Hornigold’s Paradox: The incumbent fossil fuel industry is too large for our near-term energy needs, but far too small for our long-term requirements.

Hello all – we have been away in 2021 doing many other things – check out new automotive , our attempt to accelerate the electrification of transport in the UK, ember climate  our support to shifting the world power system from coal, and Subak – a new start up accelerator for climate change companies across many sectors.

Anyway  – all that means very little happened on the blog – so as payback, a long post here on the state of the transition.

My thanks to the Physical Attraction podcast over the past year for inspiration – and I hope you can enjoy it too.

Here we go.


On his Physical Attraction podcast (and website) the excellent host, Thomas Hornigold, a physics and maths specialist and writer, covers a vast array of topics from the origin of the universe through the Roman Empire to bitcoin and economic theory.

One area he also covers in depth is climate science and the energy transition (he labels it Climate 201).

Throughout this theme he posits, half-jokingly, a law of the energy transition as “Hornigold’s Law”.

Stated simply it is this: “any new energy system for the planet will likely have to be as large and widely distributed as the existing one”

Or – put another way, “the system to stop producing carbon or remove it will have to be as big as the one currently producing it”.

Unpacking this “law” is helpful as it highlights the challenges and issues ahead.

But first – is the law broadly correct ?

For example, could we not build a better energy system that is far more efficient and far smaller than the current one ?

After all – we know one of the major downsides of the current system is that a large percentage of the energy we consume does not end up as useful energy – in fact roughly 60-65% of the energy we consume is lost as heat and greenhouse gases.

We consume (not use!) approximately 14 billion tonnes of fossil fuels per year – see here from oil company BP .

(and hence with the addition of oxygen put about 35 billion tonnes of CO2 into the atmosphere per year also)

Add all the weight of processing structures of platforms and pipelines, power plants and retail pumps and tankers moving it about we can assume today’s energy system is north of 15 billion tonnes in weight, with the high majority being the annual fuel additions.

Hornigold therefore looks to see what actual systems exist to do this huge act. And he highlights wind, solar, storage and batteries.

Not because they are simply as “heavy”  (excluding fuel use) as the existing system –  although you can argue “heavy” equals “real, and existing”-  but because they can be deployed at similar scale to the existing system, are mature technologies that are already developed, and are more efficient and relatively carbon-free.

And to this last point – which we will come back to – just assume CO2 was a benign molecule that poses no risk to the environment – would we still move to this new energy system ?

Likely answer – yes: it’s cheaper, more resilient, and more equitable for more nations – though in the near-term, that creates  some profound political issues.

The latest Lazard Levelised Cost of Electricity for example shows wind and solar to be as low as even the operational costs of fossil fuel power plants – and a fraction of the cost of new-build ones.

As a thought experiment suppose today’s global energy use of about 160,000 TWh was magically comprised of 50% wind and 50% solar.

To be serious we could only really get to say 70% of that total in the next 20 years – so about 110,000 TWh.

Now – we would not need 110,000 TWh – because fossil fuels only get to about 35% efficient – so we only need to get to about 40,000 TWh pa energy use. Divide by 2 then wind and solar each have to provide eventually 20,000TWh by 2042, or about 1,000TWh new capacity per year each.

Using today’s existing best technologies in wind –  eg GE’s Haliade-X 14MW wind turbine that would equate to about  1 billion tonnes of turbine weight globally, and considering solar  – top-end 500W solar panels required would be about 500 million tonnes of commercial PV panels.

(This assumes 3,000 tonnes per 14MW turbine (blades, plus nacelle and tower), and 25 kilos per 500W solar panel.)

In GW capacity terms that is about 400GW pa for solar and 200 GW pa for wind, taking into account what is already installed. That is not too far from today’s installation rates.

But here is the kicker.

Once all of that is built and installed we are done for likely 20 years as the system is in place, and does not need to be replaced for 2-3 decades until the next generation of technology is worth transplanting it.

This new structure is the energy system – not today’s vassal power buildings passively awaiting the influx of billions of tonnes of fossil fuels which are then inefficiently burnt.

Turbines and solar panels create energy from the natural resources – essentially infinite : photons and air molecules. Put turbines and panels in place, and wind and daylight create our energy.

