Contrary to the belief that its huge user base is a key asset, ChatGPT's free tier is described as a massive liability. The cost of running millions of GPUs for non-paying users is enormous, and monetization attempts like ads risk driving users to competitors in a market with low switching costs.

For many AI companies, the primary growth lever is no longer advertising spend but offering free trials and credits. This makes their CAC directly tied to expensive compute resources, elevating the financial impact of trial abuse from a nuisance to a major business risk.

Because compute theft occurs before a transaction, fraud risk for AI companies starts at sign-up, not checkout. In response, Stripe has adapted its Radar product to be integrated at the beginning of the user lifecycle, assessing risk before any costly credits are granted.

Unlike traditional SaaS, achieving product-market fit in AI is not enough for survival. The high and variable costs of model inference mean that as usage grows, companies can scale directly into unprofitability. This makes developing cost-efficient infrastructure a critical moat and survival strategy, not just an optimization.

The ARR/SaaS model, built on predictable human usage, is failing. AI agents can consume resources worth thousands of dollars for a low subscription fee, breaking the unit economics. This forces a shift to metered, consumption-based pricing similar to utilities like electricity.

The current subsidized AI subscription model is unsustainable. The inevitable shift to pay-per-token pricing will expose the true cost of inference. For tasks like coding, where AI can "hallucinate" and burn tokens in loops, this creates unpredictable and potentially exorbitant costs, akin to gambling.

Unlike traditional software's zero marginal costs, AI-powered apps incur significant inference expenses that scale with users. One founder estimated needing $25M just for 100k monthly actives, challenging the classic VC model for consumer startups.

Unlike SaaS where marginal costs are near-zero, AI companies face high inference costs. Abuse of free trials or refunds by non-paying users ("friendly fraud") directly threatens unit economics, forcing some founders to choke growth by disabling trials altogether to survive.

Mature B2B SaaS companies, after achieving profitability, now face a new crisis: funding expensive AI agents to stay competitive. They must spend millions on inference to match venture-backed startups, creating a dilemma that could lead to their demise despite having a solid underlying business.