The Stack

In northern Virginia, the world’s largest cluster of data centers hums quietly behind anonymous warehouse walls. Inside, rows of servers process searches, stream videos, and train the algorithms that increasingly shape modern life. Each facility consumes enormous quantities of electricity—so much that utilities across the region have begun planning new transmission lines and power plants simply to keep up with demand. The machines inside these buildings do only one thing. They turn electricity into computation. 

This is not a niche phenomenon. According to the International Energy Agency, data centers consumed roughly 460 terawatt-hours of electricity globally in 2022—about the same as the entire electricity demand of France. With the rapid expansion of artificial intelligence and cloud computing, that demand could more than double by the end of the decade.

The quiet electricity flowing through these buildings reflects a much larger transformation. For most of the modern era, power looked very different. Coal fired the engines of the nineteenth century. Oil powered the twentieth—fueling the ships, aircraft, cars, and chemical industries that made modern globalization possible. The industrial world ran on combustion. Energy was extracted from the ground, transported across oceans and pipelines, and ignited inside engines and turbines. From the steam engine to the jet engine, the logic of heat and pressure shaped the modern economy.

That system still exists. But quietly, another one has begun to emerge. Look closely at many of the machines that now define the frontier of technology and industry. Electric vehicles. Drones. Robots. Data centers. Even much of modern manufacturing. At first glance these seem like separate technological revolutions. In reality they share something fundamental: they are machines built around electricity.

An electric vehicle, at its core, is a battery connected to motors and software that converts stored electricity into motion. A drone is a battery that flies. Industrial robots are batteries turning electricity into mechanical work. Grid-scale storage is a warehouse of batteries stabilizing renewable power. Even the vast facilities that train artificial intelligence are, in physical terms, machines designed to convert electricity into computation.

Once you begin to see the pattern, it becomes difficult to unsee. Across industries that once seemed unrelated, a common technological architecture is taking shape. Electricity moves through semiconductors, batteries, motors, sensors, and increasingly complex software systems. These layers interact, reinforce one another, and grow together.

It is helpful to think of this architecture as The Stack—a layered industrial system that converts electrons into motion, computation, and production. The idea is less a prediction than a lens. It does not explain everything. But it helps clarify why certain technologies now seem to advance together, why manufacturing ecosystems matter more than they once did, and why the geography of industrial power may be shifting again.

Only a few years ago, the energy transition appeared to follow a relatively simple script. In the West especially, climate change was framed as a moral crisis that demanded a coordinated global response. Public mobilization surged. Climate summits were treated as historic turning points. Financial institutions adopted new language—“climate risk,” “stranded assets,” “orderly transition”—and ESG investing expanded rapidly across global capital markets. Mark Carney’s "tragedy of the horizon" framing helped elevate climate from an environmental issue to a financial-stability concern.

The expectation was that the world would move, however unevenly, toward decarbonization. But the story did not unfold in quite that way. Oil demand did not collapse. Instead, it remained stubbornly resilient. The United States’ shale revolution reshaped global energy markets and geopolitics. Russia’s invasion of Ukraine forced Europe to scramble for new energy supplies, reminding policymakers that energy security had never truly disappeared.

Rather than replacing one system with another, the global energy landscape grew more complicated. Climate policy continued to advance, but it did so alongside industrial competition, geopolitical rivalry, and the realities of supply chains. Gradually, the center of gravity shifted. The debate was no longer only about emissions or temperature targets. It was about manufacturing, infrastructure, and industrial capacity. The energy transition was becoming something larger: a transformation in how societies produce and use power.

One shorthand for the new era is “petrostate vs electrostate.” The petrostate is not simply an oil producer; it is a political economy organized around commodity rents: centralized revenue, weak taxation, patronage networks, and foreign policy shaped by supply leverage. Electricity complicates that logic. Electrons move through networks, depend on equipment and engineering, and are harder to concentrate into a single chokepoint.

But “petrostate vs electrostate” captures only a slice of the transformation. The deeper story is not the replacement of one fuel with another. It is the emergence of a general-purpose industrial base built around electrified production, computation, and advanced materials. That base changes what is scarce, what is strategic, and what it takes to maintain industrial leadership. A fuller description is combined and uneven electrification.

Electrification is often described as technological substitution: electric vehicles replacing combustion engines, heat pumps replacing furnaces, renewable generation replacing fossil power. But the deeper transformation is structural. Across the economy, electrons, semiconductors, batteries, magnets, grids, and software are becoming common inputs into an expanding range of systems—from transportation and manufacturing to data centers, robotics, and autonomous weapons. These technologies form an electro-industrial stack.

