edited on: 9/10/2025
Is there a future for software subscriptions?
Some thoughts on the future of SaaS - should every business just buy an app and self-host it?
This has been on my mind for a while - with super awesome deployment tools like Coolify
and vibe-coding being super efficient, what's the point of paying for SaaS?
It took me less than a week, but I replaced my entire subscription stack (- content) with self-built
and hosted apps. Whether it's notion, airtable or app builders like Softr
and webflow - it seems pretty pointless using them now. So, I wonder are they going to exist?
Let's face it - for most users these apps now have some kind of AI integration,
but it's still a steep learning curve for new users - not to mention pointless and time-consuming.
So let's look at app running costs, if you have 20 users constantly hammering something in to your
unloved work app, with say 500 read/writes a minute - a $6 server with Hetzner could host it easily - and
couple it with a deploy tool like Coolify, your org probably won't need any technical expertise.
The same app, let's say built on webflow/softr would cost you $49/month in hosting + $5,000 in
design costs, and it would still be much slower than an unloved app. So I almost come to my point -
what if you had every app + integrations for a fraction of the cost, and you would own your data? and no, you wouldn't need to code.
Vibe coding is probably the single biggest reason why this is possible - most design patterns are
industry standard now, as is the code behind it. For sub 100 user apps, you don't need cluster management or advanced k8s configurations, app just works out of the box. 10 instances of Postgres for an app used by 1,000 users - over over kill!
here comes my thesis (finally) - so as cost of creation nosedives, there will be a shift away from
So the world by this time next year will probably have hundreds of thousands of local one person niche software agencies
vibe coding client apps to spec and selling locally, no more cross-border large SaaS selling.
That's pretty much pre-2010, are we heading back to where it started? It's different this time though,
infrastructure is widely available and deployment tools are awesome!
So like tell us what you need?
and vibe-coding being super efficient, what's the point of paying for SaaS?
It took me less than a week, but I replaced my entire subscription stack (- content) with self-built
and hosted apps. Whether it's notion, airtable or app builders like Softr
and webflow - it seems pretty pointless using them now. So, I wonder are they going to exist?
Let's face it - for most users these apps now have some kind of AI integration,
but it's still a steep learning curve for new users - not to mention pointless and time-consuming.
So let's look at app running costs, if you have 20 users constantly hammering something in to your
unloved work app, with say 500 read/writes a minute - a $6 server with Hetzner could host it easily - and
couple it with a deploy tool like Coolify, your org probably won't need any technical expertise.
The same app, let's say built on webflow/softr would cost you $49/month in hosting + $5,000 in
design costs, and it would still be much slower than an unloved app. So I almost come to my point -
what if you had every app + integrations for a fraction of the cost, and you would own your data? and no, you wouldn't need to code.
Vibe coding is probably the single biggest reason why this is possible - most design patterns are
industry standard now, as is the code behind it. For sub 100 user apps, you don't need cluster management or advanced k8s configurations, app just works out of the box. 10 instances of Postgres for an app used by 1,000 users - over over kill!
here comes my thesis (finally) - so as cost of creation nosedives, there will be a shift away from
large SaaS providers, to niche smaller apps built by industry experts and product engineers, that
mimic parts of existing SaaS design patterns but actually solve an entire set of business problems.
So the world by this time next year will probably have hundreds of thousands of local one person niche software agencies
vibe coding client apps to spec and selling locally, no more cross-border large SaaS selling.
That's pretty much pre-2010, are we heading back to where it started? It's different this time though,
infrastructure is widely available and deployment tools are awesome!
So like tell us what you need?