Are we entering a world with no moats?
The great levelling
There is a peculiar irony in the current moment. Artificial intelligence has made business more fluid, more productive, and almost frictionless. Yet, with that same brisk efficiency, it has demolished the one thing every traditional business plan promises its investors: a defensible moat.
The arithmetic is simple enough. When any competent team can summon a product, a pitch, a pricing model, and a marketing calendar before lunch, the cost of entry collapses. And when that barrier falls, so does the value of being inside: the traditional corporate fortress now lowers its drawbridge for anyone with a subscription and a reasonable prompt.
What we are left with is a market that is simultaneously more capable and more homogeneous than it has ever been. Everyone is faster. Everyone is cheaper. Everyone, increasingly, is the same.
Consider what investors are now saying about companies with over 100 million in revenue: that they must rediscover product-market fit every year, because nothing they built last year is defensible this year. Product-market fit used to be a graduation. It has become a subscription.
Is IP dead, and did we kill the culture too?
Intellectual property is not dead, but it has been demoted. A patent still stops a competitor from copying your mechanism. It does nothing to stop them from reaching the same outcome by a different route in a fortnight, which is what the machines are rather good at. The clever idea that once bought you a decade now buys you a quarter, and the lawyers are the only ones celebrating.
The quieter casualty is culture. A firm's DNA used to be the sum of a thousand idiosyncratic decisions about how it hired, argued, wrote and sold. It was expensive, inconsistent and occasionally maddening, which is precisely why it was hard to copy. Now the same models draft the same memos in the same reasonable tone for everyone, and the idiosyncrasies get optimised out.
The result is a great regression to the mean, and the mean can be perfectly good. That is the problem. Good is now the floor, and a floor is not a strategy. The bad firms have been pulled up to competent, the great firms have been nudged down to competent, and competent is a crowded place to live.
A room full of magic
At 1648 Capital we see this every day, but it was never quite so vivid as at an AI conference we attended recently. The companies on show were impressive. Sharp founders, polished demos, genuine technical talent. And yet, by the third hour, a strange sensation set in. The solutions were similar. The go-to-market strategies were similar. The slide about the enormous addressable market was, as far as one could tell, the same slide.
Every second pitch promised an AI harness that could solve every problem. Here is the trouble with that. If your harness can solve every problem, and you are competing against hundreds of harnesses that can also solve every problem, you are no longer a solution. You are just noise. Expensive, well-funded, beautifully designed noise.
None of these founders were foolish. They were simply working from the same playbook, and the playbook had been written by the same tools they were selling. It is hard to build something nobody else has when everybody else is using the same oracle to tell them what to build.
The sharpest illustration of this doing the rounds involves a founder whose product lets you ask AI agents to build an exact replica of any existing app. Name the thing you admire and a working copy appears in a few hours. Genuinely remarkable. He had 200 followers and no idea how to sell it. When building is free, the bottleneck moves to the one thing the harness cannot do for you, which is persuading a specific human being to trust you with their money.
The MBA problem, revisited
I have seen this film before, albeit with a smaller budget. When I was a senior manager at UBS I made a point of not hiring MBA candidates, or at the very least of treating the credential as a mild warning sign. This was not snobbery, or not entirely. It was that they all arrived with the same way of solving problems. The same frameworks, the same two-by-twos, the same confident march from situation to complication to resolution. Put six of them in a room and you got one answer, very thoroughly argued.
What I wanted instead were people who had been somewhere strange and come back with a different map. The former engineer who thought in tolerances. The trader who thought in regret. The person who had run a failing business and knew what the numbers felt like before they appeared on a spreadsheet.
AI is the MBA problem at planetary scale. It is the most thoroughly trained graduate ever to enter the workforce, and it has been hired by everyone at once. It will give you the consensus answer faster and more fluently than any human could. It will not, by construction, give you an answer nobody else has.
So how does one dig a moat now?
The honest answer is that the moat has moved. It no longer sits in the technology, which everyone has, nor in the process, which everyone can generate. It sits in two unfashionable places: business focus and creativity.
Focus first. The firms that will survive the great levelling are the ones razor-focused on a very specific problem for a very specific customer, and who keep doubling down on it long after the conference crowd has moved on to the next shiny harness. Narrowness is not a limitation. It is the only place left where you can know more than the machine does, because the machine has been trained on the average and you have been trained on the particular. A company that understands the invoicing habits of mid-sized Swiss logistics firms better than anyone alive has a moat. A company that solves every problem for everyone has a brochure.
Then creativity, which is a polite word for the willingness to be wrong in an interesting direction. The consensus answer is now free. The value has migrated to the non-consensus answer that turns out to be right, and that requires judgement, which requires experience, which requires having been in the room when the last playbook failed. This is the part the conference founders were missing. Not intelligence, which they had in abundance, but scar tissue.
There is a neat inversion underneath all this. For twenty years the valuable skills were the hard ones, and the hard ones were technical. AI has made the hard skills soft and the soft skills hard. What is difficult for humans turns out to be easy for machines: code, Olympiad mathematics, the flawless memo. What is easy for humans remains stubbornly difficult for machines: reading a room, earning trust, holding a negotiation when the other side has stopped being reasonable. The moat, in other words, has moved to the parts of the business that were once dismissed as the soft bit.
The practical consequence is an uncomfortable one for most boards. The question is no longer whether you are using AI. Of course you are, and so is the firm across the street. The question is what you are doing that the firm across the street would never think to do, and whether you have the discipline to keep doing it when the consensus says otherwise.
Amplifier, not author
None of this is an argument against the technology. AI amplifies, extends and simplifies. It makes a good team faster and a focused firm formidable. It is the finest productivity tool ever handed to business, and anyone who refuses it on principle will be beaten by someone who did not.
But an amplifier only makes louder whatever you feed it. Feed it the consensus and you get the consensus at volume, which is what that conference hall sounded like. Feed it a sharp, particular, hard-won point of view and you get something nobody else can produce, because nobody else has your inputs.
The moat was never really the castle wall. It was always the people inside who knew exactly which hill they were defending and why. AI has simply removed the wall and made that rather obvious. Build the hill first. The machines will happily help you fortify it.
The strategies presented are thematic and do not constitute investment advice (or advice of any kind). No assurance can be given that the objectives of the investment above strategies will be achieved; the strategies involve risk (including, without limitation, illiquidity risk) and may incur a loss on some or all capital deployed. The opinions expressed, or indeed the information or assumptions that underpin them, may contain errors, mistakes, or omissions; no assurance or warranty can be made as to the accuracy or completeness of this information, and readers should not place any reliance on this content to execute investment decisions or for any other purpose. Readers accept full responsibility for using this content and are kindly requested to consult with their professional advisor before making any investment decision related to the same.
Comments