SaaS is not dead. What is dying is the 'seat-based pricing model'
"Will SaaS die in the AI era?" This question is actually a sign that both "technology" and "business models" are being shaken simultaneously. In this episode, we discuss how ① the decline in software development costs due to AI, ② the collapse of entry barriers and the acceleration of feature copying, ③ the decline in pricing power, and ④ the shift from seat-based to value/usage-based pricing will reshape the future of SaaS. In conclusion, SaaS itself will not disappear. However, there is little guarantee that "today's winners will continue to win"—this is the core of the discussion.
1. "SaaS will not die," but the "same faces" may not remain
The speakers are quite candid at the beginning. Even when faced with a binary question, they say, "No. Not likely," but at the same time, they emphasize: "I don't think the companies delivering that value in five years will be the same as today. Some companies will survive, but significant consolidation will occur."
What is important here is not the "form of SaaS," but the fact that the sustainability of SaaS companies' revenue—the certainty of future cash flows—is being shaken. The speaker states that "uncertainty has risen dramatically" and cites the fact that AI is rapidly lowering the "ability to build software" as the background.
2. The essence of the "software economy" that AI is changing
2-1. The collapse of entry barriers: "Delivering" is harder than "building"
This is the crux of the discussion. AI models are advancing and becoming cheaper, making it easier to "build" software. As a result, the advantages that SaaS traditionally held—feature development capability and development speed—are relatively diminishing.
However, at the same time, the speaker also says, "Software does not automatically get adopted or established in companies." That is precisely why "implementation" and "integration into operations" that increase the productivity of intellectual labor within companies still retain value. In short, in a world where you can "build," the difference shifts to "being able to operate, being used, and producing results."
2-2. Feature copying and the "agility" game
The faster development speed becomes with AI, the more intense the competition will be. The statement is symbolic: "Features are copied immediately. The team that is agile and can ship fast wins."
On the other hand, the common mindset in existing SaaS—"Customers keep renewing, so we can use that revenue to fund next year's S&M and R&D"—is suggested to be dangerous. A stable model based on renewals is difficult to use as a "defensive" strategy in the AI era.
3. Is the valuation collapse "fear" or "rationality"?
In the discussion, the decline in SaaS multiples is not dismissed as "market fear" but is viewed as a rational re-rating. An example that comes up is the discussion of "payback period" using the Bessemer Cloud Index for SaaS stocks.
Roughly speaking, the number of years it took for investors to hold a company and recoup their investment through free cash flow was extremely long at its peak. Even if that shortens, the speaker still asks: "Isn't 12 years still too long? In a world of increased uncertainty, do you believe you can recoup over two economic cycles?" This is from an investor's perspective, but when translated to the company side, it means that "future pricing power" and the "premise of recurring billing" have been shaken.
4. From "seat-based" to "value-based" pricing: The biggest hurdle is the business model shift
As AI agents become widespread, usage will skyrocket. However, if you stick to seat-based pricing, as stated: "If you charge for an agent as one seat, the agent works at 'a million miles per hour,' but you cannot monetize the added value." In other words, SaaS companies have no choice but to shift to pricing linked to usage, results, and engagement.
This is trickier than the "technology" itself. This is because it involves everything: existing customer contract structures, sales organization KPIs, product design, and revenue recognition. Just as the shift from "perpetual license to subscription" created "turbulent quarters" in the past, the view is that the next rough wave will be the shift from "subscription to value-based pricing."
5. The winning path is "companies that turn AI tokens into value"
The most concrete success example discussed is the type of company that organizes corporate data and combines it with LLMs to transform it into business value. Symbolically shown is the idea of "contextualizing internal corporate data with an 'ontology,' making it possible to build agents on top of that, and selling productivity."
The point here is not to treat AI as "feature addition (adding a little LLM)," but whether you can design it as
AI = a device that converts new raw materials (digital intelligence) into value within the company's unique context.
6. Will existing major players sink like the newspaper industry? — 'Core products' vs. 'Bundleable features'
The metaphor at the end is interesting. Will existing SaaS decline like the 'newspaper industry,' or will it adapt like banking and retail? The decision-making axis that emerges here is,
'Is this a core product? Or is it something that can be bundled as a feature into a larger suite?'
For example, massive suites (like Microsoft) may survive by embedding AI into their core functions. On the other hand, vertical SaaS specialized for specific tasks is easily bundled, facing stronger pressure—this nuance is close to the conclusion of the discussion.
Conclusion: SaaS in the AI era is determined not by its 'form' but by its 'monetization design'
What this episode shows is the reality that the future of SaaS is not a binary choice of 'die or live,' but is being reorganized along multiple axes: 1) lower barriers to entry, 2) lower pricing power, 3) redesign of billing models, and 4) the ability to contextualize AI and turn it into value.
'SaaS will not die.' However, if a company cannot change from a 'company that sells seats' to a 'company that can measure value and charge for it,' its lifespan as a SaaS company will be short—that is the cruel, yet simultaneously opportunistic, conclusion of the AI era.
