The Brutal Truth About AI Market Sizing Nobody Wants to Hear
We had three weeks of cash left.
That was it. Twenty-one days until my co-founder and I would have to look our team of 12 in the eyes and tell them the dream was dead. The late nights, the caffeine-fueled sprints, the sheer obsession that had consumed our lives for two years—all for nothing. We were about to become another startup ghost story told in hushed tones in Silicon Valley cafes.
Our AI-powered platform for remote team collaboration, which we’d poured our souls into, was a technical marvel. It was also a commercial disaster. A solution desperately searching for a problem. We had the cutting-edge tech, a slick UI designed by a guy who used to work at Apple, and a team of engineers who could probably build a sentient coffee machine if you asked them to. What we didn’t have was a single customer who gave a damn.
And it was all my fault. I had fallen for the oldest, most seductive lie in the book: the siren song of a massive Total Addressable Market (TAM).
The TAM Trap: How I Made a Fool of Myself in a Pitch Meeting
Every founder knows the slide. It’s the one with the hockey-stick graph pointing to the heavens and a number so big it looks like a typo. Billions, sometimes trillions. “The market for AI is projected to be $1.5 trillion by 2030!” We had that slide. I presented it with a straight face.
I remember one pitch meeting vividly. We were in a sleek Sand Hill Road office, the kind with minimalist furniture and expensive art. I was on fire, or so I thought. I got to the TAM slide and delivered my big line: “We’re targeting the global remote work market, a $500 billion opportunity!”
A partner at the firm, a woman known for her brutal honesty, leaned forward. She wasn’t smiling. “Sahin,” she said, her voice dangerously quiet. “How many of those remote workers are in China and can’t use your software? How many work for companies with less than 10 employees and have no budget? How many are already locked into a 3-year contract with Microsoft Teams? What’s your actual market?”
I froze. I mumbled something about a bottom-up analysis we’d done, but the damage was done. She saw right through me. I had presented a fantasy, a vanity metric designed to make VCs feel smart and founders feel like they’re building the next Google. It had almost no bearing on the real, obtainable market for our very specific product.
Think about it. The “AI market” is a meaningless abstraction. It’s everything from a self-driving car navigating a blizzard to an algorithm recommending your next Netflix binge. Just because your product uses a transformer model doesn’t mean you get to claim a piece of that entire pie. It’s like opening a single pizzeria in San Francisco and claiming your TAM is the entire global food industry.
Our top-down analysis was a joke. We took the total number of remote workers, multiplied it by the average price of a SaaS subscription, and—poof!—a multi-billion dollar market appeared out of thin air. It was lazy, and it was dishonest.
Here’s what that lazy math missed:
- The “Good Enough” Problem: Most teams were, and still are, perfectly happy with a messy combination of Slack, Zoom, and Google Docs. Our all-in-one, AI-supercharged platform was objectively better. But was it 10x better? Was it worth the headache and cost of migrating everything and retraining the entire team? For most, the answer was a resounding no.
- The Budget-Owner Dilemma: We were selling a dream to HR managers and team leads. They loved the demo. But the person with the actual checkbook was the CTO or CIO, who had a completely different set of priorities. They didn’t care about our slick UI; they cared about security, compliance, and integration with their existing tech stack. Our sales cycle became a multi-headed hydra, and we were getting eaten alive.
- The “AI for AI’s Sake” Trap: We were nerds in love with our tech. We had a novel approach to sentiment analysis and a slick co-pilot that could auto-complete tasks. We shoehorned it into every feature. But our customers didn’t care about the intricacies of our NLP model. They just wanted to know if our tool could create a transcript of their damn meeting.
We had spent two years and $3 million of other people’s money building a beautiful, high-tech product for a market that didn’t really exist. And the bill was coming due.
The Painful, Necessary Pivot
With our backs against the wall, we had two choices: let the company die a quiet death, or get brutally honest with ourselves. We chose the latter.
We booked a conference room for a week, armed with a whiteboard, a case of Red Bull, and a single, desperate goal: find a real problem we could solve for a specific group of people who would actually pay us. We threw out everything—our pitch deck, our market research, our egos.
We started with that one question: “What is the single most painful, annoying, and persistent problem that remote teams face?”
We became anthropologists. We interviewed over 50 companies in two weeks, from two-person startups to divisions of Fortune 500 giants. We didn’t pitch them. We just listened. We heard their frustrations, their jury-rigged workarounds, and their secret wishes. And a pattern started to emerge, clear as day.