Of course, we will still need fossil fuels for other requirements such as petrochemical manufacture and hard-to-abate sectors such as shipping, heating and air flight.

But to focus on these – as the incumbents like to do – misses the giant paradigm shift in energy underway.

So Hornigold’s Law is correct in a deep sense – we are indeed having to build a comparable heavy-weight system to replace the old.

But once it is done and in place, it will be paradoxically about a hundred times lighter than the current one.

Even if perhaps more sprawling it does not need to feed standing structures with heavy-weight fuels, and fuel supply chains: it gathers the energy from the cosmos around us, and converts it right there.

Hornigold’s law is predictive but also practical: we might be able to create an even more efficient energy system via some vast energy source like cold fusion that up-ends all our ideas of energy production and consumption – but the law is betting that is not the way it is going to happen.

Today it is clear that to re-create a carbon-free energy system we are going to need, at least in the short-term, a big new set of infrastructure to replace it


Humanity is now a vast energy consuming system – we use over 160,000TWh of energy per year which is 20,000 kWh per person per year – at home, travelling, at work.

As noted, that takes billions of tonnes of fossil fuels per year to maintain, and triple that in terms of CO2 emissions – stopping that, replacing that, improving that system is not (likely) the work of a simple singular new technology suddenly introduced.

The current system has a lot of inertia as Hornigold puts it – it will take a large effort to replace it, large in terms of effort and materials and planning and policy.

If we want to make this energy transition efficient and effective and timely – the quick deployment of an equivalent system of energy generation that can replace the incumbent energy system better be one that more or less exists at scale today: wind, solar, storage fits this bill.

We cannot rely on a magical “technology X”, or the fossil fuel industry’s favourite of carbon capture – adding dead weight to a groaning technology.

To conclude this section – Hornigold’s Law highlights that the new energy system should be an equivalently heavy (mature/ existing) one that aligns with our existing capabilities – ie mining, manufacturing, marketing and investment and policy support.

We may envisage new alternatives that break the fossil fuel paradigm, but the wind / solar / storage / battery system passes the “billion tonne” test, and then shows how it surpasses the old system in a major new way: constant learning, CO2-free, equitable, fast to implement and so on.

On that last point a final word on metrics of progress.

We generate today about 3,500 kWh of useable energy per tonne of energy system per year.

Tomorrow the system can generate ten times this much – perhaps 40,000 kWh per “tonne”.

It is a crude measure, but it highlights the energy paradigm shift – and as we will see now, there are better measures and better impact of the net energy system such as jobs and global impact.

The false hope of a bigger balloon, and the magic of learning

In another of Hornigold’s podcasts he mentions the parable of the air balloon trying to go to the moon: a civilisation on Earth attempts to get the moon by building a bigger and bigger balloon, and of course it ultimately fails, whilst convincing itself of progress.

So an energy paradigm shift is required, not incremental incumbent energy technology progress.

The energy system based on fossil fuels seems to fit this story.

If we are to have an energy system fit for centuries, not just the next few decades, and free of the downsides of CO2 emissions and the increasing political risks of widespread import dependence, then fossil fuels can, like an ever bigger balloon never provide this.

And whilst we can discuss carbon capture and storage (CCS), or electrifying oil platforms or improved petrol efficiency in cars these all look like just larger balloons, doomed to never reach the moon.

The paradigm shift in energy is from extraction to manufacturing.

From coal, oil and gas to wind, solar, batteries and storage and the electrification of the global system.

This moves energy from the limits of fossil fuel’s finite nature: reducing returns to scale, and vast CO2 emissions,  to its polar opposite: infinite fuel, manufacturing learning curves and zero CO2 emissions from energy generation.

A recent paper from Oxford University’s New Economic Thinking unit outlines the power of the learning curves of (large scale) manufactured energy ie wind and solar  – almost magical in their outcome in terms of cost and deployment (recall that compound interest is often called the eighth wonder of the world for similar reasons).

