For the past several decades, energy was something most corporate leaders could afford not to think about much. Turn on a factory, build a data center, expand a warehouse: the power would be there.

That assumption is disappearing.

Across major economies, electricity demand is rising again. AI and data centers are grabbing the headlines, but they are only the leading edge of a much larger transformation. Manufacturing is expanding and changing. Transportation is electrifying. Cooling demand is rising. Industrial processes are beginning to shift from fuels to electricity. And digital technologies are becoming embedded in nearly every sector of the economy.

Suddenly, access to reliable, affordable energy can determine where a company locates, how quickly it grows, what it produces, and whether an investment gets built at all. The instinctive response is to find more energy. But that is the wrong place to start.

The more interesting question for corporate leaders is how much more economic value can we create from the energy system we already have? And how much new infrastructure can we avoid, defer, or improve by designing demand differently? These questions turn energy from a procurement problem into a business strategy.

Fifty years ago, RMI cofounder Amory Lovins challenged an energy industry focused overwhelmingly on producing more supply. He argued that we should start instead with the services people actually want — warm homes, mobility, light, industrial production — and work backward to find the least costly, most resilient way to provide them. In his original “soft energy path” argument, Lovins argued that the most consequential energy choices were about design, not fuels. Today, the technology is different, but the principle remains the same.

We have extraordinary new tools: cheap solar and batteries, electric vehicles, heat pumps, flexible loads, virtual power plants, advanced transmission technologies, artificial intelligence, and increasingly sophisticated ways to manage energy in real time. Recent RMI analysis finds that fast, efficient solutions can meet most of the anticipated US electricity capacity gap over the coming decade. The opportunity is no longer about producing energy differently. It is to design whole systems differently.

For business leaders arriving in New York for Climate Week, that leads to five propositions worth considering.

The opportunity is no longer about producing energy differently. It is to design whole systems differently.

1. Your energy supply is both a strategic asset and a strategic liability.

For decades, companies optimized supply chains around low-cost labor, logistics, capital, and proximity to markets. Energy was usually somewhere further down the list.

It is moving up fast.

Energy security today is inseparable from economic security. Wars can disrupt globally traded fuels. Extreme weather can interrupt electricity supplies. Grid congestion can delay new factories and data centers. Volatile commodity prices can turn apparently attractive investments upside down.

But there is another side to this story.

Every unit of fuel a company doesn’t need to buy is a unit of commodity exposure it doesn’t carry. Every megawatt of peak demand it eliminates is grid capacity it doesn’t have to wait for. Every flexible process gives it another option when electricity is scarce or expensive. And every locally available energy resource reduces exposure to events thousands of miles away.

So the most resilient energy strategy isn’t necessarily the one with the most supply. It may be the one with the fewest dependencies. That’s why efficiency, electrification, distributed energy, storage, flexible demand, and diversified clean generation belong in conversations about business continuity and competitiveness — not simply sustainability.

2. Don’t buy your way out of energy constraints. Design your way out.

A single proposed data center campus can demand as much electricity as a small city. In some regions, the volume of proposed projects vastly exceeds the existing grid. RMI notes that AEP Ohio reported 30 gigawatts of pending data-center requests in 2024 — enough to more than triple its peak load.

The conventional response is straightforward: More computing requires more electricity. More electricity requires more power plants. More power plants require more transmission. So build everything as quickly as possible.

But there is a problem with that logic. A data center is not a force of nature. Its demand can be designed.

How efficiently chips compute, how servers are utilized, how facilities are cooled, where data centers are located, whether workloads can move across hours or geographies, whether batteries can shift demand, and how facilities interact with the surrounding grid all affect the amount of infrastructure ultimately required.

A data center is not a force of nature. Its demand can be designed.

3. Electricity is becoming a competitive input, not just a utility expense.

The AI power crunch matters for another reason: it is a preview.

Data centers are today’s highly visible new load, but electricity demand is also growing as transportation, buildings, cooling, manufacturing, and industrial processes electrify. That means the infrastructure we build to meet today’s needs can either become an expensive bottleneck or the foundation for a much more productive economy. The companies that understand this will stop treating electricity as simply another line item in operating expenses.

They will ask where abundant electricity creates competitive advantage. They will consider power availability when making location decisions. They will coordinate capital planning with energy planning. They will redesign industrial processes around the characteristics of electric technologies rather than simply swapping one fuel for another.

