THE PHANTOM ASSET: How economic claims on things that do not exist yet are already shaping the future
THE PHANTOM ASSET
How economic claims on things that do not exist yet are already shaping the future
There is a strange kind of asset emerging in the middle of the artificial-intelligence boom. You cannot touch it. You cannot walk into it. It has no walls, no servers, no employees and, in some cases, no certainty that it will ever be built. And yet it can occupy something real: electricity capacity, a place in a grid queue, a piece of land, a future connection, a reservation on infrastructure that somebody else may desperately need.
It sounds like a contradiction. How can something that does not exist already take up space in the real world?
Britain is discovering the answer. America is discovering it too—and the American version may be even more revealing.
The country's electricity system has become crowded with applications from companies seeking connections for enormous new loads, particularly data centres. Demand-connection applications surged from 41 gigawatts to 125 gigawatts in less than a year, with data-centre projects accounting for at least 80 gigawatts of that demand. The British government says speculative applications have helped cause the transmission demand queue to grow by 460 percent in just six months.
The numbers are difficult to comprehend because they describe something that is partly imaginary. Not imaginary in the sense of fraudulent. Not every project is fake. Many will be built. Some may become enormous industrial facilities employing thousands of people and consuming extraordinary quantities of electricity. But a queue application is not a data centre. It is a claim about a possible data centre. And that distinction may become one of the most important economic distinctions of the AI age, because the grid has to react to the claim before it knows whether the building will ever arrive.
That is the beginning of the phantom.
And then look across the Atlantic. In June 2026, the U.S. Federal Energy Regulatory Commission explicitly warned that speculative large-load requests are clogging electricity interconnection queues, diverting resources and distorting forecasts. FERC said data-centre developers can be incentivized to “shop around” prospective projects with different utilities to find the fastest or cheapest connection. The consequence is not merely bureaucratic clutter. Projects that may never materialize can be counted in forecasts, creating double counting, distorted market signals and potentially higher costs for consumers.
PJM, the largest U.S. power market, put the problem even more bluntly in a 2026 filing. It described “phantom” or speculative load requests as a major source of uncertainty and said data-centre developers were submitting requests in different jurisdictions while effectively pitching the same projects in different locations. In the filing, a quoted former Texas regulator offered the darkest summary of the incentive: when it is cheaper to buy a queue position than not use it, developers have little reason not to buy the position.
Now the paradox becomes impossible to ignore. America is not merely preparing for data centres. Parts of its electricity system are being asked to prepare for proposed data centres that may not materialize. The claim is entering the planning model before the building enters the landscape.
And the consequences are already physical. PJM is simultaneously dealing with rapidly rising data-centre demand and a capacity market struggling to bring enough new power online; Reuters reported in August that its 2028/29 capacity auction fell short by 6.8 gigawatts while more than 50 gigawatts of power projects were considered ready and roughly 220 gigawatts remained under evaluation.
This is the moment the phantom stops looking like a spreadsheet problem. A future load can help determine which generation gets built, which transmission gets reinforced, which projects wait, how much capacity costs and how utilities plan years ahead. The building may be absent. The consequences are not.
Imagine an empty field. There is no data centre there. There are no servers humming inside concrete buildings. No cooling towers. No security gates. No artificial intelligence models being trained at three in the morning. There is only land. Then somebody announces a project. The developer applies for hundreds of megawatts of electricity. The grid operator has to consider the request. Future transmission requirements may now be modelled around it. Other applicants may find themselves behind it. Infrastructure investment may be planned with its projected demand in mind. Nothing has been built. But something has changed.
The future has entered the queue.
That is why the British regulator Ofgem is now explicitly trying to remove speculative data-centre projects from the electricity-connection system. Its July 2026 proposal says the explosion in demand applications may include a significant number of projects that never ultimately proceed, creating misleading signals for network investment and potentially delaying viable developments. Ofgem is proposing a commitment fee of £237,500 to £712,500 per megawatt for large data-centre projects, refundable when a project reaches energisation but forfeited if it exits the queue early.
Think about what that means. A hypothetical 500-megawatt project could face a commitment exposure of roughly £119 million to £356 million under that proposed range. The building may not exist. The servers may not exist. The customers may not exist.
But the right to remain connected to the possibility of that future can already have a price.
