Today is Wednesday, February 19th. No long jogging this morning — just me at the computer thinking about what people are calling the “SaaS apocalypse.”
It sounds dramatic. But there’s something real happening underneath the headline.
Let’s break it down.
What Is the “SaaS Apocalypse”?
The term refers to the sharp market correction hitting many SaaS companies. Valuations are dropping. Growth multiples are shrinking. Investors are asking harder questions.
Companies that once charged premium subscription fees are suddenly facing a new reality:
Engineers can now build 80% of your product in a weekend with AI.
Think about platforms like:
- Adobe
- Hosting and publishing platforms
- Security SaaS vendors (SAST, DAST, IAST scanners)
- Collaboration and productivity tools
These categories always had competitors. For example, GIMP has existed for years as a free alternative to Photoshop.
But historically, the barrier wasn’t availability — it was complexity and learning curve.
That’s what’s changed.
Why Is This Time Different?
There have always been cheaper or open-source alternatives. So why now?
Here’s my take:
1. Generative Coding Removes the Learning Curve
Before AI, replacing a SaaS tool required:
- Deep expertise
- Time investment
- Maintenance burden
- Infrastructure overhead
Now?
You can prompt your way into a working internal tool.
You don’t need to master the entire stack. AI scaffolds it for you. It writes the glue code. It builds the API layer. It even generates UI.
The “build vs. buy” equation just shifted dramatically.
2. The Market Is Regulating Toward True Value
For years, SaaS pricing often floated far above actual usage value.
Let’s use security as an example (a space I know well).
A typical enterprise security scanner might cost:
- $100,000
- $500,000
- $1,000,000+ annually
Then vendors shift to:
“Per seat pricing.”
$20–$50 per engineer per month sounds reasonable — until you multiply it across:
- 1,000 engineers
- 10,000 engineers
- 30,000 engineers
Suddenly you’re back in the hundreds of thousands or millions per year.
It’s the same cost — just disguised differently.
And here’s the real kicker:
Most companies only use ~20% of the product’s capabilities.
The classic 80/20 rule shows up again:
- 20% of features drive 80% of value.
- But companies rarely reach even that 80%.
So you’re paying enterprise pricing for partial usage.
AI exposes that mismatch.
A Personal Example
I pay $20/month for ChatGPT.
My primary use case?
When I jog, I dictate blog ideas into it. I literally run and brainstorm.
Recently, usage limits were introduced on certain features. That forced a realization:
If I hit constraints, I could:
- Stand up my own hosted LLM
- Remove arbitrary limits
- Tailor the workflow to my needs
- Own the pipeline end-to-end
Yes, I’d absorb infrastructure overhead.
But I’d get exactly what I want — not a generalized product with usage caps.
That’s the shift.
The Real Threat to SaaS
It’s not that SaaS is dying.
It’s that customization is now cheap.
If I can generate:
- A tailored vulnerability dashboard
- A lightweight scanner wrapper
- A content workflow engine
- A document processing pipeline
And get 80% of the value…
Why would I pay 100% pricing?
Especially if I avoid:
- Vendor lock-in
- Per-seat scaling traps
- Feature bloat
- Multi-year contracts
The economics are shifting from:
“Buy the platform.”
To:
“Generate the feature.”
What This Means Going Forward
Here’s where I think this lands:
1. Pricing Must Compress
Companies heavily leveraged with debt will struggle to lower pricing.
But the market is pushing hard toward:
- Modular pricing
- Usage-based transparency
- True value alignment
The days of bloated enterprise contracts may shrink.
2. SaaS Must Move Up the Stack
If your product is:
- Just an interface over data
- A thin orchestration layer
- A workflow wrapper
AI can replicate it.
SaaS companies will need to:
- Offer deep integrations
- Provide regulatory coverage
- Deliver operational guarantees
- Own risk and compliance layers
The higher you are in the “risk stack,” the safer you are.
3. Engineers Have Leverage Again
For the first time in a long time, engineers can say:
“If you want $1M/year, I’ll just build it.”
Not because they want to rebuild everything.
But because they now can.
And that negotiation dynamic alone forces correction.
My Take
I don’t think SaaS will disappear.
I think it gets normalized.
The market is rediscovering what something should cost.
AI didn’t create competition.
It removed friction.
And when friction disappears, pricing has to align with real value.
I’ll still pay for services that:
- Save me meaningful time
- Remove operational risk
- Deliver something non-trivial to reproduce
But if a feature is reproducible through generated code and light infrastructure?
The premium better make sense.
Because now, we can build.
That’s my take on the so-called SaaS apocalypse.
Not collapse.
Correction.