Why Google’s VP Warns Against Two Popular AI Startup Models

In a recent TechCrunch "Equity" podcast, Darren Mowry, Google Cloud’s VP for global startup business, warned that two once‑trendy generative‑AI startup models are now showing warning lights.

LLM wrappers —companies that simply layer a product or user‑experience on top of existing large language models such as Claude, GPT, or Gemini—are losing patience in the market. Mowry said that merely white‑labeling a model with a thin IP layer offers no real differentiation, and investors are no longer eager to back such approaches. He cited the AI‑coding assistant Cursor as a rare success, having raised $2.3 billion in a Series D round and reaching a valuation near $30 billion.

He also discussed AI aggregators , a subset of wrappers that combine multiple LLMs behind a single UI or API, providing routing, monitoring, governance, or evaluation tools. Examples include the search‑oriented startup Perplexity and the multi‑model developer platform OpenRouter . Mowry’s advice: avoid building aggregator businesses because users now expect embedded IP that intelligently routes queries to the right model, not just a generic compute layer.

Mowry drew a parallel to the early cloud‑computing era of the late 2000s, when many startups resold AWS infrastructure until Amazon introduced its own enterprise tools, squeezing out the resellers unless they added genuine services such as security, migration, or DevOps consulting. He sees a similar profit‑pressure dynamic emerging for AI aggregators as model providers expand native enterprise features.

Looking forward, Mowry is optimistic about "vibe coding" and developer platforms, naming startups like Replit , Lovable , and Cursor (all Google Cloud customers) as poised for a breakthrough in 2025. He also expects strong growth for consumer‑facing AI tools, citing Google’s AI video generator Veo for film and TV students.

Beyond AI, Mowry highlighted biotech and climate‑tech as golden opportunities, noting abundant data and venture capital that can create real value in ways previously unattainable.