SaaS Resilience and AI Application Layer Strategy
An analysis of why the SaaS apocalypse narrative is flawed, focusing on the strategic value of the application layer, shifting switching costs, and new margin dynamics in the AI era. This brief outlines actionable frameworks for investors and founders navigating the transition from seat-based to outcome-based pricing.
The SaaS Apocalypse Is a Misdiagnosis
The prevailing narrative that AI will render traditional SaaS obsolete is fundamentally flawed. Enterprise IT spend represents only 8-12% of total budgets, meaning that even if AI completely replaced core software like ERP or payroll, the savings would be marginal compared to the potential gains from applying AI to the remaining 90% of operational spend. The strategic imperative is not to rebuild legacy systems but to leverage AI as an "innovation bazooka" to extend core business advantages and optimize non-software expenditures. This perspective corrects the market's overly negative bias against established software providers, many of which are successfully raising prices and maintaining strong product-market fit.
The Rise of the Application Layer
A critical shift is occurring in the AI stack. Foundation models are increasingly becoming substitutes for one another, with 80% of capabilities overlapping. This commoditization creates a massive opportunity for the application layer, which acts as an aggregation and orchestration layer. Companies that build rich, multi-model interfaces to solve specific, fragmented user needs—such as coding tools that combine front-end and back-end models, or creative suites that blend aesthetic styles—capture significant value. The application layer is underhyped because it solves the "last mile" problem of model specialization, providing a unified user experience that raw models cannot.
Shifting Economics and Defensibility
The economics of AI-native businesses differ from traditional SaaS. Blended margins may appear lower due to inference costs, but this is often a healthy distortion driven by high-volume, low-cost acquisition of power users. These users pay premium prices for advanced capabilities, creating a durable margin profile for converted customers. Furthermore, defensibility is evolving. While network effects remain the gold standard, proprietary live data has emerged as a powerful moat. Companies that integrate real-time, proprietary data streams can leverage commodity models to outperform competitors, creating a sustainable competitive advantage that is difficult to replicate.
Strategic Implications for Investors and Founders
Investors must recalibrate their valuation models to account for these new dynamics. The focus should shift from seat-based growth to outcome-based value and the durability of power user cohorts. Founders should prioritize building in new, native AI categories where incumbents have no historical advantage, rather than attempting to displace entrenched players in existing workflows. The future belongs to companies that can effectively orchestrate the multi-model landscape and leverage live data to create unique, defensible products.
Key insights
-
The SaaS market is not dying but being re-evaluated; IT spend is a small fraction of enterprise budgets, making AI a tool for broader operational optimization rather than just software replacement. Incumbents are raising prices, indicating strong product-market fit despite AI hype.
Impact: Investors should avoid panic-selling SaaS assets and instead look for companies leveraging AI to expand into non-IT spend categories.
-
Coding agents are drastically reducing the cost and risk of switching between SaaS providers. This eliminates the "hostage" dynamic, increasing competition and forcing providers to improve product quality and service to retain customers.
Impact: SaaS companies must focus on genuine value delivery and ease of integration, as switching costs are no longer a reliable retention mechanism.
-
Foundation models are becoming commoditized substitutes, creating a high-value opportunity for application-layer companies that orchestrate multiple models. These apps provide the necessary context, UI, and workflow integration that raw models lack.
Impact: Startups should focus on building orchestration layers and specialized applications rather than competing on model performance, capturing value through user experience and integration.
-
AI-native companies are seeing a shift in margin dynamics. While blended margins may be lower due to inference costs, the acquisition of high-LTV power users who pay premium prices creates a sustainable and profitable business model.
Impact: Investors should evaluate AI companies based on the margin profile of their power user cohort rather than blended average margins, recognizing the value of high-consumption users.
-
Defensibility in the AI era is shifting towards proprietary live data. Companies that integrate real-time, unique data streams can leverage commodity models to achieve superior results, creating a moat that is difficult for competitors to replicate.
Impact: Founders should prioritize building proprietary data pipelines and live data integration as a core competitive advantage, rather than relying solely on model access or brand.
Action items
-
Re-evaluate SaaS portfolios by focusing on companies that are raising prices and expanding into non-IT spend categories. Avoid companies that rely solely on seat-based growth without clear AI-driven value expansion.
Impact: Identifies resilient SaaS companies that are successfully adapting to the AI era and capturing broader enterprise budgets.
-
For SaaS providers, invest in reducing integration complexity and improving the ease of switching. Recognize that customers are no longer locked in and will switch if the value proposition is not compelling.
Impact: Enhances customer retention by addressing the new reality of low switching costs and fostering a competitive ecosystem.
-
Build application-layer products that orchestrate multiple foundation models. Focus on creating unified user experiences that solve specific, fragmented user needs by leveraging the strengths of different models.
Impact: Captures value in the application layer by providing a superior user experience and solving the multi-model orchestration problem.
-
Adjust financial models for AI-native companies to focus on the margin profile of power users. Recognize that lower blended margins can be sustainable if driven by high-LTV, high-consumption users.
Impact: Provides a more accurate valuation framework for AI companies, avoiding undervaluation due to misleading blended margin metrics.
-
Prioritize the development of proprietary live data streams as a core moat. Integrate real-time data into AI workflows to create unique, defensible products that outperform competitors using commodity models.
Impact: Creates a sustainable competitive advantage by leveraging unique data assets, making the product difficult to replicate by competitors.
Quotes
“I think software is completely oversold. I think it's a silly story.”
“The cost of transitioning from one SaaS provider to another going dramatically down.”
“There's a lot of value in having an aggregation layer, and that is the apps company.”