Every few months, someone posts a Google Maps screenshot of Nvidia's campus in Santa Clara next to a skyline photo of Shenzhen, and the replies fill up with the same hot take: Silicon Valley, the economic capital of the world, should look more like this. Skyscrapers. Density. Ambition made visible in glass and steel. The argument mistakes the monument for the thing.

Skyscrapers didn't emerge from a city planner's vision of optimal density. They emerged from geography. The great vertical cities of the world — Manhattan, San Francisco, Hong Kong, Singapore — are almost all port cities built on isthmuses or peninsulas. The very thing that made them ideal trading hubs (water on multiple sides) also made outward expansion impossible. When you can't build out, you build up. Skyscrapers weren't an aesthetic choice. They were a workaround.

The skyscraper was only made possible by one invention. Not steel. Not glass. The elevator — the vertical car. Without it, buildings top out at about six floors, the limit of what most people will walk. The elevator made the 40th floor as accessible as the 4th, and in doing so created the economic logic for vertical construction.

This is why the Sun Belt is cheap. Houston, Phoenix, and Dallas can simply keep building outward. Urbanists call this suburban sprawl and treat it as a moral failing — but these are the same people who invoke markets and revealed preferences when it suits them, and the revealed preference is obvious: given a choice, people spread out. They want yards. They want quiet streets. They want to park a car.

Dubai has built one of the most dramatic skylines on earth and is, by any meaningful measure of human flourishing — free speech, rule of law, freedom of expression — a developing country. The towers are a costume. Architectural cargo-culting. You don't become a world city by building like one.

The Santa Clara Valley is not Manhattan, but it's not Houston either. It's something in between: a pseudo-peninsula. The San Francisco Bay constrains it to the north, the Diablo Range and Santa Cruz Mountains wall it on both sides, and the corridor narrows toward Gilroy to the south. Buildable land is genuinely finite.

So why not build up? Several reasons, none of them irrational.

The airport. Mineta San José International sits directly under the flight paths over downtown San José. Height restrictions are real and enforced. You could, in theory, relocate commercial traffic to Moffett Field, SFO, or OAK — but the market has consistently valued the airport where it is. That's not bureaucratic failure. That's a revealed preference.

The geology. Silicon Valley has actual silicon in it. Semiconductor capital equipment companies like KLA, Applied Materials, and Lam Research are still here. Optical benches and hydrofluoric acid fume hoods don't function well on the 50th floor of a vibrating steel structure. The industrial heritage of the valley shapes its physical form in ways that don't show up in a Google Maps screenshot.

The hills. The surrounding ridgelines aren't buildable at scale. Roads there are narrow, in constant repair from landslides, and a serious fire risk. This isn't NIMBYism — it's geology.

The market. Nobody is forcing Nvidia to stay in a suburban office park in Santa Clara. They choose to. Apple built a $5 billion campus that is, at its heart, a giant ring of offices around a park. They could have built a tower. They built a garden.

When Silicon Graphics collapsed in the early 2000s, Nvidia absorbed much of its engineering talent, becoming one of the direct beneficiaries of SGI's intellectual legacy. SGI's physical campus in Mountain View, meanwhile, was taken over by a small search company called Google. The Googleplex — now one of the most famous corporate addresses in the world — is a converted suburban office park.

The tradition runs deep. HP started in a garage in Palo Alto. Bell Labs — which produced the transistor, the laser, information theory, and Unix — worked out of a suburban campus in Murray Hill, New Jersey. RenTech, the most successful investment fund in history by a wide margin, operates out of a low-slung office park in East Setauket, Long Island. Innovation has never looked like a Mad Men set on Park Avenue. It looks like whiteboards and parking lots.

The economic data confirms what the anecdotes suggest. Silicon Valley's GDP per capita ranks among the highest of any metro area on earth — comparable to Luxembourg or Switzerland — produced almost entirely from single-story and low-rise campuses. Shenzhen generates impressive absolute output through sheer population scale, but on a per-worker basis it isn't close. The towers are not the source of the productivity. They're the Instagram of it.

OpenAI just opened a new office in Mountain View. Presumably they know something.

The vertical car enabled the skyscraper. Now consider what the horizontal car will do.

Waymo is already transforming San Francisco, and urbanists broadly love it — safer for pedestrians, smoother for cyclists, genuinely useful transit for people who can't or won't drive. But the fuller implication hasn't landed yet. A reliable, affordable, autonomous car available on demand is a horizontal elevator. It decouples where you sleep from where you work just as completely as the vertical elevator decoupled the floor you enter from the floor you work on. The friction of commuting — cognitive load, unpredictability, the requirement that you be sober and alert — disappears. Living 45 minutes from the office in a quiet suburb with good schools and a backyard stops being a compromise and becomes a genuine first choice.

