The recent boom in generative AI led to a rapid proliferation of startups, but as the initial excitement subsides, two specific business models are now signaling caution: LLM wrappers and AI aggregators. Darren Mowry, who oversees Google's global startup operations across Cloud, DeepMind, and Alphabet, indicates that startups relying on these approaches are encountering significant challenges, metaphorically having their “check engine light” on.

LLM wrappers are essentially companies that build a product or user experience layer around existing large language models, such as Claude, GPT, or Gemini, to address a particular problem. A common example would be an AI-powered study assistant for students. However, Mowry emphasizes that the industry is losing patience with startups that merely white-label a backend model, stating, “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.” He stresses the necessity for startups to establish “deep, wide moats” through either horizontal differentiation or specialized vertical market solutions to ensure growth and progress. Notable examples of successful, differentiated LLM wrappers include Cursor, a GPT-powered coding assistant, and Harvey AI, a legal AI assistant. This signifies that the era where simply applying a UI atop a GPT model guaranteed traction, as might have been possible with the launch of OpenAI’s ChatGPT store in mid-2024, is over; the current imperative is to build sustainable product value.

AI aggregators, a subset of wrappers, are startups that integrate multiple LLMs into a single interface or API layer. Their purpose is to route user queries across various models, providing access to a diverse range of AI capabilities. These platforms often incorporate an orchestration layer, offering monitoring, governance, or evaluation tools. Examples include the AI search startup Perplexity and the developer platform OpenRouter, which facilitates access to multiple AI models through a unified API. Despite some aggregators gaining initial traction, Mowry’s advice to new startups is unequivocal: “Stay out of the aggregator business.” He observes that aggregators are generally seeing limited growth or progression, primarily because users seek “some intellectual property built in” to ensure optimal model routing based on specific needs, rather than routing determined by backend compute or access constraints.

Mowry, with decades of experience in the cloud industry, including tenures at AWS and Microsoft before joining Google Cloud, draws a parallel between the current situation and the early days of cloud computing. In the late 2000s and early 2010s, as Amazon’s cloud business expanded, numerous startups emerged to resell AWS infrastructure, promoting themselves as easier entry points with added tooling, billing consolidation, and support. However, when Amazon developed its own enterprise tools and customers became adept at managing cloud services directly, most of these reseller startups were eliminated. Only those that provided genuine value-added services, such as security, migration, or DevOps consulting, managed to survive. AI aggregators today face similar margin pressures as model providers increasingly extend their offerings into enterprise features, potentially marginalizing the middlemen.

Despite these cautionary tales, Mowry remains optimistic about certain sectors. He is particularly bullish on vibe coding and developer platforms, which experienced a record-breaking year in 2025. Startups like Replit, Lovable, and Cursor—all Google Cloud customers—attracted significant investment and customer adoption. Mowry also anticipates robust growth in direct-to-consumer technology, especially companies that empower customers with advanced AI tools. He highlighted the potential for film and TV students to utilize Google’s AI video generator, Veo, to bring their creative stories to life. Beyond AI, Mowry identifies biotech and climate tech as industries experiencing a significant moment, evidenced by substantial venture investment and the unprecedented availability of data that enables startups to create substantial value.