AI companies built to fill feature gaps on top of foundation models are at high risk. As core models rapidly improve, they often absorb these adjacent features, disintermediating the "wrapper" companies. Their early-adopter customers are also the quickest to switch to better tools.

In today's market, achieving massive growth is seen as the hardest problem to solve. Investors are comfortable backing companies with initially poor retention or margins, like early ChatGPT, as long as they demonstrate hypergrowth. The belief is that growth is paramount, and other metrics can be optimized over time.

Contrary to assumptions about user stickiness, consumers of AI models will quickly switch to a better-performing or cheaper alternative. The 22% drop in ChatGPT usage after new Gemini models were released demonstrates that brand loyalty is low when model performance is the key value proposition.

Early-stage AI startups should resist spending heavily on fine-tuning foundational models. With base models improving so rapidly, the defensible value lies in building the application layer, workflow integrations, and enterprise-grade software that makes the AI useful, allowing the startup to ride the wave of general model improvement.

A key trend TinySeed observes among AI-focused applicants is extremely high churn, often 10-20% per month. Even with rapid top-line growth, this level is deemed "catastrophic," indicating many new AI products struggle with defensibility and long-term customer value, making them risky investments despite the hype.

Lin warns that much of today's AI revenue is 'experimental,' where customers test solutions without long-term commitment. He calls annualizing this pilot revenue 'a joke.' He advises founders to prioritize slower, high-quality, high-retention revenue over fast, low-quality growth that will eventually churn.

While individual AI companies see slightly lower retention than SaaS, Stripe's data reveals customers often churn from one provider directly to a competitor, and sometimes switch back. This indicates the problem being solved is highly valued, and the churn reflects a rapidly evolving, competitive market, not a lack of product-market fit for the category itself.

Despite significant history and memory built up in platforms like ChatGPT, power users quickly abandon them for models like Claude or Manus that provide superior results. This indicates that output quality is the primary driver of adoption, and existing "memory" is not a strong enough moat to retain users.

For consumer products like ChatGPT, models are already good enough for common queries. However, for complex enterprise tasks like coding, performance is far from solved. This gives model providers a durable path to sustained revenue growth through continued quality improvements aimed at professionals.