End to End

The electrical energy system doesn’t just replace fuels. It replaces the entire architecture.

The fossil fuel system is a chain of conversions, each one losing energy. Drill crude from the ground. Ship it across an ocean. Refine it into diesel. Truck the diesel to a depot. Pump it into a vehicle. Ignite it in a combustion chamber at 20–30% thermodynamic conversion. The rest is waste heat, noise, and particulates.

The electrical system works differently. It creates DC. It sends DC. It uses DC. And at every stage, the energy stays under digital control.

This is what I mean by “End to End.”

A solar panel on a factory roof generates DC electricity directly from photons. A battery stores it as DC. An EV charges on DC. A heat pump, at its core, runs on a DC-powered compressor controlled by a variable-frequency drive. The inverter switches to AC-only when the legacy grid demands it, and even that is changing. Data centres, the fastest-growing electricity consumers on earth, are now designing 800-volt DC architectures end-to-end because every time you convert AC to DC and back again, you lose 5–18% of the power. Eliminate the conversions, eliminate the waste.

The old system was built around combustion. The new one is built around control.

Controlling Heat Digitally

The largest single use of energy in the global economy is heat. Not transport. Not lighting. Heat. Industrial processes, space heating, and hot water. And for a century, the answer was simple: burn something.

Heat pumps change the question entirely. Instead of generating heat by breaking chemical bonds in gas or oil, a heat pump moves heat from one place to another, using electricity to move thermal energy against its natural temperature gradient. A modern heat pump delivers 3–5 units of heat for every unit of electricity it consumes. That is a 300–500% conversion rate. No combustion process in history has come close.

But the real shift is not just the device. It is the control layer.

An industrial heat pump paired with factory load control software doesn’t simply replace a gas boiler. It becomes a digitally controlled thermal asset. The software reads electricity prices in real time, adjusts the heat pump’s output to consume power when it is cheapest (often when renewable generation peaks), preheats thermal mass during low-price windows, and curtails output during grid-stress events. The factory’s heating system becomes a flexible participant in the electricity market, earning revenue for providing demand response whilst cutting its own energy bill.

The industrial heat opportunity is enormous and barely discussed. Below 100°C — where food processing, chemicals, pulp and paper, and district heating operate — electric heat pumps are already cost-competitive and could displace 200 billion cubic metres of gas per year globally. Between 100°C and 150°C, the technology is scaling rapidly, with another 150 bcm of gas demand in the mix. Between 150°C and 200°C, the next generation of high-temperature heat pumps is entering commercial production, targeting a further 110 bcm. That is 460 bcm of annual gas demand that can be electrified with technology that either exists today or is in advanced deployment. To contextualise: 460 bcm is more than three times the volume of Russian gas that Europe imported before the Ukraine war.

The gas boiler doesn’t know what electricity costs. The heat pump does. That difference defines the transition.

The Grid Symphony

Stand back from the individual technologies, and something larger comes into focus. The electrical system is not a collection of separate devices. It is an orchestra.

Consider what is being assembled, piece by piece, across the leading markets.

20 million solar rooftops in Europe alone, each generating DC power during daylight hours. Behind a growing share of those rooftops, a home battery — Australia expects 520,000 residential battery installations in 2026 alone, adding 12 GWh of distributed storage to one country’s grid. Scale that globally, and the numbers become formidable. Add grid-scale batteries: 130 GW of new battery storage capacity is expected to be installed worldwide in 2026, with systems like Zenobe’s 300 MW / 600 MWh facility in Scotland providing both energy storage and synthetic inertia to replace the spinning mass of retired coal and gas turbines.

Now add EVs. Not as passive consumers, but as grid participants. A fleet of 1 million EVs, each with a 60 kWh battery, represents 60 GWh of mobile storage. Those cars cost consumers roughly $40 billion to purchase. But the distributed battery resource they create — 60 GWh — would cost $12 billion to build as a standalone grid battery farm. That is a $12 billion grid infrastructure asset that’s been created for free, a by-product of consumers buying cars.

With bidirectional charging and coordinated software, those EVs become virtual power plants. They charge when the grid has surplus renewable power — often at night, often at the lowest prices — and discharge during peak demand, earning their owners revenue. In the UK, Octopus Energy’s Power Pack tariff already pays EV owners £320 per year for grid services. In the US, pilot programmes show $1,000–5,000 per vehicle per year in grid revenue. A CEC-funded variable pricing pilot in California achieved 98% off-peak EV charging through price signals alone.

