Rejected Energy: What the Livermore Chart Actually Shows

The number everybody quotes and nobody reads

Lawrence Livermore put their 2024 energy flow chart out the other day. It says America got through 94.61 quads of energy and threw 62.27 of them away.¹

Right. A quad is a quadrillion British thermal units. Not a quid. There’s no money anywhere in this article, and if your eye keeps reading it as pounds sterling then mine did too. The Americans keep their national energy books in quads and the rest of us work in watt hours, so here is the conversion that makes it mean something: one quad is about 293 TWh, a shade more than the whole UK grid delivers in a year.

Rejected energy is the polite name for those 62.27. The bit that did no useful work. Heat up a chimney, heat off a radiator, heat out of an exhaust. Nearly all of it is waste heat from burning summat.

Now mind the units, because I had this wrong at first and so does most of the internet. 62.27 is quads, not per cent. Put it against the 32.34 quads that did something useful and the share is 65.8%. Two thirds of everything drilled, dug, piped and shipped, gone as warm air.

That last column is the one that landed for me. Great Britain’s grid delivers about 271 TWh a year.² So America’s rejected energy for 2024 alone comes to sixty-seven years of the entire British grid, thrown away as heat, in twelve months.

Put another way: run this country’s electricity from now until 2093 and you still would not have generated what the United States wasted last year.

That number gets quoted all over. How the chart counts gets quoted hardly at all. That’s the interesting bit.

One housekeeping note, this being a piece about America written by a Yorkshireman. Petrol is gasoline. A forecourt is a gas station. Aircon is AC. And when I say nowt or summat, I mean nothing and something.

Wind and solar are counted at 100%

Here’s the bit that catches folk out. Livermore uses the Energy Information Administration convention, and under it four sources go in with no conversion losses at all.³

A gas plant is counted on the energy in the gas, so the two thirds it chucks away as heat lands in the rejected block. A turbine is counted on the electricity it delivers. There’s no “energy in the wind” line to lose against.

So on this chart, every terawatt hour that shifts from a thermal plant to a wind farm does two jobs at once. Adds to useful, takes away from rejected. It counts twice.

Which means the argument that scrapping wind and solar would cut waste runs backwards through the very diagram it gets made about. I’ve had that one put to me in earnest by people who ought to know better. It’s not a close call. It’s the sign inverted.

Where the waste actually is

Split the chart by sector and it stops being an abstraction.¹

The power stations lose another 19.21 quads up the cooling towers before any of it reaches those four.

Transport is the worst by a distance. It takes the largest share of the energy and turns 21% of it into movement. Those 22.35 wasted quads are 36% of everything America throws away. More than a third of the national waste, in one sector.

A petrol engine, gasoline to American readers, turns maybe 16–25% of the fuel into motion. The rest is heat and noise.⁴

That’s not a marginal gain. That’s the difference between a machine that mostly warms the sky and one that mostly moves the car. Same journey, different physics.

And it runs through the whole chain, not just the vehicle. Getting liquid fuel to a forecourt, a gas station in American, costs energy before a drop of it is burned.

Road transport is around half of global oil demand. Electrify that and a fair slice of the tanker fleet has nowt left to carry.

Worth being straight here, because overclaiming is how you lose an argument you were winning. Grid losses are real and they’re resistive heat. An EV pays for cabin heat in winter that an engine gets for free. But that free heat is only free because the engine already binned three quarters of the fuel. It’s free the way warmth off a house fire is free.

The biggest single thing America could do

Light-duty vehicles are 58.5% of transport energy, the bit that electrifies without waiting for new technology.⁸ Run the numbers on swapping the drivetrain and leaving everything else alone.

Even charging every one of them off gas turbines, you save about seven and a half quads. Off wind and solar it is twelve and a half. Eight to thirteen per cent of everything the United States burns, from one swap.

That is the largest efficiency gain available anywhere on the chart, it needs no invention, no breakthrough, no pilot scheme and no new physics, because every single one of the vehicles required to do it is already being built and sold today in volume. The cars exist. That’s the whole trick.

Except a better engine doesn’t fix the layout

An EV still has to cover the distance, and this is where the other half of the problem sits.

Look at the third row, because it takes the geography excuse away. About 30% of daily trips are under a mile on both sides of the Atlantic. Same errands, same distances. Americans drive seven in ten of them. Europeans walk, cycle or catch something for seven in ten of them.

Geography didn’t do that. Weather didn’t either. Zoning did, putting the houses here and the shops three miles over there, backed by parking minimums that ended up building nearly eight spaces for every car in the country.⁹ Between the 1920s and the 1960s American cities were rebuilt around the motor car and much of western Europe copied them. From the late 1960s Europe stopped, and started undoing it.⁹

So the twelve and a half quads is the ceiling on electrification alone. Halve the miles as well and you halve what is left. One is an engineering job and the other is a planning job, and the planning job is the one nobody can buy their way out of in a single purchase.

And the cheapest passenger-mile is a shared one

There is a third lever, and America has more or less stopped pulling it.

Four to six times better, before anybody changes a drivetrain. A single-occupant car puts out 7.7 times the CO₂ per passenger-mile of a full coach.¹⁰

Now the state of play.

Nought point four per cent. Not a transport system with a transit component. A country that drives, with some buses in it.

