Has It Held Up?

MacKay wrote this book in 2007–2008. He died in 2016, so he has not been able to revise it himself. This edition tracks the numbers instead — see Changes for the mechanical, chapter-by-chapter list of every figure that's been updated, with sources.

This page is the other half: not what changed, but what it means. Eighteen years is long enough for a "state of the art" number to become a museum piece, and short enough that the book's argument — physics first, adjectives never — should still be checkable against what actually happened. So: has it held up?

Short answer: the method held up close to perfectly. A good number of the numbers did not — and they didn't move randomly. Almost everything MacKay was skeptical of got cheaper and bigger faster than he assumed possible, while almost everything he was resigned to as a stopgap (coal, "clean coal" chief among them) got smaller and more clearly finished than he assumed too. The sections below go through the book's major bets one at a time.

Solar: the single biggest miss in the book

MacKay's solar chapter is the one that has aged the most, and it isn't close. In 2008 he compared solar to the biggest farm he could find — a 6.3 MW site in Bavaria — and concluded that covering enough of Britain to matter was "beyond the bounds of plausibility," that solar electricity cost four times the market rate, and that the panel capacity you'd need for 50 kWh/d per person was more than 100 times all the photovoltaics in the whole world. World solar capacity in 2007 was about 10 GW.

By 2024, world solar capacity was around 1,900 GW — roughly 180 times bigger — and unsubsidised utility-scale solar cost about $61 per MWh, cheaper than new gas generation at $76/MWh.¹ Panel efficiency, which MacKay put at "about 10%, expensive ones 20%," is now 20–22% for a standard panel. The 100-times-the-world comparison that structured his whole skepticism about solar's scale is now inverted: the capacity his UK scenario needs is comparable to, not smaller than, what the world has already built.

The physics MacKay used — watts per square metre of panel, of roof, of desert — hasn't moved; that arithmetic is why this edition hasn't touched most of his per-area figures. What moved was the cost curve, and it moved further and for longer than almost any 2008 forecast, MacKay's included. See 6 Solar and D Solar II for the specific figures, five and one of which now carry inline updates.

Wind: the same story, one order of magnitude smaller

World wind capacity went from 74 GW at the end of 2006 to about 1,131 GW by the end of 2024 — a 15-fold rise, not the doubling-of-the-world MacKay's UK scenario implied. Offshore wind is the sharper case: in 2008 there was one deep-water UK installation, an experimental two-turbine prototype at Beatrice sending all its power to a nearby oil rig, and MacKay quoted the government's own 33 GW ambition being dismissed in the press as "pie in the sky." By 2024 the UK had about 15 GW of offshore wind actually built, global offshore capacity was about 83 GW, and the 2024 UK auction cleared offshore wind at roughly £73–91/MWh² — cheaper, at that point, than new nuclear.

Unlike solar, wind's underlying resource density (MacKay's ~2–3 W/m²) hasn't moved much; what changed is turbine size (a "typical" 1 MW machine in 2008 is a museum piece next to a 15 MW offshore turbine today) and, again, cost. See 4 Wind, B Wind II, and 10 Offshore wind.

Cars and batteries: the prediction that came true

This is the chapter where MacKay went furthest out on a limb, and it's worth reading his exact words from 2008: "It thus seems to me that the range problem has been solved by the advent of modern batteries… an energy density of 120 Wh per kg is already good enough." At the time, electric cars were a footnote — he introduces them as something to "discuss in Part II," not something anyone drove.

By 2024, EVs were over 20% of new car sales worldwide (45% in China, ~25% in Europe),³ and commercial battery packs deliver 150–350 Wh/kg, one and a half to three times MacKay's "good enough" figure. A production 2024 Tesla Model 3 gets more range from roughly the same battery mass MacKay used in his worked example.⁴ Whatever else has dated in this book, "the range problem has been solved" reads today as a correct call made fifteen years early. See 3 Cars and A Cars II.

Nuclear: the physics case held, the economics case flipped

Global nuclear output is almost unchanged since 2008 — about 364 GW of capacity now versus 369 GW then, generating roughly the same share of world electricity. What changed is the story around it. In 2008 the fast-building country was France; now it's China, building at a pace the West hasn't matched since the 1980s. Identified uranium resources have been revised up substantially (4.7 to 7.9 million tonnes) since the 2005 data MacKay used, so the fuel-abundance case is, if anything, stronger than he presented it.

But the book's own cost argument has reversed. Chapter 27's "Plan E" bets that nuclear beats wind on price — MacKay writes "when 'clean coal' and nuclear go head to head on price, it's nuclear that wins… I've assumed that onshore wind costs about the same as nuclear." In the UK's 2024 auction, offshore wind cleared at about £73–91/MWh against Hinkley Point C's roughly £92.50/MWh base price (considerably higher once escalated to current prices).⁵ Wind is now the cheaper option MacKay assumed nuclear would remain. And the book predates Fukushima (March 2011) entirely — its death-rate framing draws only on Chernobyl. The post-Fukushima consensus figure, about 0.03 deaths per TWh, still supports MacKay's underlying claim that nuclear is among the safest sources by any honest count; it's the citation that needed catching up, not the conclusion. See 24 Nuclear?.

