Harry Stebbings interviews Anisha Charya, General Partner at Andreessen Horowitz, discussing the future of enterprise software, AI's impact on SaaS markets, and where startups versus incumbents will win in the current technology cycle.

The conversation challenges the popular narrative that "AI is going to vibe code everything" and incumbents are doomed, with Charya arguing this misses the bigger picture of where AI should be deployed for maximum impact.

They explore topics including the overhyped SaaS apocalypse, why switching costs are decreasing, the underappreciated value in the applications layer, and how defensive moats are evolving in an AI-native world.

The SaaS Apocalypse Narrative is Wrong

Software represents only 8-12% of enterprise budgets, making the "vibe code everything" narrative misguided when AI could target the other 90% of spend instead.

75% of public SaaS companies have raised prices since ChatGPT's release, with mean increases of 8-12% and many raising 25% or more - indicating strength, not weakness.

"Price is a measure of product market fit. If you have enormous competitive pressure, you are not raising prices. You're typically cutting prices" - Anisha

Capable incumbents like ServiceNow "just went public and they raised guidance" - they're not the legacy IBM-style companies many assume.

Switching Costs Are Dramatically Decreasing

"Some companies have hostages, not customers" - traditional enterprise software created multi-year, high-risk switching processes that rarely succeeded.

Coding agents are making transitions between providers like SAP to Oracle "dramatically lower" in complexity, speed, and risk.

This creates "decreased switching costs, more customers, less hostages, which is a positive incentive for the entire ecosystem."

The result will be more competition, better products, and increased innovation across the enterprise software landscape.

Applications Layer Creates Underappreciated Value

Foundation models are becoming substitutes for 80% of use cases but specialists for the valuable 20%, creating aggregation opportunities.

Multi-model orchestration provides value - "if you're vibe coding your project, you probably want to use both and you don't want to switch between two CLIs all the time."

Creative tools show specialization: Midjourney and CRIA create "aesthetically opinionated" imagery while Ideogram serves graphic designers with intentionally unopinionated output.

Apps companies provide access to both specialized and general models through single interfaces, similar to AWS/Google Cloud oligopoly rather than Uber/Lyft pure substitutes.

Consumer AI Breaks Traditional Pricing Ceilings

Traditional consumer products had a "price ceiling of $20 to $25 a month" but AI products command 10x higher prices plus consumption revenue.

Grok Heavy costs $300/month, ChatGPT $200/month, Gemini Ultra $250/month - showing dramatically higher willingness to pay for power users.

"Power users are so much more powerful than they ever have been" - they're paying incredibly high subscription rates plus consumption revenue on top.

This makes customer acquisition costs for power users "very wisely invested" with much better unit economics than traditional consumer models.

Startups Win in 'Weird' Categories Big Tech Avoids

"I think weird wins" - AI enables products touching disagreement, persuasion, and sexuality that big tech committees explicitly avoid.

Companionship products like Replica have been "well received by customers and a little uncomfortable for the labs to build."

Contextual companions could provide valuable services like AI Minecraft partners that "model pro-social behaviors and is still cool and chill."

These products help people "be more self-reflective and explore aspects of themselves and human relationships" they don't have other outlets for.

Venture Strategy in the AI Era

"I don't think we're allowed to believe in luck at Andreessen. We have to see 100% of the deals in our domain and that we win 100% of the deals that we go after."

Series A is the optimal investment stage providing "information provided slash entry ownership and price" compared to seed uncertainty or later-stage competition.

"Very elastic on price" below $100M valuations but "not very elastic on ownership because that is our whole model."

Early AI leaders from 2023-2024 like Harvey, Gamma, and Cursor have "maintained their lead" unlike previous product cycles where early winners were displaced.