Twice a week, Jensen Huang gathers around sixty of Nvidia’s most senior executives in one room and runs one large meeting that seems to violate every corporate rule. No agenda item is sent, no pre-briefing, no follow-up one-on-one. Huang, who runs the company Forbes just ranked third on its new Most Innovative Large Companies list , has explained that he skips one-on-ones with his direct reports on purpose, because they would hand a handful of people private access to information the rest of the room needs. “I don’t have one-on-ones with them because it’s impossible,” he told Entrepreneur . “We present a problem, and all of us attack it.”

That is not how a $5 trillion company is supposed to run its top ranks. It is how a ten-person startup runs its Monday stand-up.

Forbes measures innovation on this list with what it calls an innovation premium: the share of a company’s market value that investors are betting on innovation it has not delivered yet, a methodology built by Innosight’s Nathan Furr and Jeff Dyer after they found that R&D budgets and patent counts predicted almost nothing about which companies kept innovating. Size was never the variable. Speed and appetite for the next bet were. And speed is exactly what shrinks as companies grow, unless they fight it on purpose, the way Huang does.

According to Glilot Capital Partners’ State of Micro-Unicorns report , the fight used to require deep pockets. Companies that reached a $1 billion valuation before 2022 needed a median of 400 employees to get there. The 26 companies in Glilot’s dataset that crossed $1 billion after 2022 needed a median of 34, generating $118 million in value per employee against $5.5 million for the typical unicorn, a 21-fold gap. Time to $1 billion has compressed from 5.1 years to 1.7. The advantage that used to be reserved for whoever could afford to run lean at scale is now available to whoever has a laptop and a subscription.

Why Do The Biggest Companies Act Small?

Amazon, ranked sixth on the new list , built this principle into its org chart two decades ago. Jeff Bezos’s two-pizza rule holds that no team should be larger than two pizzas can feed, in practice under ten people, each team owning one product end to end instead of routing decisions through layers of approval. Amazon has said the model keeps coordination overhead from creeping in as headcount grows, and it is not a sentimental choice: teams under ten people report engagement of 42 percent or higher, compared with under 30 percent in larger groups, by the company’s own data.

Mastercard, ranked tenth , takes the same logic a step further and points it outward. Instead of trying to originate every idea inside its own walls, its Start Path program , ten years old this year, has run more than 400 startups through it, which have collectively raised over $25 billion in the capital that followed. “We’re not looking for founders who are trying to build the next buzzy consumer brand,” Sabrina Tharani, who leads the program, has said, “but rather building deeply important technologies that are solving for real pain points in our industry.” A company with Mastercard’s balance sheet could try to build all of that internally. It has decided a ten-person team outside its walls will usually get there faster than a two-hundred-person team inside them.

What Happens When AI Erases Team Size?

The Glilot report names its outliers directly. Inflection AI reached a billion-dollar valuation with seven employees, or $176 million in value per person on staff. Safe Superintelligence did it with ten. Cursor got there with 29. Cognition with 50. When OpenAI bought io Products for $6.5 billion, the company had 55 employees on the payroll, or $118 million in exit value per person. None of that required exceptional luck. It required an AI stack that used to demand a department now running on a laptop.

Annie Liao felt the same shift at a smaller scale. The 24-year-old founder of the AI startup Build Club chose to go it alone after watching a previous employer’s three co-founders slow every decision down by leaning on each other for consensus. “I love that I can take risks and move fast,” Liao wrote in a personal essay published by Yahoo Style . “I have a very clear vision and mission for my startup, so being a solo founder allows me to make quick decisions and keep us moving in a streamlined direction.” She is describing, without the corporate vocabulary, exactly what Amazon spent two decades engineering on purpose.

Where Do Ten-Person Teams Actually Win?

None of this means a ten-person company is about to out-earn Eli Lilly, which topped the new list on the strength of a drug pipeline that took a decade and billions of dollars to build. Certain kinds of scale still cost what they cost. But the specific currency Forbes is measuring, an innovation premium built on the market’s confidence that a company will keep reinventing itself rather than just defend what it already has, has only ever come from three things: fast decisions, direct ownership of the outcome, and outside ideas let in the door. A small business does not need a Start Path program or a flat org chart memo to get any of that. It already has it, by default, as long as its owner does not hire it away by adding the very layers that make big companies slow.

I don’t have one-on-ones with them because it’s impossible. We present a problem, and all of us attack it. Jensen Huang, CEO, Nvidia

That is the calculation worth running before assuming the safer move is staying inside a company large enough to make this list. The same tools compressing headcount at Inflection AI and Cursor work exactly as well for the manager sitting three approval layers from her own idea inside one of these eighteen companies right now. The only real question left, for a small business owner or for the person weighing whether to become one, is whether they would rather compete for a seat at the table these companies built, or build the smaller, faster table these companies are already trying to copy.