Two types of AI startups may not survive, Google VP warns

The AI ​​boom has created a startup in a minute. But as the dust begins to settle, two once-hot business models are becoming more like cautionary tales: LLM wrappers and AI aggregators.

Darren Morey, who leads Google’s global startup organization across Cloud, DeepMind and Alphabet, says startups using these hooks turn on their own “check engine light.”

LLM wrappers are basically startups that wrap existing large language models, such as Claude, GPT, or Gemini, with a product or UX layer to solve a specific problem. An example is a startup that uses artificial intelligence to help students study.

“If you’re really relying 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,” Morey said on this week’s episode of Equity.

Wrapping “very thin intellectual property around Gemini or GPT-5” indicates that you’re not distinguishing yourself, Mori says.

“You have to have deep, wide moats that are either horizontally differentiated or something really specific to the vertical market” so a startup can “advance and grow,” he said. Examples of the deep trench LLM cover type include Cursor, a GPT-powered coding assistant, or Harvey AI, an artificial intelligence legal assistant.

In other words, startups can no longer expect to add a UI on top of GPT and get traction for their product, as they could, perhaps, in mid-2024 when OpenAI launched its ChatGPT store. The challenge now is to build sustainable product value.

AI aggregators are a subset of wrappers — they are startups that bundle multiple LLMs into a single interface or API layer to route queries across models and give users access to multiple models. These companies typically provide an orchestration layer that includes monitoring, management, or evaluation tools. Consider AI research startup Perplexity or the developer platform OpenRouter, which provides access to multiple AI models via a single API.

While many of these platforms have gained popularity, Mori’s words are clear for upcoming startups: “Stay out of the aggregator business.”

In general, aggregators aren’t seeing a lot of growth or progress these days because, he says, users want “some built-in intellectual property” to ensure they’re routed to the right model at the right time based on their needs — not because of computing or access limitations behind the scenes.

Mowry has been in the cloud game for decades, gaining significant experience at AWS and Microsoft before setting up shop in Google Cloud, and he’s seen how it’s done. The situation today reflects the early days of cloud computing in the late 2000s and early 2000s, as Amazon’s cloud business began to take off, he said.

At that time, a group of startups emerged reselling AWS infrastructure, marketing themselves as easier entry points that provided tools and consolidated billing and support. But as Amazon built its own enterprise tools and customers learned to manage cloud services directly, most of those startups were left out. The only survivors are those who have added real services, such as security, migration, or DevOps consulting.

Today’s AI aggregators face similar margin pressure as model providers expand into enterprise features themselves, potentially marginalizing middlemen.

For his part, Mowry expressed optimism about programming and developer platforms, which had a record year in 2025 with startups like Replit, Lovable, and Cursor (all Google Cloud customers, per Mowry) attracting significant investment and attracting customers.

Morey also expects strong growth in direct-to-consumer technology, in companies that are putting some of these powerful AI tools into customers’ hands. He pointed to the opportunity for film and television students to use Google’s AI-powered Veo video generator to bring stories to life.

Beyond AI, Mori also believes biotech and climate tech are having a moment – ​​both in terms of venture investment in the two industries and the “amazing amounts of data” that startups can access to create real value “in ways we’ve never been able to do before.”