And increasingly they will see their own assets as part of the energy system. A factory that can shift production by several hours has value to the grid. A fleet of plugged-in vehicles is potentially a large flexible load and eventually a source of stored power. Buildings with batteries, smart controls, thermal storage, and flexible equipment can behave less like passive consumers and more like power-system resources.

This is already happening. RMI finds that virtual power plants already serve tens of gigawatts of peak demand and can provide utility-scale services when they are properly incorporated into planning.

That means the infrastructure we build to meet today’s needs can either become an expensive bottleneck or the foundation for a much more productive economy.

4. The biggest companies can do more than just buy from markets. They can create them.

The same shift in thinking applies beyond electricity. Companies buy enormous quantities of steel, cement, chemicals, fuels, transportation, and other commodities. Traditionally, procurement teams have optimized those purchases around price, quality, availability, and risk.

But purchasing decisions can do something else: they can determine which technologies reach commercial scale. Many emerging technologies face a chicken-and-egg problem. Producers hesitate to build because they don’t know whether customers will buy. Customers hesitate to commit because the product hasn’t yet reached scale.

Large buyers can break that deadlock. Aggregated purchasing commitments can give producers confidence to invest. Long-term contracts can help make projects financeable. Better product-level information can allow customers to distinguish among products that once looked like interchangeable commodities. RMI’s recent work on advance market commitments shows how buyers can help create markets for technologies that do not yet have one.

This turns procurement from a back-office function into an instrument of industrial strategy.

The provocative question for the world’s largest companies is therefore no longer “what should we buy?” but rather “what market do we want to exist?”

Companies that answer that question collectively can accelerate technologies and supply chains that no individual company could create alone.

5. Stop planning for one future. Build for multiple options.

Perhaps the most dangerous assumption in today’s energy system is that we know what happens next.

We don’t.

No one knows precisely how quickly AI electricity demand will grow. We don’t know where the next geopolitical shock will occur. We don’t know exactly how quickly batteries, advanced geothermal, long-duration storage, nuclear technologies, artificial intelligence, or other innovations will improve. And we don’t know which new products and industries cheap electricity and computation will make possible.

Companies routinely make 20-, 30-, and 40-year infrastructure investments based on a single view of the future, but there is another approach: design for optionality.

Stage investments. Favor modular technologies where appropriate. Build flexible loads. Preserve multiple supply options. Use existing infrastructure more intensively before replacing it. Coordinate investments that are too often made separately. And place a premium on technologies whose value persists across multiple possible futures.

The uncertainty surrounding data center demand makes this especially important. RMI’s planning framework emphasizes better information, least-regrets investments, staged procurement, and risk allocation as uncertainty resolves.

That principle is just as useful inside a company. Consider a manufacturer planning a new facility. Power procurement, process heat, efficiency, storage, production scheduling, transportation, and backup power might traditionally be managed by different teams. Optimize them together and the economics change.

Efficiency reduces required capacity. Electrification reduces exposure to fuel prices. Storage can shift demand. Flexible production can avoid expensive peak periods. On-site resources can improve resilience. Long-term power contracts can stabilize costs.

What began as a “decarbonization strategy” becomes a capital-efficiency, capacity, resilience, and growth strategy.

That is whole-system design applied to the corporation.

The energy strategy hiding in plain sight

There is an understandable temptation at this moment to think that energy abundance means producing enormous quantities of energy. At RMI, we think there is a more interesting possibility.

Abundance is not measured by how many units of energy an economy consumes. It is measured by what that energy enables people and businesses to accomplish.

The objective isn’t more megawatts. It is more computation, more mobility, more comfortable buildings, more industrial production, more resilient supply chains, and ultimately more prosperity — delivered with less cost, risk, and waste. That distinction matters enormously now.

The world is in the process of investing trillions of dollars in the infrastructure underlying AI, electricity, industry, buildings, and transportation. The easy path is to take today’s systems as given and build enough energy infrastructure to feed them. The more powerful path is to ask whether those systems themselves can be redesigned.

RMI’s recent work on the emerging electro-industrial economy reinforces the scale of this shift: electricity and computerization are becoming foundational to the next wave of economic development, tying energy strategy ever more tightly to investment, infrastructure, manufacturing, and competitiveness.

Fifty years after the original soft energy path, the technology has changed. The corporate imperative has changed. The stakes have changed. But the question remains surprisingly similar: before we build more, have we designed better?

For the companies that answer that question well, energy may cease to be a constraint on growth. It may become one of their greatest sources of competitive advantage.