S2. The Ultimate Scarcity

When buildings can be constructed for pennies, does value vanish—or sink into the soil beneath? AGI fundamentally inverts real estate value: structures get cheaper, land gets pricier. Automation makes building trivial but can't create more prime locations. This drives a massive potential wealth transfer to landowners unless policy intervenes via land value taxes or redistribution.

   [AI can’t make land](#ai-cant-make-land)

   [Winners by location](#winners-by-location)

   [The new land rush](#the-new-land-rush)

   [Policy shapes place](#policy-shapes-place)

   [Next 5 years (2025-2030)](#next-5-years-2025-2030)

   [Medium term (2030-2045)](#medium-term-2030-2045)

   [Long term (2045+)](#long-term-2045)

   [Structures depreciate, location appreciates](#structures-depreciate-location-appreciates)

Buy land, they’re not making it anymore.

— Mark Twain

If you’re enjoying this deep dive and want to keep up to date with my future writing projects and ideas, subscribe here.

Real estate remade

When people talk about AGI, they usually think about how it will change jobs, wealth, and technology. But it’s also going to change something much closer to home: where we live, what our homes are worth, and what land costs.

Housing and land prices move differently when the world changes fast. Structures are easy to build if you have enough machines. But land is fixed. No matter how smart AI gets, it can’t make more coastline, more city centers, or more sunny valleys.

In this section, I look at how housing and land prices might shift during an AGI takeoff. It’s not one story, but a series of stages: first a period of rising rates and mild automation, then a turbulent middle as AGI hits, and finally a world where AI does almost everything and land becomes the ultimate bottleneck.

The main thing to watch is this: over time, houses get cheaper, land gets scarcer. And that changes almost everything about how real estate works.

There are a few ways this story could get complicated. Governments might step in with heavy land taxes, or widespread UBI could reshape who can afford what. And if space colonization ever really takes off, it could blow the scarcity model apart¹. But unless something big like that happens, land looks set to become the main store of wealth in an AI world.

Timelines, side by side

As economist Henry George observed in the 19th century, technological progress ultimately concentrates value in fixed assets that cannot be replicated—a principle that AI will very likely amplify to new degrees.

Here's how this divergence unfolds across three critical periods:

2025-2030 (Partial AI)

• Buildings: Flat to mild growth, constrained by higher interest rates. Limited automation means supply remains tight. Most people still afford housing, though affordability worsens.
• Land: Mixed performance with slight overall growth. Remote work shifts demand from prime urban land toward secondary markets and rural areas. Land remains about half of property value on average.

2030-2045 (AGI Takeoff)

• Buildings: Volatile and bifurcated. Automation-driven job losses reduce housing demand for middle/lower classes while elite wealth drives luxury housing in tech hubs. Construction automation creates cheap new supply. Housing becomes more affordable relative to income as structures' value erodes.
• Land: Strong upward trend with significant variation. AGI makes intelligence abundant, so land becomes the scarce factor. Enormous new projects create soaring demand. Prime locations in thriving regions surge in value while industry-dependent areas collapse. Land's share of total property value grows substantially.

2045+ (Post-AGI Economy)

• Buildings: Low/commodity value. Structures can be produced at near-zero marginal cost through advanced automation. Housing no longer appreciates; most structures decline in real value or are simply replaced as needed.
• Land: Becomes the key source of value. With all else abundant, land emerges as the main store of value and basis of economic rent. Prime locations command extremely high prices unless curbed by policy. Land in less desired areas may have minimal value if population concentrates elsewhere.

As structures democratize through technological abundance while location stratifies through intensified scarcity, we face questions about spatial justice and the meaning of place in human experience.

The resolution of these questions—through policy intervention like land value taxation or through market concentration of spatial privilege—may determine whether AI's spatial change liberates us from housing scarcity or intensifies territorial inequality.

The big shifts

Agriculture made arable land the primary asset. Industrialization privileged capital and machinery. The information age rewarded intellectual property and knowledge.

