Allocator language changed faster than most portfolios did. Two years ago the framing was abundance — cheap compute, cheap energy, cheap everything downstream of both. Now real-asset desks open with scarcity: power, water, food, housing, and the land underneath all four.
The shift is real. The trade most portfolios use to express it is not the same thing.
what real assets actually are
Real assets are holdings whose value comes from what they physically are and what they physically do — not from a claim on someone else's promise to pay. A corporate bond is a promise. A water right is a thing.
| category | examples | what you actually own |
|---|---|---|
| Land and resources | Farmland, timberland, ranchland, water and mineral rights | Productive capacity, plus the legal right to use it |
| Infrastructure | Generation, grid, pipelines, ports, fiber, powered land | A regulated or contracted stream of function |
| Real estate | Residential, industrial, net-lease | Space someone needs and pays to occupy |
| Commodities | Crude, copper, corn, gold — usually held via futures | Price exposure, plus roll yield or roll cost |
Three of those rows are capacity. The fourth is a price. That distinction does most of the work in this post.
why the essentials bid is real
Four forces moved real assets from a diversification footnote to the front of the allocation conversation. None of them are seasonal.
Inflation pacing. Financial assets discount future cash at a rate that moves against them when prices rise. Assets whose replacement cost rises alongside everything else tend to keep pace. That is a structural property, not a forecast.
Electricity and water demand from the compute buildout. Data-center campuses need power, cooling water, and land near both. Where those three intersect, land clears at prices agriculture cannot match. USDA put average US farm real estate at $4,350 per acre in 2025; land assembled for data-center campuses in power-constrained markets trades far above that, and the binding constraint has shifted from acreage to grid connection.
Chronic housing shortage. Shelter is the least elastic line in a household budget, and supply has been structurally short across most developed markets for over a decade.
Food and land politics. Sovereign funds buying agricultural supply chains, export restrictions, and farmland competing against solar leases and server halls are all the same phenomenon: essentials being secured rather than traded.
Put together, that is a flight to essentials — a durable bid for the things people cannot stop needing.
three layers of inelastic demand
The bid is not one thing. Separating the layers is what keeps the analysis honest.
| layer | what it is | examples | who prices it |
|---|---|---|---|
| 1. Human essentials | Need-to-live goods and services | Food, water, power, shelter | Markets — real assets, utilities, infrastructure, "living" |
| 2. Enabling natural capital | The systems that produce layer 1 | Pollination, soil biology, water cycles, watershed function | Academia and national accounts — mostly not markets |
| 3. Narrative weather | AI, war, drought, tech cycles | Who captures the rent in a given year | The commentariat |
Layer 3 is the loudest and matters least. Confusing it with layer 1 is how scarcity theses get mispriced. Ignoring layer 2 is how they get underbuilt.
you can be right about scarcity and still lose money
This is the part a commodities PM will test you on, so take it head-on.
Futures are cyclical, not a wealth machine. High prices are the cure for high prices. They invite supply, substitution, and demand destruction — that is the commodity cycle working, not failing. Over multi-decade windows, broad commodity indexes have generally lagged equities.
Consensus scarcity longs carry a logic bug. If scarcity drives inflation, inflation invites tighter policy and a stronger dollar, which compresses the very prices the trade depends on. Crowded scarcity positioning unwinds on its own success.
Compute demand is not a one-way ratchet. Near term it is real: power, water, copper, land. Medium term, efficiency and substitution cut the other way, and the narrative mean-reverts faster than the physical assets do.
Scarcity is not free private upside. When something becomes essential, it becomes political. Rent freezes, windfall taxes, export bans, and outright seizure are all ways a scarcity thesis can be exactly right and still pay nothing. Ownership form and rule of law are part of the underwriting, not context around it.
None of this argues against the essentials thesis. It argues against one specific way of expressing it. Conceding these points is the case for capacity over ticks.
the ownership ladder
Same view, different vehicle, very different asset.
| form | what you own | fit with the thesis |
|---|---|---|
| Commodity futures and indexes | Price exposure and roll | Weak — cyclical, no durable claim |
| Producer equities | Operator beta | Mixed — firm risk is not resource scarcity |
| Utilities and infrastructure | Regulated or contracted cash flow | Strong for power and water |
| Land, timber, farmland, water rights | Real productive capacity | Strong, if the cash-yield story is honest |
| Availability payments and offtakes | Contracted function, paid for existing | Strongest — you are paid for capacity being there |
"Price spikes in corn are not the same as owning the capacity that produces corn. The first is a quote. The second is an engine."
