Does AI really create a competitive moat for your company?
Is AI a core part of your strategy? Or better yet, does AI provide you protection from your competition?
It’s hard to find a company these days which in various degrees is not involved with AI. There are companies building AI models like OpenAI, Anthropic or Mistral AI. Then you have the next tier which extends the functionality of these basic models with a specific functionality. Examples might be Cognition AI, Replit or Perplexity.
And then you have companies which already have a product and are using AI functionality to enhance their product or services. Examples would include Klarna, Duolingo or Shopify.
Since we know that AI is the future — or we are told — these companies should have an enormous advantage over their competition. They have a moat — a structural advantage — which protects their market share and profits.
Any business school will teach you that the main types of competitive moat are:
network effect — the more users are using your system, the more valuable your product or service is
cost advantage — your company can produce things cheaper than anyone else
switching cost — it is very expensive to switch from your product to the competition
intangible assets — you have brand recognition, patents, licensing deals which prevents newcomers from entering your space
efficient scale — your company is so big that the market can’t support more companies
proprietary data — specialized data which is difficult, expensive to obtain
What does a moat give you?
You get:
pricing power — you can raise prices without fear of losing customers
investor attention — higher valuation, easier access to capital
better margins — difficult for others to compete
more time to respond to competitive threats
Now, if you use the above definition to evaluate current players in the market, let’s see how well they are ready to fight their competition. BTW: if you find a different and/or better definition of what a moat is and what benefit it provides, please use it for your own evaluation.
OpenAI — I don’t see there is any network effect in place. Yes, OpenAI claims millions and millions of users, but few pay for the service. The only value these users can deliver are eyeballs for the advertisers, but more advertisers don’t deliver any value to the users of ChatGPT. There’s no cost advantage, either since they fight with Anthropic and are undercut by all the free models. The switching cost is negligible. Maybe there is some brand recognition and patents and the market is big enough to support many companies like OpenAI.
The only advantage OpenAI might actually have is in their proprietary data. A recent announcement from OpenAI points that way, except that it is using data provided by others. The fact that there is no moat is highlighted with no pricing power (remember, OpenAI tried to change the pricing and had to very quickly revert). It has no margins since it is losing money on every subscription. Yes, it is trying to respond to competitive threats with limited success. The only thing which is going for OpenAI is investor attention because of promises of a rosy future.
Let’s pick one from the next tier. Perplexity is a real-time Answer Engine which is using AI provided mostly by others and augmenting the answers with information crawled from the Internet. No network effect, no cost advantage, no switching cost, no intangible assets, no efficiency scale and no proprietary data to speak of. Similar to OpenAI and Anthropic, Perplexity is around only because investors hope that it will go public.
By way of comparison with a very different kind of company, what about Shopify? Network effect — totally. The more people go to the website to shop, the more people will try to set up shop there. Shopify has scale and adding an extra few thousands customers/sellers makes no difference. For any incumbent to compete that’s the hill they will die on. The switching cost is high. The bigger the seller, the more expensive it is to switch. Proprietary data — check. Shopify sits on tons of data which can be used to further develop the platform, create better marketing, and improve new customers onboarding. The result — a public company with a market cap of 167 billion dollars.
Where does the AI come into the picture here?
Developing an open standard with Google to allow AI Agents to do the shopping. Agentic Storefronts to help sellers to distribute their products to AI platforms. Shopify Sidekick to make it easier integrate with third party app eg. increasing network effect, increasing switching cost, building brand, getting more data.
That’s what moat looks like.
The companies building ‘AI’ exist only because of investors piling money into a technology. The investors are counting on these companies to develop the tech and hoping it would be a reality-altering one. Except for the money, these companies have no moat.
The recurrent pattern? Giving technology a fancy name doesn’t change the fact that it is still just a technology. A thing which — when adopted by business — can only support your competitive advantage, not be one.