a tendency to move towards zero price to users. It digs into the dynamics of this, with examples from all sorts of industries, & suggests an interesting thought experiment. What would happen to you (or your business) if the cost of doing CFD went to zero?
Too cheap to meter
The book hinges on what happens when something becomes “too cheap to meter.” It’s based on an idea from the 50’s that electricity would soon become too cheap to meter (thanks to nuclear fusion). We’re still waiting on that one. But the idea found a better home in computing, where processing power, storage & memory costs have all been falling for decades. Therefore, computing should eventually become too cheap to meter.
Extending that thought, what happens when the cost of compute, memory & storage drops to the point that a useful CFD simulation can be completed at negligible marginal cost?
If you’re buying licences &/or compute today, then you might feel like this is a long way off.
If you’re selling licenses &/or compute today, then you’ll hope this a long way off 😶 But is it?
“Useful & negligible” are relative concepts, varying from project to project or from application to application. In my case, I’ve actually breached the “useful at negligible marginal cost” barrier in a couple of recent projects 🤫
That could be unusual, I don’t know. But, if it’s the general direction of travel, then working out how to get to “free” and how to make it work commercially, are about to become super-important.
Picture this
I doubt the big players will make the first moves toward zero, but here’s an idea for you to chew on, in case you’re thinking “this’ll never happen in CFD” – suspend disbelief for a moment…
Imagine Microsoft bought Ansys, then decided to wave the license fee for anyone running on a Windows machine or on Azure (Microsoft’s cloud). Alternatively, they could bundle unlimited usage on Azure into your license renewal?
Would either of those offers be enough to persuade you to switch to “Microsoft Fluent” &/or to Azure? Would free licenses or free compute make you look twice at your next license renewal or hardware refresh?
What if you’re using an open-source code, would it turn your head?
Or maybe not…
Perhaps the march to zero, as seen in other markets, doesn’t apply to CFD. We’re a greedy bunch. As soon as more processing power, storage or memory becomes available, we use it all. Our costs tend to stay the same, we just use more of everything. There’s no real concept of “good enough” which could be offered for free.
Many CFD users also have an affinity with their code of choice that goes way beyond the financial. I mentioned this to one of your fellow readers & we came to the conclusion that switching codes would be like switching your football team (or any other deep-routed, slightly-irrational allegiance). Your choice of CFD code says something about who you are & how you do your work. Same goes for the other tools & hardware you use, or choose not to use.
Too deep for these emails, back to the main question…
Will the cost of a useful CFD simulation ever fall far enough to become negligible for you? If so, what will happen to your usage, or sales, or to the industry in general. Who’s well placed to take advantage of it if it happens? Who’s well placed to make it happen?
Genuinely interested in your thoughts on this & on these emails in general. Drop me a note anytime, I welcome it.
Until next time, CFD safely,