Autonomous vehicles will be the single greatest enabler of suburban expansion in American history. The urbanists championing Waymo are, without realizing it, building the infrastructure that will eventually render their entire project obsolete.

This section won't convince anyone who's already decided. For a certain kind of urbanist, the question of why a place doesn't look like New York or Tokyo is already answered before it's asked: NIMBYism, car dependency, failed politics, moral cowardice. The skyline is the verdict and any defense is just rationalization. It's less a policy position than a religion, with density as the good and sprawl as the sin, and anyone who doesn't share the faith is either corrupt or in denial. If you're reading this to have your priors confirmed, you'll be disappointed.

For everyone else: none of this is an argument for arbitrary NIMBYism. Every neighbor should not have veto power over every new apartment building on their block. Silicon Valley should build more condos, more mid-rise housing, more density along transit corridors. Obstruction dressed up as community character is still obstruction.

But here's what you'd actually get if you swept all those obstacles away: probably something closer to Tuscany or the French Riviera than Tokyo. Low-rise, spread out, built around the landscape rather than against it — because that's what Mediterranean climates with finite buildable land and people who actually like living there tend to produce. The Santa Clara Valley was orchard country before it was tech country, and the bones of it are still there. You don't turn Provence into Shenzhen by removing a zoning board. The critics want to skip to the conclusion — Shenzhen's skyline — without asking why Shenzhen looks like that, which is because it's a tropical boomtown built on a swamp with a population of 18 million that needed somewhere to put them.

The deeper problem with the urbanist argument is that its central promise — build enough and prices will fall — runs into network effects. Metcalfe's Law holds that the value of a network grows proportionally to the square of the number of nodes. Cities work the same way. Every new person who moves to Silicon Valley increases the value of the network: more talent, more capital, more serendipity. That increased value attracts more people. Housing supply, by contrast, grows linearly. You build one unit, you house one household. You cannot build your way to Texas prices in a place people genuinely want to be, because building more and attracting more people simultaneously increases the desirability of being there. Manhattan has been adding housing for a century and remains extraordinarily expensive. The demand isn't a policy failure. It's the signal that the place is working.

This is not a reason not to build. It's a reason to be honest about what building accomplishes. Adding highway lanes doesn't solve traffic — induced demand fills the void — but it does allow more people to make more trips, and those trips have economic value. More housing in Silicon Valley won't make it affordable like Tulsa, but it will let more people participate in one of the most productive networks on earth, and that's worth doing. Just say that, instead of promising something the math won't deliver.

Before any of this — before Shockley Semiconductor, before the Traitorous Eight, before Fairchild and Intel and everything that followed — the Santa Clara Valley was called the Valley of Heart's Delight. The largest fruit-producing region in the world. Apricots, plums, cherries, in a Mediterranean climate that most of the world's great cities can only dream about.

The physical form of Silicon Valley is partly nature and partly a specific set of institutional choices. When Frederick Terman at Stanford began actively encouraging faculty and students to commercialize their research in the late 1940s and 50s, and when the university started licensing its intellectual property openly rather than locking it up, it created a feedback loop that Boston's Route 128 corridor — home to MIT and Harvard, with tighter IP norms and more insular corporate culture — never could replicate. The Valley's dominance over Boston wasn't weather or counterculture, though those were real. It was Stanford building an open ecosystem instead of a closed one, and that choice expressing itself physically as a sprawling, permeable, constantly cross-pollinating collection of campuses, parks, and garages.

Hong Kong, Shenzhen, and Singapore built vertically because they were land-scarce, tropical, and in some cases literally built on swamps. They didn't have much choice. Given options, people spread out. Manhattan's population density is lower today than it was at its 1910 peak, when immigrants packed the Lower East Side at over 300,000 people per square mile. That density was a product of poverty and constraint, not preference. The density people romanticize was what you got before you had an alternative.

And the alternative has costs of its own. Central Park exists because New York City seized and demolished Seneca Village, a thriving community of predominantly Black landowners, in the 1850s. "Build more" is not a values-free position. State-backed urbanist central planning has a history worth reckoning with before anyone reaches for the master plan.

So why doesn't San José look like Shenzhen?

Because of where it is, what it was built to do, and what the people who built it actually wanted. The geography constrained it. The industrial history shaped it. The institutional culture of openness and experimentation spread it outward rather than upward. And the people who live here, given every opportunity to choose otherwise, keep choosing oak trees and office parks and room to breathe.

Dubai built the tallest building in the world — just don't ask about the human rights record. Bell Labs invented the transistor, the laser, Unix, and information theory out of a suburban campus in New Jersey that looked like a community college. The monument and the achievement have essentially no relationship to each other.

Silicon Valley is the highest-value economic region on earth per capita, and it looks like a suburb. The critics want a city that looks like it's winning. Silicon Valley built things that actually won.