To survive against subsidized tools from model providers like OpenAI and Anthropic, AI applications must avoid a price war. Instead, the winning strategy is to focus on superior product experience and serve as a neutral orchestration layer that allows users to choose the best underlying model.

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 successful strategy for AI startups is to initially leverage state-of-the-art foundation models to acquire users and data. Once sufficient high-quality, domain-specific data is collected, they can train their own specialized models to drastically cut costs and latency.

Companies like Z.ai are not abandoning open source but using it strategically. They release lightweight models to attract developers and build a user base, while reserving their most powerful, agentic systems for proprietary, revenue-generating enterprise products, creating a clear monetization funnel.

Counter to fears that foundation models will obsolete all apps, AI startups can build defensible businesses by embedding AI into unique workflows, owning the customer relationship, and creating network effects. This mirrors how top App Store apps succeeded despite Apple's platform dominance.

Contrary to past momentum, the most advanced AI startups are increasingly adopting and fine-tuning open-source models. This shift is driven by the need for cost-effective speed and deep customization as their workloads mature and scale.

Like Kayak for flights, being a model aggregator provides superior value to users who want access to the best tool for a specific job. Big tech companies are restricted to their own models, creating an opportunity for startups to win by offering a 'single pane of glass' across all available models.

Open source AI models don't need to become the dominant platform to fundamentally alter the market. Their existence alone acts as a powerful price compressor. Proprietary model providers are forced to lower their prices to match the inference cost of open-source alternatives, squeezing profit margins and shifting value to other parts of the stack.

When ChatGPT commoditized AI writing assistants, AI21 pivoted. They leveraged their proprietary foundation model to build a new product (Maestro) for the enterprise, solving the emerging problem of multi-model orchestration rather than defending a now-obsolete consumer app.