The coordination layer is what makes it work. Grid-forming inverters on batteries provide synthetic inertia — the electrical equivalent of the flywheel effect that spinning turbines used to deliver. Germany is launching a formal inertia services market in 2026, paying battery operators €8–17,000 per MW per year for this capability. Software platforms from companies like Kraken (Octopus Energy’s technology arm) orchestrate millions of assets in real time: solar panels, batteries, heat pumps, EV chargers, and industrial loads, each responding to price signals, grid frequency, and weather forecasts.

This is not a speculative future. Every element I have described is either commercially deployed or set to enter commercial deployment in 2026. The solar panels exist. The batteries exist. The heat pumps exist. The V2G chargers are shipping. The software is running. What is happening now is the wiring together — the moment the individual instruments become an orchestra.

End to End also means the End of the Old Model

The fossil fuel system was linear. Extract, ship, refine, distribute, burn. Value was captured at each intermediary step, and the fuel was destroyed in a single use.

The electrical system is circular. Generate, store, distribute, use, and feed back. The assets last 25 years. The fuel — sunlight, wind, ambient heat — is free. And because electrons are controllable in ways that molecules are not, the entire system can be continuously tuned.

In a typical data centre today, power enters the building as AC, is converted to DC by the UPS, then back to AC for distribution, then back to DC at the server power supply, and finally to different DC voltages inside the chip. Each conversion wastes energy and generates heat that must be cooled, using yet more electricity. Total losses: up to 18%. The new 800-volt DC architectures being developed by Nvidia, Vertiv, and Eaton eliminate most of these steps. SolarEdge is building a 99%-conversion-rate solid-state transformer for exactly this purpose. End-to-end DC. Fewer conversions. Less cooling. Less waste.

The same logic applies at every scale. A house with solar panels, a battery, a heat pump, and an EV charger — all running on DC behind a single inverter — is a miniature end-to-end electrical system. A factory with rooftop solar, battery storage, industrial heat pumps, and energy management software is larger. A national grid with distributed generation, coordinated storage, flexible demand, and digital control is the full expression.

This is the E-Flip made physical. Not just a shift from molecules to electrons, but from a linear fuel supply chain to a circular, digitally controlled energy network. The performance gains are not incremental. An electric motor converts over 90% of input energy into motion, compared with 20–30% for combustion. A heat pump delivers 3–5x the thermal energy per unit of electricity. DC-to-DC power transfer incurs 1–3% loss, compared with 10–18% through multiple AC-DC conversions. Compound these gains across an entire economy, and the result is structural deflation in the cost of energy services.

The $12 billion question

I keep returning to this number because it captures something that most energy analysis misses.

1 million EVs with 60 kWh batteries = 60 GWh of distributed storage. Purchase cost to consumers: ~$40 billion in new car sales. Value as a grid battery asset: ~$12 billion at 2025 battery storage prices. That $12 billion in grid infrastructure was built for free. Nobody budgeted for it. No utility commissioned it. No government funded it. It appeared as a side effect of people buying cars.

Now scale it. The UK has 34 million registered vehicles. If even half convert to EVs with an average 60 kWh battery, that is 1,020 GWh — a terawatt-hour — of distributed storage. The grid value: roughly $204 billion. Created, again, as a by-product of consumer transport spending.

This is why the end-to-end framing matters. When you view the electrical system as isolated components — a solar panel here, an EV there, a battery somewhere else — it looks like a collection of subsidised replacements for fossil technologies. When you view it end-to-end — DC generation connected to DC storage connected to DC consumption, all coordinated by software — it becomes something different. A system where every asset serves multiple purposes, where transport is also storage, where heating is also grid balancing, and where a rooftop is also a power station.

The fossil fuel system could never do this. A tank of diesel serves one purpose: it burns. An EV battery serves three purposes: it moves you, it stores grid energy, and it provides frequency response. A heat pump also serves three purposes: it heats or cools your building, and it provides demand flexibility. Every electrical asset is multi-functional because electrons are controllable. Molecules are not.

As I have written before, changes occur in S-Curves. First, we saw Norway become the first to reach 50%+ EV share in new car sales; now we have the World’s largest car market, China, also at 50%+ EV share. With heat pumps, Norway was again the first country to reach 50%+ share of heat pump sales, and in 2025, heat pumps outsold gas systems for the first time in the huge German market.

End to end. That is the architecture of what comes next.

Gradually, then suddenly.

end

Nadim Chaudhry is the author of ElectroState: How the Electrification E-Flip, China, Geopolitics will Reorder the Global Economy, examining the global transition from fossil fuels to electrification through geopolitical and systems lenses.

ElectroState.com offer premium intelligence services and consulting for the electrical energy transition