As the Co-Founder and CTO of one recently acquired AI startup and the former head of engineering of another, I’ve recently come to a realization: there’s no such thing as an AI startup. I have three reasons for this, and a takeaway at the end.

Reason 1: AI is everywhere

Every serious company is now using Generative AI. This rapid adoption is exciting, but it also means AI is no longer a differentiator. Big tech companies are trying to stay ahead, non-tech enterprises are following investor demands to incorporate it quickly, and small companies are adopting AI for convenience, even on limited budgets.

If everyone is incorporating AI into their business, it’s no longer a meaningful label to distinguish one set of businesses from another.

Reason 2: Products, not AI, are what matter

Artificial intelligence is just a concept: machines learning and interacting like humans. What “AI Companies” actually sell is different. OpenAI and Anthropic offer proprietary model services, Nvidia sells GPUs for training and running models, and Cursor and Windsurf sell developer coding tools.

Look at the value delivered to users; it’s never just AI. It’s something specific, enhanced by AI technologies or the output of AI technologies. Various companies using AI technologies look very different from each other - because they are different.

Two “AI Startups” will have more differences than similarities, making it a useless label for understanding a company’s fundamentals.

Reason 3: No AI monopolies

Generative AI is not a secret, so no single company will capture all or most of its value. Venture-backed startups try to capture market value, but no one will capture the entire “AI market.”

Open scientific research created current generative AI and continues to drive its improvement and application. For example, open research into model distillation led to the surprise performance of the Deepseek R1 model. Research into input processing and memory optimization led to foundation models increasing their context windows. While individual labs will have advantages, I argue that there’s no fundamental bottleneck on AI technologies. Every time one player pulls ahead, others will catch up.

With no way to capture the entire AI market, it makes no sense to associate the total value delivered using AI with any startup’s “total addressable market.” AI’s value will be too big and ill-defined. Meaningful products will serve specific markets, best described without “AI” in the name.

Strategic Takeaway: Focus on Outcomes, Not Labels

The best way to understand AI’s impact and strategic effects on business is to look at the delivered outcomes. Look at “AI Search” rephrasing questions, delving deeper into sources, and compiling results - then you’ll understand how search is changing. Look at “AI Coding Tools” augmenting the process of going from ideas to code - then you’ll understand how software development is changing.

When looking at a company or product, ignore the “AI” label. Soon, “AI Company” will be as meaningless as “Digital Company” or “Internet Company.” To better understand the future, focus clearly on what these companies are actually doing and producing.