AI will trigger an inversion of economic value: as it makes human intelligence and labor abundant, scarcity shifts to what cannot be created even by the most advanced minds, natural or artificial—physical territory. This will remain the case until we become a multi-planetary species.

AI can’t make land

There's something fascinating about how AI eliminates scarcity in intellectual domains. As it does this, the bottleneck of production inevitably shifts to irreproducible resources. As economist Tyler Cowen² puts it simply: "First is land."

With AI unlocking "hundreds or thousands of brilliant new ideas," what limits us won't be human creativity but physical space—territory for new industries, energy generation, computing infrastructure, and applications we haven't even imagined yet.

This echoes what economist Henry George observed³ more than a century ago: technological progress ultimately concentrates wealth in fixed assets like land. AI accelerates this principle. Recent formal modeling by economists Anton Korinek and Joseph Stiglitz⁴ confirms this pattern. Their 2024 report projects that while technology-driven abundance reduces costs for most goods and services, the prices of irreproducible resources⁵—land, energy rights, natural resources—will appreciate, potentially concentrating AI's tremendous productivity gains among those who control these fixed assets.

Winners by location

This appreciation won't distribute evenly. Established technology hubs⁶ will likely experience soaring land values as AI development clusters expertise and capital. Meanwhile, cities without technological specialization—places like Hartford or Minneapolis—may decline as automation eliminates traditional employment⁷ without creating equivalent replacement opportunities.

Brookings Institution research confirms this likely divergence, projecting that AI's impact will concentrate in major metropolitan areas with substantial knowledge-economy sectors. This pattern threatens to accelerate the hollowing of small cities and former industrial centers. Globally, this could further widen disparities⁵ between countries leading in AI development and those relegated to technological dependence.

The new land rush

Computing infrastructure is driving an accelerating land rush in regions with abundant power and space. Places like Northern Virginia⁸ and Texas⁹, where data centers already consume vast territories, will likely see this trend intensify as AI's computational demands grow exponentially.

Meanwhile, traditional commercial properties¹⁰ face a testing moment. Office space faces¹¹ existential threat as automation eliminates routine knowledge work, while specialized facilities for vertical agriculture, logistics, research, and entertainment appreciate.

We're heading toward a bifurcated future where mundane office space becomes obsolete while warehouses, research facilities, and experiential venues gain value in an economy where physical presence becomes either unnecessary or uniquely valuable.

Policy shapes place

As land becomes the primary repository of AI-generated value, policy choices become consequential. Will this change serve broad prosperity or concentrate advantage? Korinek and Stiglitz¹² advocate land value taxation as a particularly efficient mechanism for redistributing AI-driven gains—capturing the zero-sum component of land appreciation without distorting productive incentives.

Some propose using these land value taxes to fund universal basic income¹³ programs, effectively recycling technology's productivity gains as broadly shared prosperity. Cowen emphasizes the critical importance of zoning reform and permitting streamlining, arguing that rigid land use regulations² could become civilization-limiting bottlenecks if they prevent development at the scale and speed AI enables.

Alternative technologies like telepresence, virtual reality, and new modes of transportation might theoretically reduce location premiums, but most analyses suggest physical proximity, natural advantages, and regulatory frameworks will maintain land scarcity¹³ for the foreseeable future.

As artificial intelligence transforms the economy, land becomes the new gold—the ultimate store of value in a world where intellectual capital becomes abundant. Whether this change widens prosperity or concentrates it will depend less on technological inevitability than on the governance choices we make in response.

Structures depreciate, location appreciates

The structures we erect will steadily lose value while the earth beneath them becomes precious¹⁴. This may alter how we understand property, wealth, and belonging in the age of AGI.

In the near term—these next five years—the change will begin subtly. Interest rates and early AI applications will soften housing markets while land values fluctuate across regions. But when AGI emerges, this gentle tide becomes a tsunami: automated construction will render buildings abundant and affordable, while simultaneously intensifying competition for prime locations¹⁵.

