Own or Rent? How AI Changed the Maths on Vendor Lock-In
Software was expensive to build, so renting it made sense. AI has cut the cost, and for the systems that earn your money, owning an open stack is often cheaper.
A lot of businesses run their software the way you might run a shop where the till, the shelves and the customer list are rented along with the building. The till only works while you keep paying the till company. The shelves get rearranged overnight because head office pushed an update you didn't ask for. Prices go up every January "due to increased platform costs". And the day you decide to leave, you find your customer list, your product photos and ten years of sales history are written in a language only the landlord can read.
Nobody would accept that for a physical shop. Plenty of businesses accept almost exactly that for their software, and have done for years. It's called vendor lock-in, and for a long time it was the sensible choice, because building your own was so expensive. That is what has changed.
Why renting used to make sense
For most of the internet's life, custom software was a luxury. Developers were slow and expensive, so you rented. You paid Shopify or Salesforce or whoever a monthly fee and got something that mostly worked, without running an engineering team of your own.
That was a fair trade, and usually the right one. The catch was that you were renting a relationship as well as a tool, and the landlord set the terms.
Tech pull-out: what "lock-in" actually means. Lock-in is anything that makes leaving painful. Your data is in a format you can't export cleanly. Your workflows are built around features only that vendor has. Your integrations break the moment you switch. The more of the business you build around their product, the harder it is to get out.
Where renting hurts
The price is never quite the price. Platforms grow by making it more expensive to stay. Transaction fees, "premium" tiers for things that used to be included, per-seat charges that climb as you hire. The headline price gets you in; the real bill arrives once leaving has become unthinkable.
The terms change under you. You signed up for the product as it was three years ago. Since then the terms have been quietly updated, features have been retired and an API you built around has been switched off with ninety days' notice. Anyone whose business leaned on something a platform then "sunset" knows how that goes.
Your day depends on someone else's system. When the platform goes down, you go down with it. When their priorities change, your plans change too, without anyone asking you. A company whose interests only partly overlap with yours now has a hand in how your business runs.
None of it is malicious. It's what renting is. The landlord looks after the landlord.
What AI changed
AI has cut the cost of building software sharply, and that changes the sums. An experienced developer using AI tools now gets through in a day roughly what used to take the best part of a week. The slow, dull bulk of the work (the wiring, the plumbing, the endless odd cases) goes much faster than it did. Good engineers are still needed. Each one just gets a lot more done.
Once custom software is cheap enough, "just rent it, building's too dear" stops being true. For twenty years build-or-buy came out as buy. It is swinging back towards building and owning, at least for the parts of a business that bring in money or protect it.
Tech pull-out: nobody is saying build it all from scratch. You don't rebuild a payment processor in your garage. You start from open foundations, the mature, free software that thousands of businesses already run, and have developers using AI shape it to fit you. You get the head start a platform gives you, and you can leave whenever you like.
The obvious example: Shopify
Shopify is the one everyone mentions. It's a good product and a perfectly reasonable place to start a shop. It's still worth being clear about what you're renting: a fee on every sale, a shop that looks a lot like everyone else's, a stack of app subscriptions on top, and your whole catalogue living in someone else's box.
The alternative used to cost too much to bother with. Increasingly it doesn't.
- You can take WooCommerce, the free, open commerce software behind a large share of the web's shops, and own it outright, with no platform taking a cut of your sales.
- You can give it a fast shop built around your brand, one that loads quickly and looks like you instead of a template.
- You can run the whole thing on a server you control, for a fraction of platform fees that only ever go up.
Tech pull-out: a separate shop front, and why your neighbours matter. Running WooCommerce behind a separately built shop front (usually made with something like Next.js) keeps the pages customers see apart from the system that handles orders. The shop is faster, easier to change and entirely yours. There's a second reason too. On most rented platforms you share servers with thousands of other businesses, so a traffic spike next door can slow your checkout at the worst moment. On your own server the capacity is yours, and on a busy day you serve the rush.
You end up with lower running costs, a faster site, your data in your hands and next January's fees decided by you. The money that paid rent now buys the freehold.
Shops aren't the only case
The same thing happens wherever a vendor has put a wall around something that ought to be yours.
Low-code platforms like PowerApps are great for a quick prototype. Then the per-user licences add up, performance hits a ceiling, and your business logic is stuck in a tool you can't take with you. A purpose-built internal app is now cheap enough to commission. It runs faster, costs less to keep and belongs to you.
Off-the-shelf dashboards and BI tools charge per viewer and still never show the numbers the way your team thinks about them. A dashboard built on your own data answers your questions, and adding the next one is a small job instead of a sales call.
So when a subscription is really rent on something central to how you operate, it's worth asking the question out loud.
The question worth asking
The old question was whether you could afford to build it yourself. For twenty years the answer was usually no, so you rented and lived with the lock-in.
AI has changed the answer. For the things that matter most to your business, would you rather own them, or keep renting them from a company whose priorities were never quite the same as yours?
You don't have to rip everything out tomorrow, and plenty of everyday tools are fine rented. Next time a platform raises its fees, rewrites its terms or holds your own data behind an export wall, remember the trade-off has moved. An open stack you own used to be the expensive, unusual choice. More and more often it's the cheaper one.
At Revitt we help businesses decide which systems to own and which to rent, and build the ones worth owning on open foundations. If a platform is setting your terms and you're fed up with it, get in touch.