# The Future of Energy Is Arriving Faster Than Our Mental Models
By 

## Article Content
I spent part of today on stage talking about the future of energy, but the most important point I wanted to get across was not really about energy at all. It was about speed. The title of the keynote was Thinking Big, Starting Small, Scaling Fast: The Future of Energy. My central argument was that we are entering a period in which multiple forces are accelerating at the same time: electricity demand, electrification, data centre growth, renewable deployment, battery storage, artificial intelligence, distributed energy resources, and the software layer that increasingly sits on top of the physical grid. That changes the nature of the challenge. We are not dealing with one trend that can be studied in isolation. We are dealing with the interaction of many trends, each moving at a different speed, but increasingly colliding with one another. And that led to the most important line in the keynote: The biggest risk isn't that the future won't happen. It's that it happens faster than your mental model. The biggest risk: the future happens faster than your mental model. For years, I have used stories from science fiction, NASA, commercial space, and technology history to make a simple point. We tend to place the future comfortably in the distance. We imagine that the big change is still ten years away, that there will be plenty of warning, and that our existing planning assumptions will give us enough time to respond. Then the future moves. Something that seemed experimental becomes practical. Something that seemed too expensive becomes economically inevitable. Something that seemed like a niche technology begins to scale. Something that was a long-range strategic issue suddenly becomes an operational one. That is why I began the keynote with ideas such as the declining half-life of knowledge, instant obsolescence, and velocity. The real management challenge is not simply understanding what is coming. It is recognizing when the timetable has changed. Electricity demand is entering a very different era One of the most important numbers I used was the projection that Ontario electricity demand could grow by 65% by 2050. Ontario electricity demand: +65% by 2050. The number matters, but the composition of that growth matters even more. Electrification is moving transportation, heating, industrial processes, and other activities toward the grid. Electric vehicles continue to scale. New digital infrastructure requires enormous amounts of power. Economic growth itself increasingly depends on access to reliable electricity. For decades, many electricity systems were designed around relatively stable assumptions. Demand grew, but it often grew in ways that utilities understood and could plan for over long infrastructure cycles. That environment is changing. The emerging problem is a mismatch between the speed of demand and the speed of infrastructure. This is why I described data centres not as the villains of the story, but as a stress test. They reveal what happens when new load can materialize far faster than the infrastructure required to serve it. A data centre project can move on a one-to-three-year cycle. EV charging infrastructure can appear quickly. New buildings, industrial facilities, and digital services can add load in relatively short periods of time. Grid infrastructure does not always move on those timelines. Substations, transformers, transmission upgrades, permitting, interconnection, and major construction projects can take years. In some cases, many years. That creates what I called the speed-to-power problem. The grid bottleneck: the speed-to-power problem. The issue is not simply whether we can generate enough electricity. It is whether the system can connect, move, manage, and deliver that electricity at the speed the economy increasingly requires. That is a very different strategic problem. The economics underneath the grid are changing too At the same time that demand is accelerating, the economics of energy technology have been undergoing a remarkable reset. In the keynote, I summarized the change in the cost of new renewable power with a deliberately stark phrase: the cost didn't fall, it collapsed. The slide used a roughly 90% decline since 2010 to make the point that the economics of utility-scale solar and other renewable technologies are radically different from what they were only a decade and a half ago. The cost of new power didn't fall. It collapsed. ~90% since 2010. The same pattern is visible in batteries. Average battery pack prices have fallen dramatically since 2010. The slide I used summarized the decline at approximately 93%. Battery costs didn't fall. They collapsed. -93% since 2010. This is one of the most important things leaders need to understand about periods of technological acceleration. A technology can appear to be moving slowly until economics cross a threshold. Once that happens, deployment can move very quickly. Cost collapse changes behaviour. It changes capital allocation. It changes customer expectations. It changes which projects are financially viable. It changes the assumptions buried inside long-term planning models. It changes the point at which one technology begins to compete with another. And then deployment responds. I used a slide highlighting 800 GW of renewable capacity and 108 GW of battery storage added in one year as a way of showing how rapidly the market can react when economics, policy, technology and demand begin reinforcing one another. Deployment responded: 800 GW renewables / 108 GW battery storage in one year. The strategic lesson is bigger than any one technology. Yesterday's economics can change. A business case that made perfect sense five years ago can become questionable. A project that once looked premature can suddenly