And 1.7% electrification means American rail is still, overwhelmingly, diesel. So every argument for shifting freight and passengers onto rail is being made about a network that still runs on oil. Electrify the track and you get the mode shift and the fuel switch out of the same job.

Here is the honest bit, because it cuts the other way and someone will raise it. Transit is only efficient when it is full. Bus occupancy in the States has been falling for decades, and energy per passenger-mile on buses has gone up 63% since 1970.¹⁰ A near-empty bus on a fifty-minute loop round a subdivision is worse than the car it was meant to replace. That’s a real number and a big one.

But look at what makes a bus empty. Nobody within walking distance of the stop, nowhere worth walking to at the other end, and a layout that puts eight parking spaces by every door. Empty buses aren’t a fact about buses. They’re a fact about what got built around the stop.

Which brings you to the thing that actually shifts people out of cars, and the gap there is wider than the mode-share number suggests.

Thirteen. In a country of three hundred and forty million people. Europe has sixty, and has been laying new track three times as fast since the turn of the century, so the gap is widening rather than closing.

The mode matters, too. Work on European cities found metros shift people out of cars in a way tram networks largely do not, which fits what you would expect: a metro is faster than driving at rush hour and a tram usually is not.¹² Speed is the whole product. Build something slower than the car and you have built a subsidy for people who have no choice, not an alternative for people who do.

That is the one genuinely expensive item on the list. Zoning reform costs political will and a redraft. Tunnels cost billions. But it buys what the other two can’t. You move people across a dense city on electricity, at 0.3 to 0.6 MJ a passenger-kilometre, and faster than they could have driven it. At that point leaving the car at home stops being a sacrifice and starts being the obvious move. That’s when people actually do it.

Which means the fixes are the same fix wearing different hats. Electrification takes 7.5 to 12.5 quads off the drivetrain. Zoning takes the miles down. Density is what makes the transit worth running, and the transit is what makes the density liveable. Pull one lever and you get one lever’s worth. Pull all three and they multiply.

America is currently arguing about the first one.

None of it starts, though, without the cheapest step of the lot, which also looks like the hardest. Somebody has to say out loud that there’s a problem.

The diagnosis isn’t missing. It gets published every year by a federal laboratory, for free, on a public website, in a diagram plain enough to read in a minute. Sixty-five point eight per cent wasted. Transport a third of it. Twenty-one per cent efficiency on the largest block on the page. Nobody needs to commission a study or wait on the science. The science came out in August. People read the headline number, got it wrong, and moved on.

That’s the bit that ought to sting. No country spends billions tunnelling under its cities to fix a thing it reckons is fine, and America has quietly filed 22.35 wasted quads a year under fine. Not argued over and dismissed. Just never put on the table.

The leftovers have to go somewhere

Heat is only waste if there’s nowhere to put it. That’s a planning decision, not a technology gap, and it were mostly made decades back.

Europe runs about 111,650 commercial and industrial sites on transcritical CO₂ as of 2025, roughly a third of all food retail.¹⁵ Meta’s Odense data centre has been pushing around 100,000 MWh a year into the local network since 2019. That is heat which would otherwise have gone up a dry cooler. It warms 12,000 homes instead.¹³

We’ve got about 14,000 heat networks in the UK and they still only meet 3% of heat demand.¹⁴ Fourteen thousand of the things and almost nowt to show for it, because they’re small, fragmented and mostly bolted onto social housing. Ofgem took over regulation in January and zoning starts this year. Target 7% by 2035, about a fifth of building heat by 2050.¹⁶

Denmark isn’t doing summat clever we can’t. Denmark put pipes under the streets. We didn’t. That’s it. That’s the difference.

Heat pumps, and the refrigerant nobody expects

Same logic at the small end. A resistive heater can’t beat a coefficient of performance (COP) of 1.0. That is the definition of the thing. A heat pump moves heat rather than making it, so it does better.

ADAC put 28 EVs through a winter test at −7 °C. Heat pump models averaged 22% less range loss than the resistive-only ones.¹⁸

The R-744 line is the one worth a second look. That’s carbon dioxide itself, working as the refrigerant, in the VW ID.3 and ID.4. Higher suction vapour density keeps capacity up as it gets colder, which is exactly when you want it.¹⁷

And refrigerant grade CO₂ is a byproduct of ammonia, ethanol and fertiliser production, captured and cleaned up rather than vented.¹⁹ A waste stream doing the heating and cooling, at a global warming potential of 1 against R-134a’s 1,430. When that one leaks, nowt happens.

It’s not carbon capture and I’ll not pretend it is; the charge is under a kilo. The point is narrower and better. The working fluid is summat we already had too much of.

What I actually run

Nowt exotic in that list and nowt new. Same principle as matching storage media to an IO pattern. Put the energy where it earns its keep, measure what you actually get, and stop trusting the label on the box.

The bit that surprised me was how much of the saving came from not moving fuel about. No tanker, no forecourt, no refinery taking its cut on the way through. The panels are thirty feet from the car.

The chart is a mirror

Livermore have published these for years and they’re good. Honestly built, genuinely useful, free. Worth an hour of anybody’s time.¹

But a Sankey diagram has no opinion. It shows you what a country decided to do with its energy, drawn to scale. The 65.8% isn’t a law of physics. It’s a picture of choices about engines, pipes and planning, taken one at a time over about seventy years, and it’d look different if the choices had been.

Denmark’s chart looks different because Denmark dug.

Sources