Peter Thiel: AI Monopolies Will Be Fewer Than Internet Monopolies

PayPal co-founder Peter Thiel, during the Axel Springer Award ceremony in Berlin, told The Times that monopolistic companies created by the AI revolution may be fewer than those in the internet era; a significant portion of the value created by technology will flow to consumers and be distributed more evenly. He compared this to the internet boom of the 1990s, where there were massive returns to scale and giant platforms that dominated the market; he doubts that AI will ultimately lead to the opposite direction—while the market can be large, competition can be extremely fierce, resulting in fewer monopolistic traits.

This contrasts with his famous business formula. In 2014, he wrote in The Wall Street Journal that "Competition is for Losers": to create and maintain lasting value, one should build monopolies rather than engage in undifferentiated commodities. Airlines create significant value but hardly make money, while Google creates less value but has a profit margin an order of magnitude higher. In "Zero to One," he describes successful companies as unique, solving distinct problems to achieve monopolies, while failed companies do not escape competition. Proprietary technology must be at least ten times stronger than alternatives in some dimension, or it will only be seen as a marginal improvement.

He identified large platforms as the standard answer for the internet. Search, social media, and e-commerce generate increasing returns on users and data, with advertising pricing power remaining with the platforms. In the realm of AI, model capabilities are rapidly diluted by open-source weights, price wars, and the ability to replicate features within weeks, allowing the application layer to sell the same capabilities to everyone. Thiel's new judgment does not deny market size but challenges the notion that "scale automatically leads to monopoly."

He did not provide a list of companies that would be eliminated, nor did he retract the narrative of monopoly for companies heavily invested in by Palantir and Founders Fund. This statement serves to shift the investment clock from "finding another Google" to "value may leak to user bills." For laboratories still financing under the winner-takes-all model, this means recalculating exit multiples: massive capital expenditures may not yield a decade of pricing power but could result in cheaper reasoning and efficiencies taken by customers.

In the same week in Berlin, he also discussed how young German billionaires rely solely on inheritance and how Trump's second-term appointments are loyal but not smart. The reduction of AI monopolies is another slice of the conversation regarding technological structure: the state side lacks new scalable companies, and the industry side may also lack new scalable platforms. Both sides are asking the same question—can incremental value still be locked into a few balance sheets?

In market mechanisms, buyers are enterprise customers and consumers seeking cheap reasoning and interchangeable models, while sellers are model companies and cloud vendors still financing under platform monopoly valuations. The event-driven aspect is that open-source and price wars are pushing capability differences down to interchangeable levels. Capital flows from venture capital and cloud contracts into training clusters, then leaks to the usage side at lower unit costs. Beneficiaries are customers who can integrate AI into workflows like utilities; those under pressure are laboratories that must prove billion-dollar computing bills with monopoly premiums and funds that treat "Zero to One" as a fundraising bible but encounter the script of "value leaking to users."

Source: Public Information