Force 08 / Power
This force hub tracks who controls the infrastructure, institutions, capital, standards, and chokepoints that decide how frontier technology reshapes the world.
China’s robotics push is easier to understand as industrial strategy: embodied AI, supplier depth, factory deployment and state-directed modernization moving into one frame.
Read the briefing →Read now
Newest power stories on control, chokepoints, and institutional leverage.
EIA’s AEO2026 does not prove an AI power crisis. It does something more useful: it makes data-center server energy use visible as a modeled planning object.
AI’s electricity problem has moved beyond power contracts. The next phase is about who controls the grid territory and utility assets that compute depends on.
AI companies are exploring data centers in orbit. The real story is not science fiction, but what the idea reveals about power, land, cooling, and compute on Earth.
The AI buildout may be limited by the parts of the grid nobody can manufacture fast enough.
AI’s next constraint may not be chips or models, but access to boring, firm, clean electricity that can run all night.
Terafab is what happens when AI stops looking like software and starts looking like heavy industry.
Field note
A compact operating note for Power: what changed, what to watch, what to doubt, and where the constraint sits.
As systems scale, control moves to the layers everyone else must pass through.
Who controls frontier access and under what conditions.
Open access removes all leverage.
Capability without coordination can create fragility instead of durable power.
Corridors
Use these routes when a story crosses into neighboring fields, forces, or reading formats.
Latest corridor story
Orbital Data Centers: When AI Compute Tries to Leave Earth
Latest corridor story
Data Center Electricity Demand Is Becoming Official Grid Accounting
Archive
Use the archive for retrieval; the map above carries the orientation.
EIA’s AEO2026 does not prove an AI power crisis. It does something more useful: it makes data-center server energy use visible as a modeled planning object.
AI’s electricity problem has moved beyond power contracts. The next phase is about who controls the grid territory and utility assets that compute depends on.
AI companies are exploring data centers in orbit. The real story is not science fiction, but what the idea reveals about power, land, cooling, and compute on Earth.
The AI buildout may be limited by the parts of the grid nobody can manufacture fast enough.
AI’s next constraint may not be chips or models, but access to boring, firm, clean electricity that can run all night.
Terafab is what happens when AI stops looking like software and starts looking like heavy industry.
The real significance of AI chip sales is not hardware trivia. It is that compute is becoming a measurable form of industrial and political power.
EGS matters because the AI energy race will not be won by glamour tech alone, but by whoever can deliver reliable clean power at scale.
AI has turned firm clean electricity into infrastructure strategy, not a background utility question.