I’ve seen it a thousand times. A founder comes to me, eyes wide with excitement, convinced they’ve built the next unicorn. Their secret sauce? A revolutionary new algorithm that’s 5% more accurate than anything else on the market.
They think the algorithm is their moat. Their impenetrable defense against the competition.
I hate to be the one to break it to them, but they’re wrong.
In the world of AI, the algorithm is not the moat. It’s the table stakes. It’s the price of admission. It’s the starting line, not the finish line.
I’ve been fortunate enough to be an early investor in some of the most successful AI companies of our time—Anthropic, OpenAI, Scale AI, Hugging Face. I’ve also built and sold two of my own companies, RemoteTeam and MovieLaLa. And if there’s one thing I’ve learned, it’s that a better algorithm alone won’t save you.
The Real Moats in AI
So if the algorithm isn’t the moat, what is? The answer is a lot more nuanced than a single piece of code. It’s a combination of factors that, when woven together, create a powerful and defensible business.
1. Proprietary Data
Data is the lifeblood of AI. The more high-quality, proprietary data you have, the better your models will be. This is a classic flywheel effect: better models attract more users, who generate more data, which in turn leads to even better models. It’s a virtuous cycle that’s incredibly difficult for competitors to replicate.
Look at Scale AI. They didn’t just build a better algorithm for data labeling. They built a massive, human-in-the-loop data engine that produces high-quality training data at a scale that no one else can match. That’s their moat.
2. Distribution and Network Effects
You can have the best AI model in the world, but if no one uses it, it’s worthless. Distribution is key. How do you get your product into the hands of users? How do you build a network of users that creates a defensible moat?
Hugging Face is a perfect example of this. They built a community of developers who contribute and share models, datasets, and tools. This has created a powerful network effect that makes it the go-to platform for anyone working in AI. Their moat isn’t just their technology; it’s their community.
3. Brand and Trust
In a world where AI is becoming increasingly powerful and pervasive, trust is paramount. Users need to trust that your AI is reliable, unbiased, and secure. Building a brand that is synonymous with trust is a powerful moat.
This is something we’re seeing with companies like Anthropic. They are not just focused on building powerful AI models, but also on building safe and ethical AI. This focus on trust and safety is a key differentiator and a source of competitive advantage.
4. The Full-Stack Solution
Another powerful moat is to own the entire stack, from the underlying infrastructure to the end-user application. This gives you more control over the user experience, allows you to capture more value, and makes it harder for competitors to displace you.
Think about what OpenAI is doing. They are not just providing an API for their models. They are building applications like ChatGPT that are used by millions of people every day. They are building a full-stack solution for AI, and that is a very powerful position to be in.
My Own Journey: From Algorithm to Exit
I learned this lesson the hard way with my own startups. At MovieLaLa, we had a pretty good recommendation algorithm. But what really made us valuable was our user base and the data we collected on their movie preferences. That’s what ultimately led to our acquisition by Gfycat.
At RemoteTeam, we were building tools for remote teams long before it was the new normal. Our moat wasn’t a single algorithm, but a deep understanding of the needs of remote workers and a suite of tools that addressed those needs. That’s what made us an attractive acquisition for Gusto.
The Bottom Line
So, if you’re an AI founder, I urge you to think beyond the algorithm. Don’t get me wrong, you need a great algorithm. But that’s just the beginning.
Focus on building a business with multiple, overlapping moats. Obsess over your data, your distribution, your brand, and your user experience. That’s how you build a truly defensible AI company.
The algorithm might be the spark, but it’s not the fire. The fire is the business you build around it.
Frequently Asked Questions
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.
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.
How has this view evolved over time?
My thinking on most topics has changed significantly over the years. Early in my career, I held many conventional views that experience proved wrong. I try to update my beliefs when the evidence changes.