Google VP warns two AI startup models may not survive

The era of easy AI startup creation is facing a structural shift. Darren Mowry, who leads Google’s global startup organization, warns that two common business models—large language model (LLM) wrappers and AI aggregators—are facing significant viability challenges. While the generative AI boom previously minted “a startup a minute,” Mowry indicates that industry patience for companies that simply white-label existing models is wearing thin. “If you’re really just counting on the back-end model to do all the work and you’re almost white-labeling that model, the industry doesn’t have a lot of patience for that anymore,” Mowry said. He argues that startups need deep, wide moats—defined as competitive advantages that protect a business—to progress and grow.

The warning distinguishes between two types of thin software layers. LLM wrappers build a basic product or user experience layer on top of existing models like Claude, GPT, or Gemini. AI aggregators are a subset of wrappers that route queries across multiple models, providing an orchestration layer for platforms like AI search startup Perplexity or developer platform OpenRouter. Mowry’s advice for new founders is direct: “Stay out of the aggregator business.” These aggregators face margin pressure as model providers like Google, Microsoft, and AWS expand their own enterprise features, potentially sidelining middlemen. Mowry compares this to the early days of cloud computing in the late 2000s/early 2010s, when startups that merely resold AWS infrastructure were squeezed out once Amazon built its own enterprise tools.

Instead, Mowry highlights growth in developer platforms and “vibe-coding”—which refers to AI-powered coding and development platforms—that offer deeper differentiation. Startups like Cursor (a GPT-powered coding assistant), Harvey AI (a legal AI assistant), Replit, and Lovable are building the necessary competitive advantages to succeed. In 2025, developer platforms experienced a record-breaking year for investment and customer traction. Mowry also points to opportunities in direct-to-consumer applications, such as film students using Google’s AI video generator Veo, as well as data-rich sectors like biotech and climate tech.

Why it matters

The AI startup ecosystem is maturing as investors and incumbents shift focus from simple model-wrapping to businesses with deep, wide competitive advantages. This transition signals a much tougher environment for startups operating as thin software layers.