3 runway planning frameworks SaaS leaders are using in 2025
Knowledge
September 2, 2025
.avif)
You’ve seen it:
SaaS is dead.
It’s a bold claim circulating in LinkedIn threads, podcast panels, and market commentaries. Often it’s based on a simple causal chain:
Therefore, SaaS is finished.
This framing is too simplistic.
Yes, software valuations have reset, and the stock prices of many large platforms are lower year-over-year. But a reset is repricing, not extinction. And there’s solid data showing SaaS continues to expand globally. The global SaaS market is forecast to grow strongly in 2026, with estimates ranging from ~$408B to over $430B and projected to continue expanding robustly through the decade.
What we’re seeing isn’t the death of SaaS, it’s a filtering and evolution of the model.
One of the biggest analytical missteps in the “SaaS is dead” narrative is averaging across the whole category.
SaaS today includes:
AI doesn’t impact all of these equally. Treating them as a single asset class misses the nuance. What matters is which segments have deep embedded value, structural moats, and mission-critical workflows, and which do not.
AI doesn’t disrupt software in just one way; it does so through a few different paths, each affecting different kinds of products differently:
Disruption follows incentives.
If you can take 5–10% of revenue from a huge enterprise platform, that’s worth billions and that’s exactly the kind of target deep-pocketed teams chase.
But trying to win 5–10% of a €15–€20M vertical SaaS tool? That’s a very different economic equation. The upside is too small to attract serious competition.
So big horizontal platforms are the natural focus for aggressive disruption. Many smaller, specialised SaaS businesses simply aren’t big enough to matter to would-be competitors, and that protects them.
AI makes it easy to sketch prototypes, automations, or simple utilities. That’s real.
But turning a sketch into production-grade software that runs critical operations requires more than code generation:
A small startup might build a basic UI in a weekend with an AI tool. But rebuilding a logistics, finance, legal, or compliance system internally? That takes real investment and changes how the company runs.
Most organisations don’t want to become software builders. They want to run their business.
General AI tools like copilots, large language models, and automation assistants excel at:
Those tools can replace simple utilities or single features.
But they don’t replace software that’s deeply woven into how an organisation operates systems tied to compliance, multi-stage workflows, domain expertise, or years of accumulated data.
In practice, AI more often augments entrenched SaaS making it smarter, faster, and more capable rather than simply displacing it.
Some of the softness you see in valuations and credit markets isn’t because SaaS as a model is broken, it's because of when and how certain companies were funded.
During the 2020-2022 capital boom, a lot of businesses were financed at very high multiples with aggressive growth expectations baked in. Those underwriting assumptions simply didn’t align with the reality of 2023-2026. What we’re seeing now is a correction of specific vintage risk, not evidence that SaaS itself is structurally flawed.
One of the clearest positive signals in the data is that AI is genuinely reducing the cost and time of building and iterating software.
AI-assisted development tools are already boosting productivity faster feature releases, shorter cycle times, and more experimentation with less friction. Because engineering is often the biggest cost item in a SaaS business, that improvement shows up as:
Think of this like what cloud did to infrastructure: it commoditised the base layer, and shifted value up the stack. Today’s value sits in workflow depth, data ownership, and customer outcomes, not just writing code.
That’s why AI isn’t a threat, it’s a competitive tailwind for many well-positioned SaaS companies, especially those with strong product-market fit.
Code is just one element of how value is created in SaaS. The things that really make companies resilient are harder to replicate:
AI may make writing code cheaper, but it doesn’t erase these kinds of embedded value. In fact, in many cases, it amplifies them because systems that are deeply integrated become even more mission-critical.
AI won’t impact every SaaS product the same way. The companies that tend to thrive have one or more of these strengths:
Products with these traits don’t get easily displaced just because code is cheaper to write. AI can make teams more efficient but it doesn’t take away deep operational value. That’s where real resilience lives.
We’re particularly bullish on vertical SaaS SMEs because they’re often set up to benefit from AI’s productivity gains:
Smaller, agile organisations can adopt AI more quickly than large enterprises weighed down by legacy systems. In this environment, speed compounds advantage more than size does, a core benefit for growth-stage SaaS.
If AI were truly killing SaaS, we would expect to see it in fundamental metrics like:
What we actually see is stability in many segments. Deeply embedded suites continue to expand and retain customers. Weak products churn first and that’s expected. That pattern isn’t structural collapse, it's filtering, where stronger products continue to thrive.
The SaaS companies best positioned for long-term success share a few consistent traits:
AI doesn’t kill SaaS.
AI raises the bar.
Cloud didn’t kill software.
Mobile didn’t kill SaaS.
AI will reshape expectations, and in doing so, it privileges execution over surface narratives.
Strong models get stronger.
Weak models get exposed faster.
That’s filtering, not extinction.
It’s entering a more disciplined, competitive, and mature phase.
AI is not the end of SaaS.
It’s a sorting mechanism and in sorting environments, structural strength wins.
.avif)