Permira's analysis suggests AI can replicate software features, eroding the value of high switching costs and recurring revenue. The new moat is whether a company owns critical data or is deeply embedded in workflows.

When evaluating AI companies, focus on customer love (gross retention) and efficient acquisition over gross margins. High margins are less critical initially, as the 99%+ decline in model input costs suggests a clear path to future profitability if the core product is sticky.

SaaS companies are trying to preserve high gross margins, but this is impossible if they want to succeed in AI, which is compute-intensive. Lower margins should be reframed as a positive signal of user adoption and AI integration, much like the successful transition from on-prem to cloud.

Unlike in traditional SaaS, low gross margins in an AI company can be a positive indicator. They often reflect high inference costs, which directly correlates with strong user engagement with core AI features. High margins might suggest the AI is not the main product driver.

AI application-layer companies are knowingly accepting negative gross margins by reselling expensive model inference. Their strategy is to first lock in users with a superior UX, then solve the cost problem later through vertical integration or cheaper models.

Application-layer AI companies can pivot rapidly with model improvements because they serve sticky end-customers. Infrastructure companies face a pickier developer audience that is more likely to churn completely to the next hot tool, making pivots riskier.

Unlike SaaS, where high gross margins are key, an AI company with very high margins likely isn't seeing significant use of its core AI features. Low margins signal that customers are actively using compute-intensive products, a positive early indicator.

Contrary to traditional software evaluation, Andreessen Horowitz now questions AI companies that present high, SaaS-like gross margins. This often indicates a critical flaw: customers are not engaging with the costly, core AI features. Low margins, in this context, can be a positive signal of genuine product usage and value delivery.

The assumption that building the most advanced AI model creates a defensible, high-margin business is collapsing. With competitors offering comparable performance at lower prices, the sustainable advantage shifts from owning the best intelligence to how that intelligence is productized and integrated.