To avoid being made obsolete by a frontier AI model, startups need a strong moat. The three most defensible moats are: 1) building hardware, which AI cannot physically replicate, 2) establishing strong network effects where value increases with more users, and 3) operating in a complex, regulated industry requiring human interaction.

While foundational AI models threaten broad applications like writing aids, startups can thrive by focusing on vertical-specific needs. Building for niche workflows, compliance, and deep integrations creates a moat that large, generalist AI companies are unlikely to cross.

While patents are important, a pharmaceutical giant's most durable competitive advantage is its ability to navigate complex global regulatory systems. This 'regulatory know-how' is a massive barrier to entry that startups cannot easily replicate, forcing them into acquisition by incumbents.

For entrepreneurs building on top of large language models, the key differentiator is not creating general platforms but achieving deep domain specialization. The call to arms is to know a vertical better than anyone and imbue that unique knowledge into AI agents, creating a defensible moat against more generalized tools.

The enduring moat in the AI stack lies in what is hardest to replicate. Since building foundation models is significantly more difficult than building applications on top of them, the model layer is inherently more defensible and will naturally capture more value over time.

As AI commoditizes software, the most defensible businesses are no longer asset-light SaaS models. Instead, companies with physical world operations, regulatory moats, and liability are safer investments. Their operational complexity, once a weakness, now serves as a formidable barrier against pure AI-driven disruption.

AI could theoretically provide world-class legal, medical, or educational advice. However, it cannot disrupt these fields because it can't get licensed, admitted to the bar, or receive insurance reimbursements. These regulatory moats will keep these professions untouched by AI's capabilities for the foreseeable future.

The venture thesis for AI is shifting towards companies that cannot be easily absorbed as features by large platforms like OpenAI. Investors are targeting startups with defensible moats derived from navigating complex regulations (e.g., medical) or owning unique, proprietary datasets that are difficult to replicate.

In the AI era, defensibility comes from building a complex system of record, not just a thin wrapper on an LLM. Companies with a 'thick application layer' that offers standalone value are unattractive for model providers to replicate, whereas thin wrappers risk being absorbed by the platform they are built on.