Google VP Identifies AI Startups Facing Survival Challenges
The generative AI explosion led to a rapid proliferation of startups. However, as the market matures, two business models that were once highly popular are now emerging as cautionary tales: LLM wrappers and AI aggregators.
Darren Mowry, who heads Google's global startup organization across Cloud, DeepMind, and Alphabet, believes that startups relying on these models are facing significant headwinds.
The LLM Wrapper Dilemma
LLM wrappers are essentially companies that build a product or user experience layer on top of existing large language models, such as Claude, GPT, or Gemini, to address specific user needs. An example could be a startup using AI to assist students with their studies.
"If you are solely relying on the underlying model to do all the work and you are essentially white-labeling that model, the industry has little patience for that anymore," Mowry stated during a recent podcast appearance.
Mowry explained that applying a "very thin intellectual property layer around Gemini or GPT-5" indicates a lack of differentiation. To achieve success and growth, he emphasized the need for startups to possess "deep, wide moats that are either horizontally differentiated or something really specific to a vertical market." Examples of successful LLM wrappers with strong differentiation include Cursor, a coding assistant powered by GPT, and Harvey AI, an AI assistant for the legal sector.
In essence, startups can no longer expect to gain traction simply by adding a user interface to a pre-existing AI model, a strategy that may have seen some success in the earlier days of the generative AI boom. The current challenge lies in building sustainable product value.
AI Aggregators Under Pressure
AI aggregators represent a subset of wrappers. These startups consolidate multiple LLMs into a single interface or API layer, routing queries across different models to provide users with access to a variety of AI capabilities. Typically, these companies offer an orchestration layer that includes features for monitoring, governance, or evaluation. Examples include AI search startup Perplexity and the developer platform OpenRouter, which provides unified API access to numerous AI models.
While many such platforms have gained popularity, Mowry's advice to new startups is unequivocal: "Stay out of the aggregator business."
Mowry suggests that aggregators are currently experiencing limited growth and progression because users desire "some intellectual property built in" that ensures they are directed to the most appropriate model for their specific needs at any given time, rather than being guided by backend infrastructure or access limitations.
Historical Parallels and Future Outlook
With extensive experience in the cloud computing sector, having worked at AWS and Microsoft before joining Google Cloud, Mowry draws parallels between the current AI landscape and the early days of cloud computing in the late 2000s and early 2010s, as Amazon's cloud business began to gain momentum.
At that time, numerous startups emerged to resell AWS infrastructure, positioning themselves as more accessible entry points with bundled tooling, billing consolidation, and support. However, as Amazon developed its own enterprise tools and customers became more adept at managing cloud services directly, most of these intermediary startups were displaced. The survivors were those that offered genuine value-added services, such as security, migration, or DevOps consulting.
AI aggregators today face similar pressures on their profit margins as model providers increasingly develop their own enterprise-focused features, potentially bypassing these middlemen.
Mowry expresses optimism for developer platforms and tools that enhance coding, citing a record-breaking year in 2025 for startups like Replit, Lovable, and Cursor, all of which have attracted significant investment and customer interest, and are Google Cloud customers.
He also anticipates robust growth in direct-to-consumer technology, particularly for companies that empower customers with powerful AI tools. He highlighted the potential for students in film and television to utilize Google's AI video generator, Veo, to bring their creative visions to life.
Beyond AI, Mowry believes that the biotech and climate tech sectors are currently experiencing significant momentum, both in terms of venture capital investment and the vast amounts of data available to startups, enabling them to create unprecedented value.
Stay Tuned to Devignitor Insights for More Updates
