Google VP warns that two types of AI startups may not survive

Source: TechCrunch

LLM Wrappers

LLM wrappers are startups that layer a product or UX on top of existing large language models—such as Claude, GPT, or Gemini—to solve a specific problem. For example, a startup that uses AI to help students study.

“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,” said Darren Mowry on this week’s episode of Equity (TechCrunch podcast).

Mowry warns that “very thin intellectual property wrapped around Gemini or GPT‑5” signals a lack of differentiation. He stresses that startups need deep, wide moats—either broad horizontal differentiation or a highly specific vertical focus—to “progress and grow.”

Examples of Deep‑Moat LLM Wrappers

• Cursor – a GPT‑powered coding assistant.
• Harvey AI – a legal AI assistant.

In contrast, the era of simply slapping a UI on top of a GPT model—like the surge that followed OpenAI’s ChatGPT Store launch in mid‑2024—has faded. Sustainable product value now requires more than a superficial interface.

AI Aggregators

AI aggregators are a subset of wrappers that combine multiple LLMs into a single interface or API, routing queries across models and often providing monitoring, governance, or evaluation tooling. Notable examples include:

• Perplexity – an AI‑search startup.
• OpenRouter – a developer platform offering access to multiple AI models via a single API.

Mowry’s advice to newcomers is clear: “Stay out of the aggregator business.” He notes that users now expect “some intellectual property built in” to ensure queries are routed to the right model at the right time, rather than relying solely on behind‑the‑scenes compute or access advantages.

Industry Context

Mowry, who cut his teeth at AWS and Microsoft before joining Google Cloud, sees a parallel with the early days of cloud computing (late 2000s–early 2010s). Back then, many startups resold AWS infrastructure, promising easier entry points, tooling, and billing consolidation. When Amazon introduced its own enterprise tools and customers learned to manage cloud services directly, most of those resellers were squeezed out—survivors were those that added real services like security, migration, or DevOps consulting.

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

Mowry’s Outlook

• Developer Platforms & Coding Tools – Mowry is bullish on “vibe coding” and developer platforms, citing a record‑breaking 2025 for startups such as Replit, Lovable, and Cursor (all Google Cloud customers).
• Direct‑to‑Consumer Tech – He expects strong growth in companies that put powerful AI tools directly into customers’ hands, e.g., using Google’s AI video generator Veo for film and TV students.
• Beyond AI – Mowry also sees momentum in biotech and climate tech, driven by abundant data that enables startups to create real value in ways previously impossible.