The big problem wasn’t collaboration. It was communication. Or rather, the complete breakdown of it. The challenge of keeping everyone aligned and informed when you’re scattered across time zones.
That’s when the lightbulb went on. It was so obvious we’d missed it. The most valuable part of our entire platform wasn’t the fancy AI project manager. It was a simple, almost-forgotten tool we’d built for ourselves in a weekend hackathon. A tool that transcribed and summarized our own internal Zoom meetings.
We’d built it because we were tired of re-watching hour-long recordings to find a single decision point. It turned out, everyone we talked to had the exact same problem. They were drowning in a sea of video files and messy Google Docs. They were desperate for a way to search their spoken conversations.
So we made the gut-wrenching decision. We were going to kill our baby, the all-in-one platform we’d bled for. We were going to pivot and build the best damn meeting intelligence tool on the planet.
It was terrifying. We were throwing away 90% of our codebase and starting over. But it was also the most liberating feeling in the world. For the first time in a long time, we had a clear, focused mission. We had a real problem to solve.
From Zero to Profitability: The Power of a Niche
The next few months were a blur. We rebuilt the product from the ground up, with a relentless focus on simplicity and user experience. We launched a beta to a hand-picked group of 20 companies and treated their feedback like gold. We iterated, we polished, we obsessed over every detail.
And then, the magic happened. People started paying for it.
It started with a single Stripe notification. Then another. Then a trickle, then a stream, then a flood. Within six months, we hit profitability. We went from the brink of death to a thriving, sustainable business. We eventually sold that company to Gusto.
What changed? We stopped chasing the mythical, multi-trillion-dollar market and focused on a tiny, unsexy niche with a burning, unmet need.
Here are the lessons that were seared into my brain:
- Fall in Love with the Problem, Not Your Solution. I see this constantly in the AI startups I invest in. Founders get so enamored with their clever model or algorithm that they lose sight of the actual problem they’re supposed to be solving. Your tech is a tool, a means to an end. The end is always, always solving a painful problem for a customer.
- Niche Down Until It Hurts, Then Scale. Instead of trying to be everything to everyone, we became the absolute best solution for a very specific person: a manager at a remote company who was tired of wasting time in meetings. That focus allowed us to build a product that was 10x better than the generic transcription services out there. We built a passionate community of early adopters who became our best salespeople.
- The Money is in the Boring Stuff. Everyone wants to build AGI or a world-changing AI model. But the most durable, profitable businesses are often the ones that solve the unsexy, everyday problems that everyone else is too proud to tackle. Transcription, data entry, scheduling, compliance—these are the areas where AI can provide immense value right now.
The Real AI Market Is a Thousand Tiny Markets
So what is the real market for AI? It’s not one big thing. It’s a collection of thousands of tiny, niche markets, each with its own unique problems, language, and opportunities.
The winners in this new era of AI won’t be the ones with the most sophisticated models or the biggest TAM slides. They’ll be the founders who have the humility and the focus to identify a real customer pain and use AI to solve it in a way that feels like magic.
So if you’re an AI founder, I’m begging you: delete the TAM slide from your pitch deck. Stop talking about the trillion-dollar market. Instead, go out and talk to 100 potential customers. Find their biggest, most hair-on-fire problem. And then, and only then, figure out how your AI can be the elegant, indispensable solution.
That’s the brutal truth about AI market sizing. It’s not about the size of the market. It’s about the depth of the pain. And focusing on that might just save your company. It certainly saved mine.
Frequently Asked Questions
How can I apply this thinking to my own situation?
Start by identifying the core principle behind the opinion, not the specific example. Then ask yourself: does this principle apply to my context? If yes, test it in a small, low-risk way before going all in.
Do all experts agree with this view?
No, and that's fine. The best ideas in business are often contrarian. I share my perspective based on my experience and data, but I encourage you to seek out opposing viewpoints and form your own conclusions.
What's the most common pushback you get on this?
People often push back by citing exceptions or edge cases. And they're usually right that exceptions exist. But building a strategy around exceptions rather than patterns is a losing game for most founders.
What experience informs this perspective?
This perspective comes from over a decade of building companies in Silicon Valley, two successful exits (RemoteTeam to Gusto, MovieLaLa to Gfycat), and investing in 200+ startups including Anthropic, OpenAI, and Scale AI. I write about what I've lived.