AI might not kill the big companies. It might actually save them.
The startup model is breaking. Here is what that means for the companies it was supposed to disrupt.
For the past two years, corporate leaders have been told the same story. Move fast or get eaten. Startups are coming, and AI is their weapon.
That story is not wrong. But it is incomplete.
Because while everyone has been watching startups accelerate, something else has been happening at the same time. Quietly, the startup model itself has started to strain. And that shift may end up changing the balance between startups and incumbents more than people currently expect.
Not because the threat from startups has disappeared. It has not. But because the economics of building new companies are evolving in ways that unexpectedly favor some of the assets large organizations already possess.
To see why, it helps to zoom out.
What is actually happening to startups
When you step back from the daily noise, three changes are happening simultaneously. Most people treat them as separate issues. In reality they reinforce each other.
The first is that building software has been dramatically commoditized. AI tools now make it faster and cheaper than ever to create products. On the surface this sounds like great news for startups.
But when everyone can build, everyone does, and everyone will.
The barrier and cost to creating a product is collapsing. The barrier to being noticed, chosen, and retained is not changing nearly as much. If anything, it is becoming significantly harder.
The second is more uncomfortable to say out loud.
Most of what is being built right now, especially using AI, are tools, not businesses. And there is a big difference.
Think of it this way. A builder gets paid to solve a specific problem for a specific customer. The tool is what he uses to do it. A hammer does not care who picks it up. It works for any builder, on any job. That is its strength — and its problem. Because if anyone can use it, anyone can sell it.
A business, when done right end to end, grows by itself. Software worked as a business because you could build it once and sell it a million times. The more customers you added, the more profitable you got. That is what made venture capital work — investors funded the early losses because the growth curve justified it.
Tools do not work that way. Someone builds a tool. Someone else copies it. Then ten more people copy it. The price drops. The margins disappear. You are back to competing on acquisition cost.
Most AI products being built today are tools. But they are being funded, priced, and measured like software businesses. That gap between what is being built and how it is being valued is something I am not sure enough people are talking about yet.
The third shift is something operators have been feeling for a while. Customer acquisition is not getting easier. In many cases it is becoming harder.
Sam Jacobs pointed out in his latest newsletter that Customer Acquisition Costs (CAC) payback periods in parts of the SaaS market have stretched to 35–36 months, while sales and marketing still consume more than a third of revenue.
More companies are competing for the same attention with increasingly similar products. Acquisition cost is eating into whatever efficiency gains AI was supposed to deliver.
Put these three together — commoditized building, tools replacing software businesses, and ever more expensive customer acquisition — and the picture becomes difficult to look at directly. Startups are building faster into more crowded markets, spending more to acquire customers they struggle to differentiate for, while chasing growth metrics that the underlying business model can no longer reliably deliver. All of this while investors are struggling to figure out and then see the implications of new venture math.
Capital is moving. Intentionally or unintentionally — not away from AI, but toward the infrastructure and foundation models that power it. Away from the application layer where most startups live. The funding environment for the average AI startup is not what the headlines suggest.
This is not a temporary correction. It is a structural shift. And it would create a vacuum that most people perhaps have not yet named correctly.
The hand corporates are holding
This is where I want to pause and say something that feels almost contrarian right now.
Large companies — corporates, established players, whatever you want to call them — are sitting on a better hand than most people realize. Not because they are smarter or more innovative. History and innovation case studies are pretty clear on that. But because the game has changed in ways that favor exactly what they have.
Three things specifically.
The first is cash. Corporates have revenue coming in. Not runway — revenue. That distinction matters enormously right now. A startup needs to prove its model, convince investors, and show growth metrics on a timeline that is rarely aligned with how markets actually develop. A corporate can fund experiments without a pitch deck. It can absorb failures that would kill a startup. It can wait. In a moment where finding the right business model requires running many tests, the ability to fund those tests without external pressure is not a small advantage. It is perhaps the decisive one.
The second is time. Closely related but distinct. The venture capital clock is a brutal forcing function. Eighteen months to show traction. Three years to prove the model. Five years to exit. That clock is misaligned with the pace at which genuinely new markets develop. Corporates don’t have that clock. They have a different kind of pressure — quarterly results, internal politics, shareholder expectations — but they might have more calendar time to let something develop before pulling the plug. In the current moment, that time is worth more than most corporate leaders appreciate.
The third is existing customer relationships. This one is underestimated almost everywhere I look. Startups focus on every new acquisition with a high CAC, working hard to make the business viable one customer at a time. Corporates already have them. They have trust, contracts, data, and access. Building something new on top of an existing customer relationship is categorically different from building something new and then having to convince a skeptical market to try it. The distribution problem — the hardest problem in any new venture — is already partially solved.
Cash. Time. Customer relationships. In the game that is emerging, these are the assets that matter most. And they sit overwhelmingly with the corporates.
But — and this is the part that bothers me and should bother corporates — having the best hand is not the same as playing it well.
The window
So what does this moment actually offer corporates?
Not a guarantee. Not an automatic advantage. A window. And windows close.
During these years of struggle and uncertainty, the pressure on startups to survive will consume energy that would otherwise go into disrupting corporates. That is the window. Not permanent relief. Breathing room. Time to move deliberately while the challengers are preoccupied with their own survival.
But breathing room is only useful if you see it for what it is. And I am not sure most corporates are looking in the right direction.
Most are competing harder and harder for the same customers with increasingly similar offerings, spending more to acquire and retain, and calling it growth. The map they are using is not wrong. It is just zoomed in too close to see the whole board.
The real strategic question
Every business plays on the same board. The intersection of supply and demand.
On one axis, demand. How many people want what you are offering. On the other, supply. How many others are offering something similar.
The goal is simple. You want to be where demand is high and supply is low. Where many want what you have and few can provide it. That is where margins hold, customers choose you without being convinced, and growth does not require outspending everyone else.
The problem is that competitive pressure always moves in one direction. Toward the crowded corner. High demand, high supply. More competitors, more noise, more acquisition cost. Everyone technically competing, nobody really winning.
That is precisely where the AI application layer is today. Thousands of tools, competing for the same buyers.
Corporates are not immune to this pull. But they have something startups increasingly do not. The time and resources to choose a different direction.
The hardest part
Real differentiation in an AI-driven world requires three things.
First, creativity that cannot easily be commoditized. Not the creativity of slogans and campaigns, but the ability to see a customer problem more completely than others do and build around that insight.
Second, the courage to specialize. That often means deliberately focusing on specific customers and letting go of others. Large organizations find this difficult because the instinct is always to broaden and capture more of the market. Yet breadth often leads straight into commoditization.
Third, genuine ownership of the customer journey. Not as a data project or CRM initiative, but as the strategic foundation of the company. If you do not understand how customers actually experience your product or service, every moment of friction and every moment of value, you cannot improve it deliberately.
And without that depth of understanding, differentiation rarely lasts once competitors arrive.
The one thing AI cannot commoditize
There is an interesting implication hidden inside all of this.
AI can commoditize building. It can accelerate experimentation. It can even commoditize parts of knowledge.
But it cannot replace judgment.
The judgment about where to play. The courage to commit to a position. The continuous human work of understanding customers better than anyone else.
Those capabilities remain stubbornly human.
Which means they may become the most valuable capabilities companies can develop in the years ahead.
The hand is there.
The window is open.
The real question is who will recognize the moment and choose to play differently.
I'm writing a book about the pattern that puts startups and new ventures in the valley of death, and the window before each capital raise where the direction can still be changed cheaply. The book is currently titled- The Physics of Startup Failures.
Early access as an ebook will be available before it's out. Sign up here.