look obvious. An asset that was evaluated under one set of assumptions may be competing against an entirely different set of economics before the end of its useful life. Long-lived infrastructure meets fast-changing economics. That tension is going to define a lot of decisions in the energy sector. Storage changes more than storage Battery storage is often described too narrowly. We think of a battery as a device that holds electricity until we need it. That is technically true, but strategically incomplete. Storage changes the relationship between energy and time. Electricity generated at one moment can be used at another. A system can charge when supply is abundant or prices are low, then discharge when demand rises, the grid is constrained, or prices increase. That changes the value of electricity itself because value becomes increasingly dependent on when power is available, not simply how much power exists. It also changes how we think about capacity. One of the slides in the keynote made the point directly: A megawatt of capacity doesn't have to come from a power plant. A megawatt of capacity doesn't have to come from a power plant. Storage can increasingly compete with traditional generation for some peak-demand requirements. It can support reliability, respond quickly, shift load, participate in energy markets, defer some infrastructure costs, and provide flexibility to a system that needs more of it. In other words, batteries are not simply storage devices. They are becoming market participants. They can buy. They can sell. They can balance. They can respond. Once you begin thinking about storage that way, the architecture of the energy system starts to look different. The grid becomes less about a fixed chain from generation to transmission to distribution to customer and more about the continuous orchestration of assets across time. Then AI arrives, as both a load and a solution Artificial intelligence adds another layer of acceleration. The story I told in the keynote was not simply "AI is getting better." The more important story is the feedback loop behind it. Compute keeps scaling. Models improve. The cost of useful machine intelligence keeps falling. The data available to those systems keeps expanding. That makes more applications possible, which creates more demand for compute, which drives more investment, which accelerates the cycle again. The result is a powerful feedback loop: AI gets more powerful. AI gets cheaper. AI gets more data. AI innovation compounds. And AI itself becomes a massive electricity load. AI gets more powerful -> cheaper -> more data -> compounding innovation -> electricity load. That last point is what makes AI so important to the energy sector. Artificial intelligence is not just another digital trend happening somewhere else in the economy. It is becoming a physical infrastructure issue because the data centres behind it require enormous amounts of electricity. But that is only half of the story. The same technology putting pressure on the grid is becoming one of the technologies we can use to manage that pressure. AI is both the load and the solution. AI is both the load and the solution. That is the fascinating contradiction. AI can increase demand while also improving forecasting. It can make data centres a major source of new load while helping utilities optimize assets in real time. It can contribute to congestion while also helping predict where congestion will emerge. It can increase complexity while providing new tools for managing complexity. This is why I described AI as becoming an operating layer for the grid. It can sense. Predict. Optimize. Act. Learn. The progression is from a system that is primarily reactive, to one that is increasingly predictive, and eventually to one in which some actions can become autonomous within defined boundaries. From reactive to predictive to autonomous. That does not mean removing people from the system. Quite the opposite. As systems become more capable, the need for judgment, oversight, security, governance and clear operating guardrails becomes more important. More tools create more capability, but they also create more decisions. The future of AI in critical infrastructure is not simply automation. It is automation combined with accountability. The grid itself becomes distributed The next change follows naturally from everything above. The grid is becoming more distributed. Solar panels, home batteries, commercial storage, EV chargers, thermostats, water heaters, buildings and electric vehicles are no longer just things sitting at the edge of the electricity system. Increasingly, they can respond to price signals, grid conditions, software commands and local needs. Devices become participants. And that means the grid begins making millions of small decisions rather than relying only on a relatively small number of large ones. One of the most visible examples is the electric vehicle. An EV is usually described as a transportation technology. But from the perspective of the electricity system, an EV is also a large mobile battery that spends much of its life parked. Vehicle-to-grid technology opens the possibility that those batteries can become flexible grid assets. They can charge when electricity is abundant, avoid charging when the system is constrained, and in some cases return power when the grid needs support. EVs become batteries on the grid. The same logic applies throughout the system. The home can become a small power plant. A building can become a controllable load. A battery can become a market participant. A fleet of EVs can become distributed storage. A thermostat can become part of a demand-response network. This is how the grid evolves from a one-way delivery system into a network of networks. Power still matters, of course. Poles, wires, transformers, substations