The intellification thesis
Electricityis howintelligencetouchesmatter.
What we call the second electrification is real—but it is not the largest category. It is the physical subdomain of a more consequential transformation: intellification.
01 / The category error
The phrase second electrification is trying to hold several changes at once: electric vehicles replacing engines, heat pumps replacing furnaces, batteries balancing grids, electric processes replacing combustion, and data centers becoming industrial loads.
That description is accurate but incomplete. It names the carrier—electricity—while overlooking the new organizing principle. The defining feature of this era is not simply that more things consume electrons. It is that more systems can perceive conditions, infer what should happen next, coordinate with other systems, and convert a decision into physical action.
I call that larger process intellification: the insertion of adaptive decision capability into products, infrastructure, organizations, and environments. Under this definition, electrification is not a rival megatrend. It is the branch of intellification that gives decisions a fast, measurable, controllable way to act on matter.
The control loop is the real machine.
An electric endpoint becomes part of intellification when it closes this loop. Electricity powers every stage, but the transformation lies in the loop’s ability to change behavior in response to reality.
Sense
Measure temperature, motion, load, price, traffic, vibration, demand, risk, or human intent.
Model
Turn raw signals into a representation of what is happening and what may happen next.
Decide
Select an action against goals and constraints: cost, safety, comfort, speed, carbon, or reliability.
Act
Move electrons through motors, chips, batteries, relays, pumps, heaters, robots, and networks.
Because the electric conversion is the action architecture inside a broader intelligence architecture. An ordinary toaster is electrified without being intelligent. A heat pump that learns a building’s thermal behavior, responds to grid conditions, protects comfort, and shifts demand is participating in intellification. The distinction is not whether electrons are present. It is whether decisions have become adaptive and executable.
The first electrification already taught us this lesson.
Electricity’s largest contribution to the factory was not a cheaper shaft. It was freedom from the shaft.
Steam-era factories organized machines around a central drive train of belts and pulleys. Individual electric motors allowed each machine to be placed where the flow of work demanded. The technology changed the geometry of production—and then the organization of the firm around it.
Historical evidence finds that U.S. manufacturing electrification brought rapid, persistent productivity gains alongside capital deepening and organizational change. The point is not merely that factories received power. Power became divisible, relocatable, and independently controllable.
The same mistake is available to us now. We can count chargers, megawatts, heat pumps, batteries, and data centers while missing the new freedom they create. The second electrification does not only replace fuels. It makes physical processes addressable by software—and increasingly governable by models that can reason across the whole system.
Electrification makes matter addressable. Intellification decides what to ask it to do.
A common, fast, precisely metered medium for computation and physical work.
Millions of endpoints across time, location, price, capacity, and constraint.
The plan itself changes when conditions, evidence, goals, or risks change.
The numbers reveal a two-way system.
More final energy becomes electric
The IEA’s stated-policies outlook moves electricity’s share of global final consumption from roughly 20% to 26% by 2035. More ambitious scenarios reach 29% or 36%.
Intelligence becomes a major electric load
Global data-center electricity use is projected to rise from 415 TWh in 2024 to about 945 TWh in 2030, with AI the most important driver of that growth.
Intelligence can release electric capacity
The IEA estimates that remote sensing and AI-based management could unlock up to 175 GW of transmission capacity without building new lines.
Those facts are often placed in separate stories: one about decarbonization, another about AI’s appetite for power, and a third about smart grids. They are one coupled system. Intelligence consumes electricity; electricity extends intelligence into the physical world; intelligence then helps coordinate an electrical system made more complex by variable supply and flexible demand.
This is why describing AI as “just another load” is as incomplete as describing an electric motor as “just another use of current.” The load matters. But so does the reorganization it enables.
02 / What follows
Build for closed loops, not isolated devices.
- The grid becomes a coordination network.
Generation and transmission remain foundational, but value increasingly depends on visibility, controllability, flexible loads, storage, forecasting, and secure orchestration.
- The endpoint becomes a participant.
A car is transportation, a battery, a sensor platform, and a schedulable load. A building is shelter, thermal storage, occupancy data, and a controllable node. Category boundaries begin to dissolve.
- The relevant efficiency becomes system efficiency.
Tokens per watt and miles per kilowatt-hour matter. So do avoided peaks, reduced downtime, better routing, longer asset life, and useful work completed per constrained unit of energy.
- Energy policy becomes intelligence policy.
Permitting, interconnection, market design, data access, cybersecurity, technical standards, and accountability determine whether electrified assets can coordinate—or merely add load.
Intellification is not automatically wise. A badly specified objective can optimize the wrong thing at extraordinary speed. A connected endpoint can become a surveillance surface or a cyber vulnerability. Cheaper intelligent action can also produce rebound demand. The parent category is larger than electrification because its responsibilities are larger, too: judgment, governance, security, and human consent belong inside the architecture.