It is combined because the layers reinforce one another. Consumer adoption pulls scale. Scale drives learning curves. Learning curves reduce costs. Lower costs expand markets and justify further scale. Capabilities gained in one market spill into adjacent markets. It is uneven because those feedback loops spin at dramatically different speeds across countries—and within them. Some societies electrify consumption while remaining dependent on foreign manufacturing. Others accumulate industrial depth without translating it into broad-based prosperity. Within countries, adoption and investment can diverge across regions, turning electrification into an economic and political fault line.

A useful way to see the combined nature of the shift is to look at modern machines as electricity systems. An electric vehicle is a battery connected to motors and software that turns stored electricity into motion. A drone is a lightweight battery that flies. Grid-scale storage is a warehouse of batteries balancing variable generation. Robots are mobile batteries that transform electricity into mechanical work. Even data centers—the physical backbone of artificial intelligence—are vast machines designed to convert electricity into computation.

Once this architecture becomes common, the learning is transferable. The supply chain that feeds electric drivetrains supports drones and industrial automation. The factory that perfects battery packs for cars can apply that learning to stationary storage. The software that manages fleets can extend into logistics and defense applications. Electrification becomes a compounding industrial platform, not a single “clean energy” vertical.

One of the more surprising drivers of this transformation lies in consumer hardware.

For decades, analysts tended to treat consumer electronics as separate from heavy industry. Smartphones and laptops belonged to one economic world; factories and energy systems belonged to another. In practice the boundaries blurred.

Consumer technologies impose extraordinary demands on manufacturing systems. Devices must be reliable, affordable, and produced at enormous scale. Meeting those requirements forces companies to master complex supply chains, automated production lines, and tight integration between hardware and software.

As Andreessen Horowitz partner Ryan McEntush has argued, the decisive arena is often not “energy systems” in the abstract, but products that bundle performance, cost, and convenience into mass adoption.  Consumer markets impose brutal requirements: reliability, yield, cost, serviceability, and volume production. They turn engineering into manufacturing competence. Everything is a computer. When Steve Jobs first mass-marketed the iPhone, it was something foundational: “the first mass-market machine that bundled compute, power, sensing, connectivity, and software into a single, tightly-engineered package.” Once that template existed, a wide range of devices began converging toward similar electro-industrial architectures—laptops, appliances, drones, robots, and, beneath the bodywork, electric vehicles. In that sense, electrification spreads not only through policy but through performance. Consumers adopt products that are cheaper to operate, quieter, faster, and more capable, pulling entire supply chains behind them.

In that sense, consumer devices are not downstream of industry; they can be upstream of industrial capability. Smartphones trained global supply chains in miniaturization, battery management, sensors, and high-throughput electronics manufacturing. Electric vehicles are doing something similar for power electronics, motors, packs, and industrial automation—at a much larger physical scale. This is why certain firms and ecosystems matter beyond their market share. They serve as integration schools: places where design, manufacturing, software, and supply chains are forced to work as one system. 

Electric vehicles appear to be playing a similar role for the next phase of industrial technology. An EV is not simply a cleaner version of a car. It is a highly integrated electro-mechanical system combining large batteries, power electronics, software, and automated manufacturing. The factories that produce them increasingly resemble electronics facilities as much as traditional automotive plants.

The supply chains that feed those factories—from battery materials to motors and control systems—are becoming foundational capabilities for robotics, drones, and a range of emerging technologies. Consumer products, in other words, can become training grounds for industrial systems.

“The Stack” is a way to describe the layered architecture that converts raw materials into strategic capability. In simplified form, it includes:

• Low-cost electricity generation (increasingly renewable and nuclear in some grids)
• Power electronics and control systems (inverters, drives, converters)
• Batteries and storage (mobility and grid balancing)
• Critical materials and processing (from copper and aluminum to lithium, nickel, graphite, rare earths)
• Semiconductors (the control substrate of modern machines)
• Grids (transmission, distribution, interconnection, and reliability)
• Software (optimization, automation, AI, and systems management)

You could add more layers—industrial heat, electrochemistry, or even finance as the enabling bloodstream of deployment. But the key point is not the exact taxonomy. The key point is that these layers behave like a system. When they integrate, advantage compounds.

This also clarifies what The Stack is not. It is not a prediction that hydrocarbons vanish. It is not a claim that one country inevitably wins. It is not a moral narrative. It is a description of how industrial capability increasingly accumulates through scale, learning, and integration across physical technologies that share common inputs.

The rise of these technologies coincided with a world already shaped by different economic strategies. Over the past three decades the United States increasingly specialized in software, finance, and capital markets while large portions of global manufacturing migrated across the Pacific. By the late 2010s, manufacturing accounted for roughly 11 percent of U.S. GDP, down from about 16 percent in the mid-1990s, while services and finance expanded as dominant sectors of the American economy.

Europe focused on economic integration and regulatory coordination through the institutions of the European Union. The single market and the euro created one of the largest integrated economic zones in the world, but the continent struggled to produce technology platforms on the scale of those emerging in the United States and Asia. Today only two European firms—ASML and SAP—rank among the world’s largest technology companies by market capitalization, highlighting the continent’s more limited presence in the commanding heights of digital and electro-industrial platforms.