This is where the story stops being about data centres and becomes something much stranger.
We have always understood assets as things. A factory is a thing. A mine is a thing. A power station is a thing. A house is a thing. Even financial assets generally represent something that exists somewhere: ownership, debt, a claim on cash flows. But the emerging phantom asset is different. It is a claim on something that may exist later. And the more scarce the underlying resource becomes, the more powerful that claim can become.
Britain's grid is an unusually clear laboratory because connection capacity is scarce and projects can wait years for reinforcement. A 2026 peer-reviewed study by researchers including economists at Cambridge describes the grid connection system in precisely these terms: available connection capacity at a substation becomes a scarce service when demand exceeds supply. The researchers examine whether queue positions could potentially be traded and explicitly discuss the economics of secondary trading of connection rights.
That does not mean electricity queue positions have suddenly become freely tradable financial securities.
It means something more interesting.
Economists are now asking whether a place in the queue can itself be treated as something with economic value.
A position in line has entered the language of markets.
And once you see that, the phantom becomes easier to recognise.
A developer does not necessarily need the finished building to benefit from having secured a future connection. The claim itself may have strategic value. It may tell investors that the project has a path to power. It may make a piece of land more attractive. It may give a developer an advantage over somebody arriving later. It may influence the price of infrastructure around it. It may even change the behaviour of competitors. The physical asset has not arrived. But the economic consequences have.
The shadow has arrived before the object casting it.
This is why the phrase “powered land” has begun appearing in the data-centre economy. Reuters reported that land suitable for data centres in London has been commanding substantial premiums over ordinary industrial land, with suitable sites reaching roughly £8 million to £15 million per acre compared with around £4.5 million to £6 million for ordinary industrial land. Reuters also reported that only 7 percent of 61 British data-centre projects it tracked since late 2022 had started construction.
That is an extraordinary picture of an economy valuing possibility.
The land is real. The power connection may be real or prospective. The data centre may still be years away. Yet the possibility of what the land could become is already affecting its value.
This is the same transformation we began seeing with the Strategic Acre.
But the Phantom Asset goes one step further.
The Strategic Acre asks:
What might this land become?
The Phantom Asset asks:
What if the possibility of becoming it is already economically valuable?
That is a much darker question.
Because it means the future does not have to arrive before it can begin competing with the present.
A proposed data centre can compete for electricity before it consumes electricity. A proposed industrial project can compete for land before it builds a factory. A proposed mine can compete for infrastructure before it extracts a single tonne of mineral. A proposed housing development can affect land values before the first foundation is poured. The claim comes first. The physical object comes later. Sometimes it never comes at all.
And that last possibility is what makes the system so strange.
The electricity grid cannot simply assume every application will become reality. The British government is now trying to strengthen the conditions for entering and remaining in demand-connection queues precisely because speculative projects can consume planning capacity without ultimately materialising.
Ofgem's proposed reforms are effectively an attempt to force the future to prove that it deserves to occupy the present.
Show us the money. Show us the customer. Show us the equipment. Show us that you are actually going to build.
Otherwise, get out of the queue.
That sounds like ordinary regulation.
It is actually an extraordinary admission.
The system has become crowded enough that unbuilt things are interfering with things that could be built.
And that is where the phantom asset becomes politically dangerous.
Suppose two projects need the same scarce electricity connection. Project A has financing, customers and construction plans. Project B has land, an ambitious proposal and a place in the queue. If B arrived first, should A have to wait? The old answer was simple: First come, first served. But when the resource becomes extremely scarce, “first” starts acquiring economic meaning. A place in line becomes valuable.
The 2026 Cambridge research points directly at this problem. The existing first-come-first-served system does not necessarily account for the feasibility, progress or value of competing projects, and the researchers explore alternative allocation mechanisms, including potential trading of queue positions.
The queue begins to resemble a market.
And markets have a peculiar habit. They put prices on things that previously looked administrative: a permit, a licence, a spectrum allocation, a development right, a grid connection, a reservation, a place in line.
Once scarcity becomes severe enough, paperwork starts looking remarkably like property.
But there is another layer.
Artificial intelligence is creating enormous expectations about future electricity demand. Investors are pouring money into data centres because they believe AI will require extraordinary amounts of computing capacity. That expectation is causing developers to secure land, pursue power connections and assemble projects years before the full demand is visible.