Fossil fuels and carbon capture: his skepticism aged well

Global coal reserves have been revised down by roughly a third since 2008 (1,600 to about 1,074 billion tonnes), while global coal consumption kept rising to a record 8.5 billion tonnes in 2023 — both moving, but in opposite directions from what a simple "reserves running out" story would predict. Britain's own coal story ended outright: the last deep mine closed in 2015, the last coal power station in September 2024.

The more interesting case is carbon capture. MacKay's 2008 framing was a single prototype — Vattenfall's plant, opened weeks before the book's publication — treated as barely real. By 2024 there were 65 commercial CCS facilities worldwide, capturing about 57 million tonnes of CO₂ a year.⁶ That sounds like vindication for the technology, until you compare it with global emissions of roughly 37 billion tonnes a year: CCS grew from a rounding error to a slightly bigger rounding error. Tellingly, when chapter 27's energy plans got properly recomputed for this edition, "clean coal" — a real component of three of MacKay's five plans — had to be removed entirely and its capacity reallocated to wind, solar, and desert imports, because retrofitted coal CCS never happened at any scale and the UK no longer has any coal plants left to retrofit. MacKay's evident distrust of "someone will invent large-scale CCS in time" reads, in hindsight, as one of the more prescient calls in the book. See 23 Sustainable fossil fuels?.

The energy plans: the structure held, the winner changed

Chapter 27's five plans for Britain are the book's synthesis — the moment where all the earlier chapters' numbers get assembled into an actual stack. Recomputing them for this edition, the structure of the exercise needed no rework at all: you still have to pick some big things, and the arithmetic of "this many GW gets you this many kWh/d/p" still works exactly as MacKay set it up. What moved is which big things are now cheapest. Plan E's cheapest-wins logic, which pointed at nuclear in 2008, now points mostly at wind. The "clean coal" removal above reallocates roughly 16 kWh/d/p across three plans. And the 2008 starting line has itself shifted: UK renewables generated just over half of UK electricity for the first time in 2024 (50.8%), up from the roughly 4% figure quoted in chapter 18 — a number MacKay used, in 2008, as evidence that Britain's renewable effort so far was trivial. It no longer is. See 27 Five energy plans for Britain and 18 Can we live on renewables?.

What he couldn't have seen coming

Not everything moved in the direction of MacKay's skepticism being vindicated or his optimism being rewarded. US data-centre electricity use — a footnote in 2008 at 0.4 kWh/d per person — reached roughly 1.5 kWh/d per person by 2024, driven substantially by AI workloads, a demand-side surprise nobody's 2008 stack had a line item for.⁷ And Fukushima itself belongs here too: not a number that dated, but an event the book simply predates, reshaping nuclear's political trajectory worldwide in a way the safety statistics alone don't capture.

Where he was simply right

The book's method — energy in physical units, per person, compared like for like — hasn't been falsified anywhere in this audit; only re-priced. His skepticism of hydrogen cars, of biofuels at any meaningful scale, and of vague renewables boosterism ("Britain has wonderful renewables") all still read as correct. So does the book's most-quoted line, and its least dated one: if everyone does a little, we'll achieve only a little. Adding up small gestures was never going to work in 2008, and it still doesn't.

Where the world moved fastest

If there's one lesson in assembling this page, it's that the numbers which moved were almost never the physics — they were the economics, and MacKay was explicit about the difference: "This isn't a book about economics, but…" is a phrase that opens several of his most dated passages. Solar, wind, and battery costs fell further and faster through the 2010s than the International Energy Agency's own forecasts predicted, revised upward nearly every year of that decade. MacKay's physical intuition about what's possible has held up about as well as a piece of quantitative writing can. His economic intuition about what's cheap — the part he flagged, in his own words, as outside his expertise — is the part eighteen years has mostly overtaken.

---

That, in turn, is a reason to keep this page current rather than treat it as a verdict: the book's own argument is that you should track the numbers, not get attached to an answer. See Changes for the detailed, sourced, chapter-by-chapter edits this overview draws on.

Sources

Footnotes

1. Lazard, Levelized Cost of Energy+, June 2024: unsubsidised utility-scale solar PV ≈ $61/MWh vs. $76/MWh for new combined-cycle gas, 2024. ↩

2. UK Contracts for Difference Allocation Round 7 results, 2024; World Nuclear News, Hinkley Point C cost coverage, 2024. ↩

3. IEA, Global EV Outlook 2025: EV sales exceeded 17 million globally in 2024, over 20% of new car sales worldwide. ↩

4. EPA / InsideEVs, 2024 Tesla Model 3 Long Range specifications. ↩

5. UK CfD Allocation Round 7, 2024; World Nuclear News, "Hinkley Point C cost rises," 2023–2024. ↩

6. Global CCS Institute, Global Status of CCS 2024. ↩

7. Lawrence Berkeley National Laboratory, 2024 United States Data Center Energy Usage Report. ↩