China followed a different trajectory. Over several decades it expanded manufacturing capacity across a wide range of industries while building the infrastructure required to support them. For much of the early twenty-first century this global division of labor appeared stable. Globalization allowed regions to specialize according to their strengths. 

Electrification technologies complicate that arrangement. Unlike software platforms, many of these technologies depend heavily on manufacturing ecosystems. Batteries, power electronics, motors, and grid infrastructure improve primarily through scale and learning-by-doing. The companies and countries that build them gain experience that compounds over time.

China’s expansion in solar manufacturing, battery production, and electric vehicles illustrates this process. According to the International Energy Agency, China today accounts for more than 80 percent of global solar module manufacturing capacity, roughly 70 percent of lithium-ion battery cell production, and over 90 percent of the world’s rare-earth magnet manufacturing. These industries are supported by dense industrial clusters linking raw materials processing, component suppliers, and final assembly.

The scale of these ecosystems accelerated cost declines across global markets. Between 2010 and 2020 the cost of solar photovoltaic modules fell by nearly 90 percent, while lithium-ion battery pack prices declined by a similar magnitude, driven largely by manufacturing expansion and learning effects. China accounted for over 60 percent of global electric vehicle sales in 2023, with more than 8 million EVs sold domestically, according to BloombergNEF.

Industrial advantage begins to resemble gravity.

Energy transitions have always reshaped the global economy. Coal powered the factories and railways that drove nineteenth-century industrialization. Oil enabled the mechanized warfare, aviation, and global logistics networks that defined the twentieth century. Each transition reorganized not only fuel systems but the geography of industry and power. Something similar may now be underway.

Increasingly, analysts describe a shift away from an energy system organized primarily around fuel extraction toward one built around integrated electricity infrastructure. Jeff Currie, the former Goldman Sachs commodities strategist now at Carlyle, has described this emerging system as a “New Joule Order.” In this framework, the key strategic resources of the twenty-first century are not simply barrels of oil or tons of coal but the infrastructure that produces, stores, and manages electricity: power grids, transmission lines, battery supply chains, semiconductor fabrication, and industrial manufacturing capacity.

The insight reflects a broader shift in how energy advantage is measured. In a more electrified economy, industrial power depends less on access to fuel deposits and more on the ability to build and operate the infrastructure that converts electricity into useful work. For much of the twentieth century, access to hydrocarbons determined strategic leverage. Today the ability to deliver large quantities of reliable electricity—and to manufacture the technologies that use it—is becoming equally important.

Major energy companies are beginning to frame the future in similar terms. Shell’s long-term energy outlooks describe a world increasingly defined by electrification, where the share of electricity in final energy consumption continues to rise as transport, industry, and digital infrastructure shift toward electric systems. Electricity, once a secondary energy carrier, becomes the central platform around which modern economies organize their productive activity.

Seen through this lens, the emerging contest is not simply about energy supply. It is about the industrial ecosystems capable of building and deploying the technologies that convert electricity into economic output. And that returns us to the central challenge facing many advanced economies today.

Across much of the Western world, the central obstacle to electrification is no longer technological discovery. It is the capacity to build. From Thompson and Klein’s Abundance to Dunkelman’s Why Nothing Works to Wang’s Breakneck, a wide swath of literature diagnoses hollowed institutions, deindustrialization, regulatory capture, and lawyerly sclerosis. All describe decayed state capacity and lack of industrial policy.  Electrification requires physical infrastructure on an enormous scale: factories, transmission lines, renewable generation, storage systems, and data centers. Building those systems demands coordination across permitting regimes, capital markets, labor forces, and political institutions.

Where institutions function smoothly, deployment accelerates. Where they do not, progress slows—even when the technologies themselves are ready. The energy transition is therefore becoming a test of institutional competence as much as engineering.

Energy transitions have reshaped global power before. Coal powered the industrial empires of the nineteenth century. Oil underpinned the military and economic systems that dominated the twentieth. The shift now underway may prove similarly consequential.

The strategic infrastructure of the twenty-first century increasingly includes semiconductor fabrication plants, battery factories, power grids, and data centers alongside traditional energy systems. These assets determine where supply chains form, where technological learning occurs, and where industrial ecosystems deepen.

The competition that follows is not simply about fuels or emissions. It concerns the layered industrial system that converts electricity into economic and military capability. In other words, it concerns the stack of technologies and industries that increasingly power the modern world. Understanding how that system develops—and where it concentrates—may prove to be one of the defining economic and geopolitical questions of the twenty-first century.

The strategic infrastructure of the twenty-first century increasingly includes semiconductor fabrication plants, battery factories, power grids, and data centers alongside traditional energy systems