No manager, no majority: an agentic Vanguard

Pooling money was never the hard part. Amsterdam solved that in 1602. The hard part is steering it together, and there have only ever been two ways to try. You can hand the pool to managers, which is the VOC and the venture fund, and you inherit what economists call the principal-agent problem: the people deciding are not the people whose money it is, their interests overlap but differ, and the owners' only recourse is to write a letter. Or you can put every decision to a vote, and you inherit a worse problem, because under a vote a majority can do anything, including pay itself. Crypto ran that experiment at scale, and the record, which Vitalik Buterin summarized in 2021, is a history of treasuries emptied by whoever assembled 51 percent, and of votes that were cheaper to buy than to win. Until now, strangers had no third way to pool what they know along with what they own, and steer both, without trusting someone they had no reason to trust.

Now they do, and you can see the next financial revolution brewing in three new capabilities. AI agents can now do the work of investing, and this is the worst they will ever be.¹ But an unconstrained AI can no more be trusted with capital than a Spanish king with a loan, so the trust has to live somewhere else, and now it can. Decision markets let a fund's owners decide each proposal by trading on whether it will make the fund worth more, so that the people with skin in the game on a decision are the people who make it. That lets strangers govern what they own together with no manager and no majority in charge, and lets them govern an agent the same way. Internet finance makes ownership open to anyone and lets it move. Put the three together and you can build for the fund what the joint-stock company built for the enterprise: a fund organized around an idea about the future instead of around a manager, whose thesis is public, whose owners can be anyone, and whose owners steer it.

We call it agentic capital, and two things make us think it wins. The first is available on the first day. The structure expands access, cuts cost, makes ownership liquid and puts investors in control, and no partnership can match that without ceasing to be a partnership. The second has to be earned. An agent that is exposed to market feedback on every decision it makes, and to the knowledge of everyone who cares about its thesis, will compound what it learns in a way no partnership can, and communities of belief will gather around the vehicles that get it right.

In 1976 John Bogle launched the first index fund for ordinary investors. It raised $11.3 million of the $150 million he sought, and Wall Street called it Bogle's Folly. Vanguard now manages more than $13 trillion, and the index fund opened public markets to everyone at a fraction of the cost. Private markets are at their horse-and-buggy moment now, and what comes next is an agentic Vanguard: the same opening, for the companies that now grow up in private.

¹ I know this because I work in growth equity investing at the intersection of AI and healthcare, and the sourcing, the models, the memos and the diligence are being done by agents now, faster and cheaper than any associate.