The Vanishing Middle
How The Agentic Economy Is Forcing You to Choose: Be Cheap or Be Irreplaceable
In 2014, Peter Thiel made one of the most influential claims in modern startup culture. “Monopoly is the condition of every successful business,” he wrote in Zero to One. He believed that the goal of every startup should be to build something so dominant, so deeply embedded in its market, that no one can compete. The logic is straightforward: capture a small market, lock it down, then expand from a position of total control.
The idea was elegant and influential. Zero to One became an assigned book in accelerator programs, the framework that shaped how founders thought about strategy. It gave them a single objective: to become the monopoly.
Thiel’s thesis rested on a structural foundation, even if he never named it directly. In 1937, the economist Ronald Coase explained that firms exist because using the market is expensive. Specifically, finding suppliers takes time, negotiating contracts takes effort, and monitoring quality takes resources. Firms grow because doing these things inside a hierarchy is cheaper than doing them through the open market. The firm expands until the cost of organizing one more activity internally equals the cost of carrying it out externally.
That is the boundary of the firm. And Thiel’s monopoly thesis is a bet on pushing that boundary as far and as fast as capital allows.
The methodology is clear: pour money into growth to build the platform, capture network effects, and make your coordination infrastructure so embedded in your customers’ operations that leaving costs more than staying. The moat is the switching costs, not the product. Once you achieve a monopoly, the position holds because the friction of leaving protects you.
For decades, this logic delivered. Enterprise software locked customers into years of configurations and integrations. Marketplaces became more useful as they grew larger. SaaS platforms held users through data gravity and workflow dependency. The monopoly thesis worked because coordination was genuinely expensive, and the company that owned the coordination infrastructure owned the market.
Then AI agents arrived.
The Monopoly Starts Leaking
AI agents are not another feature upgrade but a structural shift in the economics that underlies Thiel’s entire thesis.
The monopoly bet works only as long as coordination remains expensive, and that is precisely what agents are dismantling. They negotiate terms, reconcile documents, source suppliers, monitor quality, and route workflows across tools, all without requiring anyone to commit to a single platform. The friction that made monopolies sticky is dissolving.
This is already visible in the platforms that connect buyers and sellers. The value of a booking site or a shopping marketplace was never the interface; it was the reduction of search costs and the aggregation of supply, so the buyer did not have to look everywhere. That is a coordination function, and it is exactly what agents absorb. When a buyer asks an agent to find the best option across every platform simultaneously, loyalty shifts from the marketplace to the agent.
The platform still exists, but it recedes into invisible infrastructure, because the user’s loyalty was never to the interface itself, only to the convenience it provided. The agent now delivers that convenience without tying them to any single provider.
This is not a temporary disruption. When an agent can coordinate across tools without friction, the rationale for being locked into any single platform weakens; once that lock-in weakens, the monopoly Thiel described ceases to be a defensible position and becomes an expensive one to maintain.
Two Kinds of Speed
Thiel’s framework rests on a single assumption about speed: move fast, capture the market, and the position becomes unassailable. Your first-mover advantage leads to winner-take-all dynamics. This is what Silicon Valley later branded as blitzscaling.
But there are two kinds of speed operating simultaneously, and most founders are optimizing for the wrong one.
The first is the speed of technology, measured by how fast you can build a platform, acquire users, and integrate new capabilities. AI accelerates everything on this axis, making every function that can be described, standardized, and automated faster and cheaper. This is the axis Thiel’s thesis optimizes for.
The second is the speed of people, measured by how quickly you can build trust, deepen domain expertise, and establish a track record of standing behind your judgment when things go wrong. AI does nothing to accelerate this axis. Capital cannot make trust-building happen faster, domain expertise cannot be automated, and there is no shortcut to the years of consistent delivery that make a community willing to stake its own success on yours.
Thiel’s prescription is to optimize for the speed of technology: grow fast, capture the market, lock customers in. The problem is that when technology accelerates everything, ventures optimized purely for that axis get overtaken by the next wave. The moat they built out of coordination infrastructure dissolves when agents make that infrastructure unnecessary, and the monopoly falls not to a competitor but to a structural shift that makes the monopoly itself less viable.
