The Opportunity
The AI economy runs on things the world has not built yet.
Power, data centres, chips. Demand for all three is racing ahead. Supply is far behind. That gap is the biggest opportunity of our time, and it is what we build.
The Pattern
Three layers. One simple problem.
Every part of the AI economy tells the same story. We have a certain amount today. We will need far more by 2030. The difference, the gap, has to be built. Here is that gap, layer by layer.
01 Energy
0
the gap to build by 2030
Today the world's data centres use about 485 TWh of electricity a year. By 2030 they will need about 950, as much power as the whole of Japan uses today. That is roughly double, in five years.
Source: IEA, Energy and AI
02 Data Centres
0
needed to build the capacity
Today the world has about 82 gigawatts of data-centre capacity. By 2030 it will need about 219, nearly three times as much. Building it will cost around $6.7 trillion.
Source: McKinsey
03 Chips
0
of the most advanced chips come from one island
Chips are the hardest layer. The market will grow from about $700 billion today to over $1 trillion by 2030. But almost all of the most advanced chips are made in a single place, with no back-up if it is ever cut off.
Source: SIA
The Build-Out
The build-out runs into the trillions.
By 2030 the world needs to build around $6.7 trillion of data centres, most of it for AI, and about $1 trillion of new chip factories. These are the sums the AI economy must deploy, and it needs places to put them.
Sources: McKinsey, SIA. Cumulative capex to 2030, US$ trillion. AI is the largest share of data-centre spend.
The Thread
It is all the same problem.
Power, data centres, chips. Three layers of one foundation. The AI economy cannot run without any of them. The world is short of all three. And they are made in too few places.
Someone has to build this foundation, in more countries, and make it safe enough for serious capital to follow. That is a hard job. It is also the opportunity.
Our Approach
We started wide. We went narrow. Now we are building the platform.
First
We began by advising on this infrastructure, data centres and power plants, across the Middle East, Africa, and Asia. We learned where the real value, and the real risk, sit.
Then
We went narrow. We took our own position, with our own balance sheet, in the hardest layer of all: semiconductors. That venture is Alpha Power.
Now
We are building an infrastructure investment platform to do this across the whole foundation, power, data, and chips, and help power the AI economy.
Where We Come In
The gap is the opportunity. We build it.
If backing the physical foundation of the AI economy interests you more than the hype on top of it, we would like to talk.
Figures drawn from the IEA, McKinsey, and the Semiconductor Industry Association, 2024 to 2026.