Is your startup developing core AI or is it just another YAWOC?

It’s no secret that venture capital has gone ga-ga over AI.

In 2024, 33% of VC dollars went into AI startups. Thought that was crazy? In 2025, 65% of U.S.-based VC investment piled into AI.

AI is sucking all the air out of the room. And all the money out of investors’ pockets. If you’re developing battery technology, or plastics recycling, or SaaS for bowling alleys, what’cha gonna do?

The answer for most founders seems to be that if you can’t beat ‘em, join ‘em: if investors are investing exclusively in AI, then you have no choice but to turn your SaaS startup into an AI-enabled, AI-powered, AI-focused startup.

One of the startups I invested in when it was an energy-saving solution for commercial offices, this year presented itself as the AI-solution for optimized building operations. Did they actually change anything? Other than the pitch deck, no. Were they actually even using AI? No. But they would, they promised, as soon as they had enough investment to hire a team of AI scientists to solve impossible problems with AI magic. Ugg.

This week I received a deck from a startup pitching an “AI-powered recycling business.” What were they using AI for? As far as I could tell, it was nothing but the AI features built into Hubspot to help identify and qualify potential customers. Okaaaaaay. Pass.

Suddenly every pitch has “⚡AI-POWERED⚡” in their one-line blurb intro. A few years ago, everything was blockchain or Web3, and before that it was SaaS up the wazoo. But is the startup really AI, and does that even matter? The answer is: it depends on what type of AI startup you are.

3 Types of AI Startups

Not all AI startups are the same. I put them into 3 buckets:

  1. Core AI. When you think of AI, what comes to mind? ChatGPT. Gemini. Anthropic. Grok. DeepSeek. The big LLMs that are changing the world. There’s also TensorFlow and PyTorch for building your own neural networks.
  2. AI Applications. There’s a million new applications that wouldn’t be possible without AI. Low-code/No-code development tools. Radiology analysis. Protein folding software. Agents for everything and anything.
  3. Everyday AI. Nowadays, is there anyone who isn’t using AI tools for market research, for sales pitches, for faster coding? Pitch decks are written by AI. Investment applications are analyzed by the same AI.

Core AI is real. This is where the vast majority of venture investment is going, though if you ask me, Microsoft putting $13 billion into OpenAI shouldn’t be categorized as venture capital investment. Still, it seems likely that OpenAI will soon be the most valuable company in the world. And all the pickleball buddies of Sam Altman are about to become billionaires. If you’re working on Core AI, expect your phone to be ringing day and night from AI-focused VCs begging to be allowed to invest at any price.

Applications using AI are exploding, and for good reason — AI enables solutions that weren’t previously possible. Where the previous craze of SaaS was simply a switch of business model and blockchain of limited value, AI applications are remaking how humans interact with the world. There are a lot of big incumbents across every industry about to be disrupted by AI upstarts, and whole new classes of solutions that never existed before. A few founders and their investors will hit goldmines.

But…these are not AI startups. These are medical startups, chemistry startups, real estate startups, and robotics. AI is a tool, like a programming language or a database. The customer doesn’t care if you solve their problem using AI or heuristics, or even oracular hamsters, so long as it works. Investors don’t care about your tech stack, so long as it solves a customer problem.

For AI applications, the investors are not AI investors — but the same investors already focused on your customers’ industry. They’re medtech investors, chemistry investors, proptech investors, and robotics industry specialists. The fact that you’re using AI sits at the center of the “why now?” slide, but not the center of the business.

And if you fit into the 3rd category of everyday AI, calling yourself an AI-powered startup is just complete b.s. Don’t do it. You might think it’ll get investors to look, but you’ll immediately lose all credibility.

The Challenges of “AI-Powered” Applications

Investors currently see a lot of pitches for AI-enabled applications (and many that say they are AI-enabled and aren’t.) These startups present a great opportunity for angels and small VCs to invest in applications that could remake an industry, or at least provide enough value to get bought out by one of the industry giants.

However, there are 2 complications that come up nearly every time:

You’re feeding proprietary data into ChatGPT (Anthropic, or other LLM model) to automatically write company press releases, prepare financial documents, review contracts, translate instruction manuals, troubleshoot computer networks, or a million other things that currently rely on humans. Great! There’s a real need and a serious opportunity.

But…what happens when the next model ChatGPT does all this itself? All those low-code/no-code startups from 5 years ago were wiped out now that all the major LLMs can turn my 90 year mother into a vibe coder.

And if you can feed data into ChatGPT, so can I. If you no longer need to hire 10 programmers to build the app, neither do I. AI has made the bar to entry ridiculously low. Every college student in the world, half the high school students, and even many junior high school students are potential competitors. How will you prevail? Hint: it’s not about the tech stack. It’s not about the AI. With low barriers to entry, it’s about the marketing and industry connections. Or about access to training data that nobody else can get.

Investors hear so many of these pitches for AI applications that we’ve started calling them YAWOC — “Yet Another Wrapper on ChatGPT.”

That’s not to say that YAWOCs are bad, just that investors are seeing so many of them and most can’t answer these two basic questions:

  • What stops the large LLMs from taking your market?
  • What’s your competitive moat?

How To Pitch Your AI-Enabled Startup

You have a startup that takes advantage of the power of AI to provide a solution to a problem that previously couldn’t be solved. Great. How do you pitch it? The rules haven’t changed.

“AI Investors” aren’t looking for you or any other YAWOC. Sorry. They’re investing in core AI functionality. Don’t waste your time.

Who is investing? The same people as before. The people familiar with your industry who understand the problem you’re addressing and appreciate the value of your solution.

The AI doesn’t matter. The tech stack doesn’t matter. What matters is the problem and product. Your product isn’t AI, it’s software or hardware that solves someone’s (hopefully a lot of someones) problem that happens to use AI.

So go ahead and mention the AI, but there’s no need to dwell on it. Focus instead on how you’ll build a competitive moat in the industry when the barriers to entry are so low. Focus on how you’re going to get to $100M in revenues and get acquired by an industry giant to make investors rich.

In other words, ignore the AI hype and all the money going into core AI. That’s irrelevant other than it’s helping you build your tech stack for cheap. Convince us you have a great business that will be a successful investment. But if you try to hoodwink us with the supposed sizzle of AI, you’ll only burn yourself.