and generation remain essential. But software, data, devices and orchestration become increasingly important to how the physical system performs. The future is not the hard part. Leading through it is. The final part of my keynote was about leadership, because every trend I had discussed carries a leadership implication. Forecasting the future is interesting. Translating it into what people should do on Monday morning is much harder. That is why I put this challenge on the screen: Your job is to translate the future into today's priorities. Translate the future into today's priorities. What matters now? What has changed? What has not changed? Where do we focus? What do we deliberately choose not to chase? Those questions become more important when the external environment is moving quickly because no organization has unlimited money, unlimited people, unlimited time, or unlimited attention. Leadership is therefore not about reacting to every new trend. It is about building the capacity to distinguish between noise, signal and strategic consequence. That requires operating in multiple time horizons at once. Good leaders operate in three time horizons at once. Organizations have to run today, delivering safe, reliable service and operational excellence. They have to grow tomorrow, expanding capacity and adapting to new customer needs. And they have to transform for what comes next, experimenting with new technologies, new operating models, AI, distributed energy and new forms of resilience. The mistake is thinking those are sequential tasks. They are simultaneous. The organization that focuses only on today eventually gets trapped by tomorrow. The organization that obsesses only about tomorrow can lose control of today. Good leadership is the discipline of doing both while preparing for the third horizon beyond them. That is also why I returned to the technology hype cycle. Every important technology goes through periods of excitement and disappointment. The peak of inflated expectations gets the headlines. The trough of disillusionment gets the cynicism. But neither of those moments tells you whether the technology will ultimately matter. What matters is what leaders do while everyone else is swinging between hype and dismissal. What we do at the peak and trough defines our success. Do we build useful skills? Do we run disciplined experiments? Do we learn what is real? Do we create guardrails? Do we invest at the right pace? Do we know which capabilities could become important before they become urgent? That is how organizations get from hype to actual productivity. Culture is a performance system The final leadership point was about culture. Culture is often treated as something soft, separate from operations. I do not believe that. Culture shows up in how people make decisions, how they communicate risk, how they respond to uncertainty, whether they raise concerns, how they recognize one another, how they handle safety, and whether people feel empowered to improve the work around them. That makes culture a performance system. Culture is a performance system. In a fast-moving environment, culture determines whether people freeze or adapt. It determines whether new technology is adopted thoughtfully or chaotically. It determines whether a team can experiment while maintaining discipline. It determines whether safety and innovation are treated as competing goals or as responsibilities that must be designed together. The most useful leadership question is therefore not abstract. It is practical: What do I do differently Monday morning? What do I do differently Monday morning? Clarify the priority. Explain the why. Define the guardrails. Recognize the right behaviours. Ask what can be improved. Those sound like simple actions. They are. But that is the point. The future is usually not implemented through one dramatic decision. It is implemented through thousands of small decisions made by people who understand what matters, why it matters, and what boundaries they are expected to work within. Think big. Start small. Scale fast. That brought me back to the title of the keynote. Think big. Start small. Scale fast. For a safety-focused organization, I added an important qualification to the middle phrase: Start small, carefully, within boundaries. Think big. Start small (carefully, within boundaries). Scale fast. That, to me, is the right operating philosophy for an era of acceleration. Think big enough to understand where the world is going. Think beyond the immediate planning cycle. Pay attention to weak signals. Challenge assumptions that were formed under yesterday's economics and yesterday's technology. But do not confuse thinking big with betting everything. Start small. Run experiments. Learn. Test assumptions. Build skills. Put the right guardrails around new capabilities. Understand what works and what does not. Then, when the signal becomes clear and the value is proven, scale fast. Because the most important lesson from the energy transition, the AI revolution, storage economics and distributed infrastructure is this: The future rarely waits for our planning cycle. The organizations that succeed will not be the ones that predict every detail correctly. They will be the ones that become very good at recognizing when the future has moved from "someday" to "now," and that have built the leadership, culture and operational discipline to respond. The future of energy is not one technology. It is the convergence of accelerating demand, collapsing costs, new forms of storage, artificial intelligence, distributed assets, software-defined systems and changing customer behaviour. And the biggest risk is still the simplest one: It may all happen faster than our mental model.

---
Source: https://jimcarroll.com/2026/09/the-future-of-energy-is-arriving-faster-than-our-mental-models/