Anish Acharya: Is SaaS Dead in a World of AI?

What It Covers

Anish Acharya, general partner at a16z, challenges the narrative that AI will kill SaaS companies. He argues IT spend represents only 8-12% of enterprise budgets, making wholesale replacement inefficient. The real transformation involves dramatically reduced switching costs through coding agents, creating more competition and better products across the ecosystem.

Key Questions Answered

• •SaaS disruption overstated: IT spend constitutes 8-12% of enterprise budgets. Using AI models to rebuild payroll, ERP, or CRM systems wastes innovation potential when companies could apply AI to their core business advantages or optimize the remaining 90% of non-software spend. 75% of public SaaS companies raised prices 8-12% since ChatGPT launched, with some increasing 25% or more, indicating strong product-market fit rather than competitive pressure.
• •Switching costs transformation: Coding agents dramatically reduce the complexity, time, and risk of transitioning between SaaS providers like SAP and Oracle. This shift converts hostages into customers, forcing companies to compete on product quality rather than lock-in. The result creates positive incentives across the ecosystem through increased competition, better products, and accelerated innovation without wholesale market disruption.
• •Apps layer aggregation value: Foundation models specialize across use cases - Gemini excels at front-end coding while Codex handles back-end, Midjourney creates aesthetically opinionated imagery while Ideogram serves graphic designers. Apps companies like Cursor provide valuable orchestration layers, allowing users to access multiple specialized models through single interfaces rather than switching between command-line interfaces for different tasks.
• •Margin structure evolution: AI companies experience indirect subsidization through zero or negative gross margin credits for user trials, creating healthy conversion into high-paying power users. This differs from 2021's empty calories of Google and Facebook ad spend. Evaluate month-two retention as the true baseline since month-one includes free trial tourists. Target 50% month-twelve retention minimum, with 70% indicating strong performance.
• •Consumer product pricing ceiling shattered: Pre-AI consumer products maxed out at twenty to twenty-five dollars monthly, exemplified by Spotify's premium family plan. AI products command 10x higher prices - Grok Heavy costs three hundred dollars monthly, ChatGPT two hundred dollars, Gemini Ultra two hundred fifty dollars. Power users pay these premium subscription rates plus consumption revenue, justifying higher customer acquisition costs.
• •Geographic network effects persist: San Francisco maintains the original network effect for technology builders. Secrets get whispered down shadowy hallways, and choosing to relocate everything to San Francisco demonstrates singular focus and uncompromising ambition. Tel Aviv succeeds as an alternative because its ten million population forces immediate international expansion, unlike London's sixty million people enabling dangerous domestic market focus.