Google Executive Sounds Alarm on Two Flawed AI Startup Models

The initial frenzy of the generative AI boom has passed, and a Google executive overseeing startups says the market is losing patience with two popular but fragile business approaches. Darren Mowry, who leads Google's global startup organization, suggests companies built as simple "wrappers" for existing models or as AI aggregators are showing serious warning signs.

Mowry describes LLM wrappers as startups that apply a thin product layer atop foundational models like GPT or Gemini to address a niche. "If you're really just counting on the back end model to do all the work... the industry doesn't have a lot of patience for that anymore," he stated on the podcast Equity. He argues survival requires "deep, wide moats," such as highly specialized vertical expertise, pointing to successes like the coding assistant Cursor.

He was even more direct about AI aggregators—platforms that route user queries across multiple models. "Stay out of the aggregator business," Mowry warned. He explained that users now demand intelligent routing based on proprietary logic, not just convenient access. Growth is stalling for middlemen as major model providers expand their own enterprise services.

Mowry draws a parallel to the early cloud computing era, where startups that merely resold AWS infrastructure were squeezed out when Amazon developed its own tools. The survivors added genuine value through services like security or consulting.

Looking ahead, Mowry is optimistic about developer platforms, which saw record investment in 2025, and direct-to-consumer AI tools. He also highlighted biotech and climate tech as sectors leveraging vast new data streams to build substantial, durable companies.