The future is therefore helping finance the infrastructure that is supposed to create the future. That sounds circular because it is.
Expectation creates investment. Investment creates infrastructure. Infrastructure makes the expected future more plausible.
And the more plausible the future becomes, the more valuable the claims on it may become.
This is how a phantom can acquire weight.
Not physical weight.
Economic weight.
There is a frightening possibility hidden inside that mechanism. What if the biggest infrastructure boom of the AI age creates an enormous market for things that are not yet real? Not necessarily fraudulent things. Not fake companies. Not imaginary assets. Something much more ordinary. Projects. Thousands of projects. Some genuine. Some speculative. Some over-optimistic. Some waiting for customers. Some waiting for financing. Some waiting for electricity. Some waiting for planning permission. Some waiting for a technology that has not matured. Some waiting for a future that may never arrive in exactly the form their spreadsheets predict.
Yet collectively they begin shaping the decisions of governments, utilities, investors, landowners and competitors.
The future becomes an input into today's economy.
And that changes the meaning of risk.
For centuries, investors mostly asked whether an asset would produce enough value after it was built.
Now an increasingly important question may be:
How valuable is the claim that allows you to build it?
That is the Phantom Asset.
It is not the building.
It is not the machine.
It is not the power station.
It is the economic value created by having a credible claim on the future physical asset.
And the closer the world gets to genuine scarcity, the more powerful that claim becomes.
We have already seen versions of this logic elsewhere.
A development right can be valuable before the building exists. A mineral exploration licence can be valuable before the mine exists. A spectrum licence can be valuable before the network exists. A water right can be valuable because of what may eventually be built or produced with it. A grid connection can be valuable because of the industrial activity it may enable.
None of this requires a conspiracy.
That is what makes it unsettling.
Nobody has to sit in a dark room and decide to manufacture scarcity.
The scarcity is already there.
Electricity takes years to generate and transmit. Transmission lines take years to permit. Transformers can take years to procure. Land cannot be manufactured. Water cannot be manufactured. Planning approvals cannot simply be printed.
And when the real world moves slowly while expectations move at the speed of financial markets, the gap between what exists and what is expected to exist becomes economically significant.
That gap is where the phantom lives.
The AI revolution has made the gap enormous.
A company can announce a future data centre today. Investors can value the company tomorrow. The land can change hands next week. The grid can reserve capacity next month. The government can begin planning transmission infrastructure next year. And five years later, perhaps, the building appears. Or perhaps it doesn't.
But by then the phantom may already have changed the landscape.
That is the part we should be watching.
Not whether every data centre gets built. But what happens to the world while everyone is preparing for the possibility that they will.
Because perhaps the most important assets of the next decade will not be the things we can already see.
They will be the claims on things we are told are coming.
The megawatt that has not yet been consumed. The factory that has not yet been constructed. The mine that has not yet produced a kilogram of mineral. The city that has not yet been built. The infrastructure that exists only on a planning document. The project that lives inside a spreadsheet. The future that has already been assigned a number.
And perhaps that is the most unsettling thing about the Phantom Asset.
The future does not have to exist to begin competing with us for the present.
It only has to be believed.
And once enough money believes in it, enough institutions plan around it, enough land is assembled for it and enough scarce resources are reserved for it, the future acquires something very close to physical weight.
You can no longer see the thing.
But you can see what it is doing.
A queue grows. A price rises. A grid is redesigned. A field changes hands. A regulator intervenes. Another project waits.
And somewhere, on a computer screen, there is still nothing but a proposal.
A thing that does not exist.
A claim on something that may never exist.
And yet the world is already moving around it.
Perhaps that is the real economic revolution now beginning beneath artificial intelligence.
Not the invention of machines that can predict the future.
But the emergence of markets capable of pricing the future before it arrives.
And if that is true, then the question is no longer whether something that does not exist can be worth billions.
The question is far more disturbing:
YOU MAY ALSO LIKE
THE NEXT BILLIONAIRES AREN'T BUYING THINGS—THEY'RE BUYING THE NEXT SCARCE RESOURCE: PERMISSION
By: Manish Kumar
Founder & Editor, Explain It Clearly
Comments
Post a Comment