Consider Sydney's harbourside suburbs—places like Point Piper or Vaucluse. A waterfront mansion there today commands astronomical prices for both its structure and position. In the AGI era, even the most opulent residence could be fabricated in weeks at a fraction of current costs, yet the finite stretch of Sydney Harbour shoreline remains irreplaceable. We may witness the once-unthinkable: a $20 million home whose physical structure accounts for less than $2 million of its value, with the rest attributed solely to those irreproducible coordinates overlooking the Opera House and Harbour Bridge.

This divergence faces countervailing forces at each stage. For housing, rising interest rates and potential job displacement push prices downward, while supply constraints and wealth concentration pull them up. For land, the scarcity principle suggests appreciation, especially as AI cheapens all other production inputs¹⁶. Yet this isn't guaranteed—policy interventions, climate change impacts, or even paradigm shifts like immersive virtual worlds could undermine land's primacy. What seems certain is that our grandchildren will inhabit a world where structures are seen as depreciating consumer goods while land remains the bottleneck and repository of value.

In conclusion, the advent of advanced AI and AGI is likely to turn the real estate equation upside-down: structures trend toward commodity pricing, while land trends upward in value. Policymakers must rethink housing and land frameworks to ensure that AI-driven abundance benefits society broadly, not just landholders. Understanding this dynamic is crucial for preparing for the AGI takeoff's impact on real estate.

Coming up next: When work and want disappear, what fills the void inside us?

Continue reading →

Want me to send you new ideas?

If you’re enjoying this deep dive and want to keep up to date with my future writing projects and ideas, subscribe here:

Subscribe to future updates →

Footnotes

1

Geographic constraints endure regardless of technological revolutions.

2

Martian homesteads replacing coastal villas? Musk's vision could be the ultimate wild card in Earth's real estate game if SpaceX cracks interplanetary living by 2045.

3

Solar farms, vertical farms, data centers—they all need dirt. Cowen's analysis explains why physical space becomes more valuable as everything else gets digitized.

4

Written in 1879 yet eerily prescient for our AI moment. George argued technological progress inflates land values and concentrates wealth in the hands of property owners rather than spreading benefits widely.

5

Rising land prices alongside falling wages? The economists model how AI's labor displacement creates this exact dynamic. Their proposed land tax solution faces massive political resistance—landlords vote, robots don't.

6

Non-scarce. information, manufactured goods, and most services. Increasingly scarce. land, energy, and food. NBER's projection warns that without policy intervention, property owners could capture most of AI's economic gains.

7

Silicon Valley thrives while Rust Belt cities empty out. Cowen predicts automation's uneven impact will dramatically redraw America's real estate map over the next decade.

8

White-collar metros hit hardest by job disruption? Brookings data challenges conventional wisdom about automation primarily threatening manufacturing regions.

9

Parekh documents how Loudoun County land prices are already spiking due to AI infrastructure demands—the canary in the coal mine for tech's reshaping of property markets.

10

Some Austin-adjacent plots doubled in value last year alone. Realtor.com reports Texas' cheap power and open spaces are creating an AI-driven land rush unlike anything seen since the shale boom.

11

Less office space, more labs and warehouses. JLL foresees a massive commercial real estate shift as AI redefines work patterns and physical infrastructure needs.

12

Empty cubicles as chatbots handle clerical tasks? Oxford Economics identifies office space as AI's biggest real estate casualty, while biotech labs could become tomorrow's premium properties.

13

Combine land taxes with UBI and you might democratize AI's economic benefits. Ord's proposal makes theoretical sense but faces the eternal challenge. convincing governments to tax wealthy landowners.

14

Your home's structure will depreciate; the land beneath it will not. Davis' study shows land already dominates property values in most markets—a trend AI will only accelerate.

15

Land values track economic growth—a relationship AI will accelerate exponentially. As wealth concentrates, Davis shows prime plots could become the ultimate positional good in a post-scarcity world.