2009: It starts with a thermostat
In June 2009, an internet-connected thermostat was enough to offer a glimpse of a different electricity system. Jim Carroll wrote about using an iPhone application to manage connected thermostats in his home and cottage. A thermostat sounds like a small thing to build a vision of the future around. But its significance was not the screen on the phone. It was the idea that millions of individual devices might eventually share information and respond to changing conditions instead of consuming energy without much coordination.
In his June 25, 2009 discussion of the emerging smart energy infrastructure, Carroll connected consumer control to intelligent heating and cooling systems, a smarter grid and analytics that could help households and organizations manage energy demand. He described both the promise and the engineering challenge. The grid had been built around relatively predictable power flows, yet demand continued to grow and constructing new infrastructure was difficult. Intelligence, connectivity and more flexible use of existing assets suggested another path.
That post did not describe the neighborhood of 2036. It was an early exploration of connected energy systems. Looking back, the remarkable thing is how much of the bigger question can be seen in a seemingly ordinary device: could electricity become something we schedule and coordinate, not merely something we consume?
2021–2025: The idea becomes much larger
By November 2021, Carroll’s published work was bringing together a different collection of technologies: renewables, energy storage, microgrids and electric vehicles. The subject had moved well beyond the connected thermostat. A car could become a battery on wheels. Solar panels could turn a building into a producer as well as a consumer. Storage might change when renewable electricity was available to use.
In his July 2025 electrification megatrend, Carroll argued that electrification was rewiring transportation, infrastructure and energy supply chains. He also described the shift toward an intelligent, interactive grid, including vehicle-to-grid technology: the potential for an electric vehicle to return power when the network needs it. A great deal of the technology was still unevenly deployed, and vehicle-to-grid business models remained difficult. But the central idea had grown from smarter individual appliances into a more connected energy system.
It helps to picture the transformation as an orchestra. A traditional grid already has skilled conductors: system operators who constantly balance production and demand. What is changing is the number of instruments they may eventually coordinate. Rather than managing only big generators and an aggregate pattern of electricity use, the system could increasingly draw useful flexibility from millions of small devices—if those devices can communicate, respond reliably and operate under workable rules.
2026: Reality grows more complicated than the vision
Ontario’s electricity planning shows why the transition matters. Its Independent Electricity System Operator forecast 65% electricity demand growth by 2050 in the reference scenario of its March 2026 outlook, down from the 75% estimated in the 2025 outlook. The earlier figure was not simply “wrong”; the outlook’s assumptions changed. Electric-vehicle expectations, economic conditions and efficiency measures can all move the numbers.
The IESO’s 2026 outlook also forecast that data centres could account for 8.6% of Ontario’s electricity demand in 2050 in its reference scenario. That is a modeled share, not observed electricity consumption in 2050. Different scenarios produce different possible futures. The useful lesson is that a bigger grid cannot be planned only by extrapolating one tidy number. The timing, location and flexibility of demand matter, too.
The same issue is visible globally. The International Energy Agency’s 2025 Energy and AI assessment projected approximately 945 terawatt-hours of annual electricity demand from data centres worldwide by 2030 in its base case, roughly twice the 2024 level. It emphasized uncertainty and a mismatch in timescales: technology facilities can be built much faster than major parts of the electricity system. The challenge is not only producing more power. It is connecting new facilities reliably while using the existing system intelligently.
That intelligence has limits. A battery can move energy from one hour to another; it cannot create energy. A flexible factory might shift some processes; it cannot always move every production run. A home might allow a water heater to operate earlier; its owner may not accept remote control under every circumstance. The emerging system is not magic. It is a negotiation among physics, economics, trust, software and individual choice.
2036: Imagine a neighborhood that participates
Now move ten years forward. Suppose a suburban neighborhood has more electric vehicles, home storage and smart equipment than it does today. On a hot afternoon, electricity demand begins to rise. The utility sends a request to an approved group of devices to avoid adding to the evening peak. Some cars postpone charging. Some batteries discharge for a limited period. Equipment that can safely run at another hour moves to a cheaper window.
This is a scenario, not a prediction that every home will be enrolled or that all of these capabilities will become commercially standard by 2036. It follows directly from the possibilities explored in Carroll’s archive and from technologies being investigated or deployed in different forms today. The future could be less centralized, but that does not mean the utility disappears. In fact, coordinating these resources without compromising safety, customer freedom or reliability could make grid operations more complex.
Imagine what the experience might mean to an ordinary household. Electricity would still come from an outlet, yet the decisions behind that outlet could become far more dynamic. The utility might offer rewards for flexible charging. A community battery might help smooth a local peak. New software could make it easier to see when a particular neighborhood needs more capacity and when it has room to spare. The household would not need to become an energy trader for any of this to matter. It would need clear choices, understandable benefits and confidence that essential appliances will work when needed.
The economic consequence could be substantial—but it must be tested locally. Deferring an expensive network upgrade may make sense if enough flexible resources perform when required. In another district, aging equipment or a hard physical capacity limit may still make new transformers, cables or generation unavoidable. The right future is not always the one with the most software. It is the one that reliably meets an identified need at a cost and level of complexity people can accept.
What could redirect the journey?
There are several ways this vision could stall. Equipment from different manufacturers may not communicate well. Utilities may lack the authority, investment incentives or cybersecurity capability to manage large numbers of customer devices. Households might refuse participation if the program is confusing, intrusive or poorly compensated. An unreliable market for flexible capacity could make an otherwise promising technical system uneconomic.
There is an equally interesting counter-possibility: the transition could accelerate in places where devices, prices and rules align. Large building owners, commercial fleets and industrial operators may prove easier to coordinate than millions of individual homes. The first visible benefits might arrive in specific districts and business parks rather than sweeping across entire countries. A very connected power system could remain unevenly distributed for a long time.
This is why the historical thread is useful. The idea first visible in a thermostat was not that one gadget would change the world. It was that information might become as important to electricity planning as hardware. Carroll’s archive lets readers see that idea take shape, broaden, and run into real-world constraints. The future remains open: the pace and the business models have never been guaranteed.
The next signal to watch
By 2036, will the neighborhood itself have become a useful participant in balancing the grid—or will flexible energy remain concentrated in a smaller number of programs and large customers? The decisive evidence will not be another polished demonstration. It will be verified, repeatable deployments that show coordinated devices reducing measured demand peaks, doing so reliably, protecting customers and producing benefits that justify their total cost.
The voyage began with the humble thermostat. The destination may be a community that helps operate the infrastructure it depends on. We cannot know the precise route. But following the idea through time makes the next chapter far more interesting to explore.