The tiny-team, billion-dollar company is not a prediction. It is already a proven model.

Elad Gil, the investor and entrepreneur, places a historical anchor where most people put a prediction. The tiny-team, multi-billion dollar company, he argues, is not a coming AI phenomenon. It has already happened. “Minecraft was like what was it five people, 10 people when it was bought for billions of dollars by Microsoft,” Gil says. “People keep talking about someday there will be like a multi-billion dollar single person company. That was basically Minecraft roughly. It already happened like 15 years ago or whenever that was.”

Gil’s point is not that the Minecraft case was common. It is that the assumption behind most objections, that a company worth billions requires a large organization to build it, was already wrong before the current generation of tools existed. None of what made Mojang valuable required institutional headcount. It required a product people wanted and a team disciplined enough not to bloat.

What has changed is replicability. Mark Cuban puts a concrete number on the shift: he has two or three people building software that, five years ago, would have cost roughly two to three million dollars a year. That is not a rounding error. It is the difference between a project that requires institutional backing and one that does not.

Minecraft was like what was it five people, 10 people when it was bought for billions of dollars by Microsoft. People keep talking about someday there will be like a multi-billion dollar single person company. That was basically Minecraft roughly. It already happened like 15 years ago or whenever that was. Elad Gil

Patrick Collison, Stripe’s co-founder and chief executive, offers a product-level example. Stripe’s internal artificial intelligence knowledge tool, Kai, was built by two people in approximately six months, by Collison’s account. He also points to a structural signal in Stripe’s own data: year-on-year usage growth of Stripe Billing is running significantly higher than usage of Stripe overall, because it disproportionately captures newly created software companies. The formation rate of software businesses appears to be accelerating, and the billing infrastructure is registering it before most other indicators do.

Anton Osika, the chief executive of Lovable, describes the process at the individual level. Employees at large companies discover tools like Lovable through colleagues, build side projects outside their day jobs, and some of those projects reach hundreds of thousands of dollars in revenue before the person becomes a founder. The on-ramp to company formation, in Osika’s observation, now runs through the corporate workplace rather than despite it.

Quinn Slack, who leads Sourcegraph, frames the organizational consequence most directly. He describes seeing companies around his size beginning to hire product managers, marketers, and similar overhead roles, and calls it “the old way of building a software business.” His concern is structural: the historical staffing pattern, where roughly 10 percent of a company’s people focus on building a great product and 90 percent manage overhead around selling and operating it, was a function of necessity. Slack’s argument is that agents dissolve most of that necessity, and companies that hire into the old ratio will be outcompeted by those that do not.

Taken together, these are not predictions about what AI might eventually do. They are descriptions of what small teams are doing now, measured against a baseline that Minecraft established before any of the current tools existed. The open question is not whether a five-person team can build something worth billions. It is whether the institutions and investors organized around larger headcounts will recognize the shift in time to respond to it.