When Economics Changes Faster Than Capital
2 August 2026
Why artificial intelligence may become the defining systemic investor story of the decade.
For decades, investors became exceptionally good at valuing scarcity.
Scarce oil.
Scarce semiconductors.
Scarce software.
Scarce expertise.
Scarce intelligence.
Artificial intelligence appeared to fit comfortably into that framework.
Build frontier models.
Charge premium prices.
Expand margins.
Generate cash flow.
Service debt.
Repeat.
It looked like another infrastructure cycle.
It may become something very different.
The Assumption
The investment case rested on one central belief.
Frontier intelligence would remain scarce.
Scarcity would support pricing power.
Pricing power would justify extraordinary capital investment.
That assumption unlocked hundreds of billions of dollars.
Not only equity.
Debt.
Private credit.
Infrastructure funds.
Power purchase agreements.
Equipment leases.
Construction finance.
Long-duration supply contracts.
Entire balance sheets were constructed around an expectation that frontier intelligence would remain economically scarce for years.
The recent acceleration in model efficiency is beginning to challenge that assumption.
Economics Can Change Overnight
Capital Cannot.
Model releases happen monthly.
Efficiency improves continuously.
Inference costs fall.
Open-weight competition accelerates.
Distillation compounds.
Every improvement reduces the cost of delivering intelligence.
The economics of AI are changing at software speed.
Capital was financed at infrastructure speed.
That difference matters.
A twenty-year financing agreement cannot adapt every time intelligence becomes cheaper.
Neither can a data centre.
Neither can a power contract.
Neither can billions already committed.
If economics moves faster than capital…
valuation assumptions begin breaking before physical infrastructure does.
Investors Don’t Own Companies
They own assumptions.
Assumptions about pricing.
Margins.
Utilisation.
Duration.
Competitive advantage.
Terminal value.
Every discounted cash flow is simply a collection of assumptions about the future.
When the underlying economics change faster than those assumptions…
markets eventually have to reprice.
This Is Why It Becomes Systemic
Modern financial markets are deeply interconnected.
Banks finance infrastructure.
Private credit finances expansion.
Pension funds own infrastructure funds.
Insurance companies own debt.
Cloud providers sign long-term commitments.
Utilities build generation.
Semiconductor manufacturers expand capacity.
None of these exposures exist independently.
They are connected through financing structures built upon similar assumptions.
The more connected the system becomes…
the more rapidly repricing can propagate once assumptions begin changing.
This is why systemic events often appear gradual…
until they aren’t.
Value Doesn’t Disappear
This is perhaps the most misunderstood part.
Economic value rarely disappears.
It migrates.
As intelligence becomes abundant…
value moves elsewhere.
Towards trust.
Towards coordination.
Towards distribution.
Towards memory.
Towards systems that reliably resolve uncertainty with the least computation.
The scarce resource is no longer intelligence.
It becomes reliable resolution.
Markets Have Learned to Value Scarcity
They May Now Have to Value Abundance
Perhaps this is the real transition investors are beginning to confront.
For decades, financial markets rewarded businesses that controlled scarce resources.
Artificial intelligence is beginning to make one of the most valuable resources in history progressively more abundant.
That does not eliminate value.
It relocates it.
The companies that understand where value migrates may define the next decade.
The companies still optimising for yesterday’s scarcity may discover that economics changed long before their financing agreements did.
History rarely announces paradigm shifts.
It simply leaves behind balance sheets that were built for a world that no longer exists.