ETX Sightline - Compounding Scarcity

Land is the one input the Tangible Economy depends on — and the one input AI still can't read

Every commodity, megawatt, structure, and supply chain begins the same way: as a decision about physical space. Land and natural resources are the most upstream input in the economy, the layer everything else is conditional on.

Case in point: the pace of AI development is now throttled by community meetings and zoning board debates about appropriate data center locations, rather than the balance sheets of the hyperscalers or sophistication of their frontier models.

And yet land is the last major asset class without machine-readable identity. We have indexed securities, shipments, and sensors — down to the tick and the pallet — but not the ground beneath them. The one input everything else depends on remains largely AI-illegible to the systems now making decisions about it.

That gap would be a manageable inefficiency in a world where land and resources were becoming more abundant, or even just holding steady.

They aren't.

Scarcity is compounding from three directions at once:

1. Demand is rising with population growth and electrification — more people, and each of them drawing more power, water, and material through the economy than the generation before.

2. Usable supply is contracting as climate change withdraws land through fire, flood, drought, erosion, and water stress, converting acreage that used to be developable, farmable, or insurable into acreage that isn't.

3. And access is narrowing as land and the resources beneath it become instruments of geopolitical competition — export controls, resource nationalism, and contested supply chains turning what used to be a market transaction into a strategic one.

Demand up, usable supply down, access tightening: the value of every remaining acre, and the cost of every wrong decision about it, are rising together.

The market underneath the thesis

This isn't an abstract concern. Three markets sit directly on top of this resource dynamic, and each is already being counted in the trillions:

• Natural resources, the baseline case. The commodities and materials that underpin everything else still have to come out of the ground, and doing that at the pace the next decade demands is itself a multi-trillion-dollar undertaking. The Future Minerals Forum puts the investment required to bring critical minerals — copper, lithium, nickel, cobalt, rare earths — to market at $5.4 trillion globally by 2035, before a single downstream factory or grid connection is built.

• The infrastructure revolution. Energy and compute are being rebuilt simultaneously, on overlapping footprints. The IEA's World Energy Investment 2026 puts total global energy investment at $3.4 trillion for the year — $2.2 trillion in grids, renewables, electrification, and efficiency, and $1.2 trillion still in oil, gas, and coal. BloombergNEF separately tracks energy-transition investment (electrified transport, renewables, grid buildout, storage) at a record $2.3 trillion in 2025, projected to average $2.9 trillion a year through 2030 — on the order of $14 trillion for the rest of the decade in that category alone. Layered on top of it, McKinsey estimates data centers will require $6.7 trillion in capital by 2030, with AI-driven demand accounting for $5.2 trillion of it. None of that gets built without real estate, a substation, a transmission corridor, and a buildable footprint underneath it — which is exactly the layer this thesis is about.

• Biospheric resilience. The same forces straining supply also create a resilience market of their own. UNEP's Adaptation Gap Report 2025 puts developing-country adaptation finance needs at $310–365 billion a year by 2035 — against international public adaptation finance that totaled just $26 billion in 2023, a gap the report calls 12 to 14 times current flows. That's the shortfall in the hardest-hit economies alone, before counting resilience spending, insurance repricing, and nature-based infrastructure in the developed world.

Add a conservative reading of those three — natural resources, infrastructure, resilience — and the applicable market isn't measured in billions. It's tens of trillions of dollars in investment and capex over the next ten years, all of it ultimately executed as a series of decisions about specific parcels of land.

The gap both conclusions point to

That leaves two conclusions pointing to the same problem. For firms operationally dependent on spatial data, competing without AI-readable land intelligence is becoming an existential disadvantage — the diligence that used to differentiate a good deal from a bad one can't keep pace with capital moving at this scale. And for a world allocating a finite, shrinking resource under accelerating pressure, allocating it well isn't optional — it's a civilizational requirement, playing out across tens of trillions of dollars of decisions this decade alone.

Both depend on something that doesn't yet exist: a foundational identity layer that makes physical space computable, the way markets, shipments, and sensors already are.

The one input everything else depends on is the only one AI can't read — and it's getting scarcer from three directions at once.