Notice where the ladder ends. The strongest form on the list is not a commodity at all — it is a contract that pays for function being available. Hold that thought.
the four essentials, briefly
Power is the most institutionalized. Generation, grid, and powered land are all ownable, with a decade of visible demand behind the buildout.
Shelter is ownable through residential and living strategies, with the caveat above — it is the essential most exposed to political intervention.
Water is the hardest to own cleanly. Utilities, rights, reuse, and desalination infrastructure are investable, but rights and regulation drive returns more than scarcity does, and the source — recharge, snowpack, watershed function — is almost never the thing being bought. That gap deserves its own treatment, and gets one in the asset that makes water.
Food is ownable as farmland and offtake. It is also the clearest illustration of the problem in the next section: the priced half of a harvest has a market, and the unpriced half does not.
the rung the ladder is missing
Every form above sits on a layer almost nobody is buying.
The Dasgupta Review states the problem most cleanly: between 1992 and 2014, produced capital per person doubled while the stock of natural capital per person fell by nearly 40%. We built the engines and drew down the fuel, and only one of those showed up in the accounts.
The spatial version is starker. Chaplin-Kramer and colleagues mapped the ecosystems supplying 90% of the local benefits people get from nature and found they occupy roughly 30% of global land. About 87% of humanity depends on that land. Only 16% of people live on it. Only about 15% of it is protected.
"Thirty percent of the world's land delivers ninety percent of the local benefits people get from nature. Eighty-seven percent of humanity depends on it, sixteen percent lives on it, and fifteen percent of it is protected. That is not a conservation gap. It is a missing market."
If you find the abundance case persuasive — yields up, calories cheaper, technology compounding — none of this contradicts it. Output can rise while the production function degrades; that is precisely what the accounts show. Abundant harvests and a depleting substrate are the same decades. Which is the whole argument in one line: owning the output is not the same as owning the engine.
how an unpriced layer becomes an ownable one
Nothing becomes an asset because it is valuable. It becomes an asset when someone manufactures a coupon — a payor, a term, and a claim.
That is what ensurance builds. In finance terms:
| piece | ordinary name | in ensurance |
|---|---|---|
| Who pays | Offtaker, ratepayer, concession authority | Dependency payors — cities, utilities, food companies, fabs, insurers who lose money when the function fails |
| What they pay | Availability payment | Premium for standing natural function, paid whether or not a loss occurs |
| How it is priced | Appraisal, rate base | RealValue — ecosystem-service value by condition, across 15 ecosystem stocks and 19 service flows |
| What you hold | The contract | Certificate — a tradeable claim tied one-to-one to a specific place |
| Where the money goes | Distribution waterfall | Proceeds routed onchain to the stewards holding the asset |
A certificate is the position. An agent is the onchain account representing a specific place or purpose. Proceeds are the routing. If you have watched an offset market fail, the last row is the one that matters: the money reaches the place, and the claim is on function you can verify standing today rather than on a counterfactual about what would otherwise have happened.
Structurally this is an availability payment on a toll road. The road exists; someone pays for it to keep existing. Here the road is a watershed.
the honest limit
Enabling-layer claims sit alongside working-land ownership, not instead of it. A protection claim is not the crop check. Anyone implying that a certificate on pollinator habitat produces farmland-like total returns is selling something — the return profile is closer to contracted infrastructure than to operating land, and that is the point of owning it.
The flight-to-essentials narrative is permission and urgency. It is not a substitute for underwriting.
where to go next
If the ownership-form argument landed, the next questions are shape and horizon.
- Who actually pays for the harvest? → food security is an ownership question
- Is this actually an asset class? → is nature an asset class yet? the honest answer
- Who can hold it for decades? → the money that can wait
- What happens when the unpriced gets priced? → the scarcity trade: what's free worth?
- What is already trading? → general ensurance, or start at ensurance