The ventures that endure are those that build on the speed-of-people axis, accumulating deep domain expertise, real relationships, and a track record of accountability that cannot be commoditized because it cannot be accelerated beyond the pace at which human trust develops.
The Barbell
The two-speeds framework explains a pattern emerging across every major market: when the speed of technology and the speed of people diverge, value concentrates at the extremes and the middle hollows out.
At one end sits the algorithmic, operating at the speed of technology. It is ultra-cheap, high-volume, and commoditized, with AI agents handling coordination, thin margins, and scale as the only competitive lever. This is where surviving platforms compete relentlessly, because switching costs are low and the next agent can route around you tomorrow.
At the other end sits the curated, operating at the speed of people. It is trust-intensive, domain-specific, and built on expertise and accountability. Margins are healthy here because the value is irreplaceable, and scale is naturally bounded by the depth of relationships the founder and team can sustain.
The middle is collapsing. Ventures that were too slow to compete on cost and too shallow to compete on trust have been hollowed out, having tried to operate at both speeds and achieved neither. This is precisely where Thiel’s monopoly was designed to live, dominant enough to set prices and broad enough to serve the mass market, and it is exactly that ground that is disappearing.
What Replaces the Monopoly
If the monopoly thesis is expiring, what takes its place?
The answer is not another technology play, because every technology-based moat is subject to the same compression: if agents can route around your platform, they can route around the next one too. The replacement is not a better monopoly but a fundamentally different kind of advantage.
Nassim Nicholas Taleb calls it skin in the game. For founders, this means building a position where revenue, reputation, and relationships are concentrated in a domain narrow enough that failure is visible and recovery is personal. A venture built on skin in the game does not compete on lock-in; it competes on the depth of trust its community places in it.
This is not a lifestyle business argument. It is a structural argument about where durable competitive advantage lives when coordination costs approach zero. When an agent can find any supplier, compare any offering, and route around any platform, the only thing it cannot replicate is the judgment of someone who has consistently delivered and absorbed the consequences when things went wrong. An agent can evaluate data, but it cannot generate trust. Trust is a human decision, earned through repeated exposure to someone’s judgment over time, and no amount of computational power can compress the years it takes for one person to decide another is worth relying on.
That is a moat that is slower to build, cannot be funded into existence, and is the only one that holds when the economics underneath Thiel’s thesis shift.
The Implication for Founders
Thiel framed the choice for founders as a binary: build something no one can compete with, or resign yourself to fighting over scraps. Monopoly or commodity.
There is a third option he did not account for, because the structural conditions of his era did not demand it. I call it Momentum Scaling. Momentum Scaling is the deliberate choice to build toward the irreplaceable end of that barbell — through domain depth, earned trust, and a track record of accountability that no agent can replicate. You build something that is neither a monopoly nor a commodity, something that holds its position not through lock-in but through the depth of trust and expertise earned in a specific domain over years of consistent delivery.
If your strategy depends on capturing a market and locking customers in through switching costs, you are betting that coordination stays expensive in your domain long enough to recoup the investment. That bet is getting worse every quarter.
If your strategy is to build deep expertise in a narrow domain, earn the trust of a specific community, and create a position where your ecosystem depends on your judgment, you are building on the axis that AI cannot compress. It will take longer, and it will not produce the hockey stick chart that impresses a Series A investor, but it will produce something more durable: a moat that strengthens as technology accelerates everything around it. The more that agents commoditize the algorithmic end of the market, the scarcer and more valuable human judgment becomes at the curated end, and every wave of automation that makes the commodity layer cheaper makes the trust layer worth more.
Zero to One was written for a world where coordination was expensive, and the fastest builder won. That world is ending. What replaces it favours the patient, the focused, and the accountable.
This essay was inspired by Howard Yu’s Substack essay “Coase vs Claude and the Future of the Firm”
Davender’s passion is to guide innovative entrepreneurs in developing the clarity, commitment, confidence and courage to enter, engage and lead their markets in a world that refuses to hold still, by thinking strategically and acting tactically.
Find out more at https://coachdavender.substack.com/about and https://linkedin.com/in/coachdavender