New LinkedIn data explains why

Local economic boosters have been predicting the imminent collapse of the Bay Area’s dominance of high-growth entrepreneurship for longer than the 20 years I’ve covered the debate. They just keep being wrong.

Surely, since spiraling housing costs and glaring wealth inequality boiled over in the pandemic, San Francisco’s time is up as America’s startup capital. Whoops! 

New LinkedIn data shows entrepreneurial activity in the San Francisco metro grew 153% over the past five years, among the country’s highest. Not the very highest, though — that was Dallas-Fort Worth at 192%, joined at the top by Austin at 173%. The new analysis reveals something counterintuitive about entrepreneurship’s geography: the hubs are still accelerating.

“I think this goes sometimes in ebbs and flows,” Sharat Raghavan, director of data science and research at LinkedIn, told me. “Certain markets get more attractive. But one trend in the post-COVID period is that people have realized you can start a business — even a technology-driven business — really anywhere.”

To be clear, entrepreneurship really has boomed everywhere. But it’s showing up differently in different places, like types of barbeque sauce and whether you call them hoagies (you should). 

One of the primary reasons entrepreneurship matters is that research shows nearly all net new jobs come from new firms. But as entrepreneurship has surged, new Kauffman Foundation research shows its job-creating effectiveness has shrunk. It’s a K-shaped entrepreneurial economy, baby. 

To date, big old economic hubs still hold structural advantages that matter, and can teach the rest of us something.

The founder-to-founder signal

Raghavan shared with me one metric that wasn’t part of LinkedIn’s public release: the percentage of new connections from founders that go to other founders.

Roughly 1 in 7 of the average San Francisco entrepreneurs’ Linkedin connections is to another founder. Austin, New York, Denver and Los Angeles aren’t far behind. By the time you reach Philadelphia, Washington DC and basement-dwelling Houston, it’s dropped to fewer than 1 in 10.

“It’s an indication of learning. They’re learning best practices, where their next customers are going to be,” Raghavan said. (Dallas-Fort Worth is the only exception of a big region that saw high-entrepreneurship gains, but relatively low founder-to-founder connections. No explanation for that yet.)

The gap between big regions widens further when you look at founders’ connections to investors. San Francisco has roughly twice the founder-to-VC connection density of other metros, a structural advantage that persists even as formation rates spread.

The implication: You can start anywhere, but your network still determines what happens next.

Gen Z entrepreneurs: Relying on AI and holding multiple jobs

LinkedIn’s research found that 68% of Gen Z entrepreneurs say AI and digital tools are important to their business — more than twice the rate of Baby Boomers (27%). Nearly one in five Gen Z founders said AI made starting their business feel possible. 

Where will they do their biggest building? Broadly, Americans are still moving to the Southeast and Southwest, and young people say for lifestyle reasons (like cheaper housing) – but big dynamic urban centers are continuing to attract Gen Z, like they did with Millennials before them. 

What economic development leaders have to care about is what they do when they arrive. If this Linkedin research holds, it looks like Philadelphia, Houston and Charlotte are attracting young people just fine. But they aren’t as economically dynamic as those young people going to Austin, Seattle and Los Angeles. Smaller regions and rural counties have similar variation: growing or shrinking in population doesn’t correlate neatly with the newest analysis on whether their entrepreneurial outcomes are growing or shrinking.

Might AI tip the scales? Raghavan sees this as the latest in a series of structural cost reductions. AWS in 2005 eliminated the need to buy server racks. COVID forced remote-first experimentation. Now AI is lowering barriers again.

“Imagine you’re a hotel owner,” he said. “People are using AI to analyze data on occupancy and pricing, to offer different rates — where before that analysis would take many hours. People are building websites, marketing collateral. Structurally, the cost of a lot of this has come down, especially for solopreneurs.”

But there’s wide geographic variation in AI adoption, and the question of whether AI favors new firms or incumbents remains open.

“In the United States, if you look at the history, it’s always been a pretty dynamic place,” Raghavan said. “Incumbents can’t move as fast as nimble startups. Entrepreneurs get close to the customer. They’re obsessed about product-market fit. They don’t have the luxury of a big revenue stream.”

He paused. “That’s the way I see it.”

LinkedIn’s data shows 73% of Gen Z entrepreneurs have multiple income streams (a twist on the gig economy), compared to 56% of Gen X entrepreneurs. People graduating from college now are likely to hold roughly twice as many jobs over their careers as someone who graduated 15 years ago.

This is sometimes called “stacking incomes.” But it can mask two very different stories.

“Some of it is born out of necessity — their jobs may have been affected,” Raghavan acknowledged. “But I don’t think that’s unique to this generation. Throughout times, if you look at the data, people holding multiple jobs has been as high or higher in other economic periods.”

What may be different is the form. The cost of forming an LLC has dropped. What once looked like multiple jobs now passes through an entity and shows up as entrepreneurship. The same underlying gig work gets relabeled.

“Maybe it’s not just one job anymore,” Kauffman Foundation researcher Robert Fairlie said in that new research. “People are piecing together income from multiple kinds of work.”

That’s not necessarily a triumphant story.

The funnel still narrows quickly

Six and a half million Americans started a business last year, but the optimism around increased formation rates deserves context. Most new businesses don’t become employers. Most don’t survive.

The question isn’t whether entrepreneurship is booming — it is. The question is what comes next.

“Whether these translate into actual employment gains is the real question we’ll have to track,” Raghavan said. “How many of these businesses actually go on and hire people and grow? That’s a fundamental question we’re really curious about.”

He offered a frame I’ve been thinking about since our conversation: “When the anecdotes and the data diverge, go with the anecdotes.”

I know people who started businesses because AI helped them bootstrap something they couldn’t have built alone. I also know people whose LLC is a fancy name for precarious gig work. Both show up as entrepreneurship.

What the geography data suggests is that networks still sort who gets which outcome. San Francisco founders connect to other founders and to investors at rates other metros can’t match. Dallas and Austin are growing fast, but connecting to the right people remains harder there. The Twin Cities, Charlotte and gobs of other midsized regions have too little economic coordination (a storytelling gap, I might say).

The cost of starting a company has dropped. The cost of scaling — in relationships, capital, and knowledge — may be as concentrated as ever.

“That connection rate kind of follows,” Raghavan said of the founder-to-investor gaps. “It does indicate that especially for businesses that need capital, they’re very, very good at reaching out to those players.”

Entrepreneurship is spreading. The infrastructure that turns starts into scale is not.