Learning – and Unlearning: Manufacuring Energy or Extracting It 

Wind and solar alone today have almost 2 TW capacity deployed and supply well over 10% global electricity – they also continue to grow at ca 10% pa

Contrast this with the unlearning curves of massive infrastructure systems especially oil and gas (and nuclear) projects which are structurally prone to cost and schedule over-runs, and often only justified by taking big bets on future energy price spikes (or high price contract terms).

And worse

To meet 2030 emissions targets, whilst sticking with a mainly fossil fuel system as per IPCC and IEA scenarios,  will require one new CCS facility to be built every week for a decade – adding millions of more tonnes of construction for no increase in energy production, reducing energy per tonne even further – a system going backwards.

So for those sceptics of how many turbines and PV panels we can build – please apply the exact same logic and simple arithmetic to CCS plants.

So far since the 1970s, when CCS technology was developed, only about a handful of these plants have ever been built, and then half-heartedly – and why bother – it just adds cost to an already costly system.

Spoiler alert – CCS will never take off as a technology – I’ll put money on that. As David Mackay noted – it is the last thing we should be doing. We have a hundred better ideas than this.

It would be as if in the balloon parable the civilisation convinced itself to stop building bigger balloons and add more weight to the basket instead. And still expected to get the moon (perhaps because incumbent balloon manufacturers convinced everyone that bigger baskets would help to get there).

The main paradigm shift in this energy transition is the dematerialisation it will entail: when the new energy system is built out it will feed itself not through a vast network of upstream and downstream infrastructures requiring the lumpen movement of coal, oil and gas, but via the quiet conversion of near weightless particles of light and air.

Yet Hornigold’s Law holds overall: thousands of wind turbines and millions of solar panels is still a vast system., and much of it will have to be replaced every few decades until we find something better or smaller.

Which leads to a final conclusion of Hornigold’s Law – the transition will likely be disruptive.

The Transition Ahead – Two Large Systems Colliding

The reason that the system will be likely highly disruptive is because any new and large and distinct energy system, which is manufactured and essentially by-passes fossil fuels, is by its nature hugely unsettling to the large existing one: the two are unalike, and therefore will be in conflict.

Many historians of energy (eg Vaclav Smil, see below) seem to expect energy to just add layer upon layer of fossil fuel infrastructure to existing ones in some grand sedimentary arrangement.

They seemingly cannot conceive of how a new system can evolve and replace it at scale and quickly. A complete failure of imagination – or put another way, a failure to work out  rudimentary exponential growth calculations.

Flyvbjerg calls this a process of “Strategic misrepresentation” – or basically political bias. To subjectively twist data toward a preferred outcome to gain favour with some invested group – whether governments or industry bodies.

How does this come about ?

The new large-scale energy system of wind and solar and storage has grown gradually, and now suddenly – and the companies and techniques in the system are totally alien to the incumbents.

This is causing major reaction as the incumbents now see the dilemma they are in – they assumed the new entrants were too small to cause significant change.

And initially the new entrants did only cause small impact  – until they grew very quickly: wind and solar are now 10% of global electricity supply from practically zero in 2010: and will be 20% likely by 2025. All that market share gone for the incumbents in the space of 60 quarters of financial reporting, barely half an individual CFO’s career.

Our current energy system is based on the combustion of fossilised fuels – thus it benefits a few countries who can export those in great quantities (essentially OPEC plus Russia), and large corporations who have massive capital and technical expertise.

Yet for all that installed capacity, the global fossil fuel energy system is a rarefied, fragile one.

Adding up the populations of the OPEC+ countries and the employees of the global oil industry tallies to less than 500 million people or about 6% of world population.

Contrast with the advent of manufactured energy: more or less everyone can now generate home-grown energy if their government or industries can afford to buy solar panels, turbines, batteries and so on, or make them themselves or find some sort of hybrid. 

It is perhaps the difference between the distribution of high-end passenger cars across the world (about 10% of people own them) and mobile phones – essentially at 100% ownership.

Hornigold’s Law is saying that not only is the new system at the same scale of the old one, but by its nature it may be large in a very different way – the current system is a dense large concentration of power (coal mines, gas plants, nuclear plants, oil platforms), in a few areas (OPEC, Russia, US); the new system is a vast sprawling array of solar and wind plus battery systems that can spread around the world and across countries.

To see this play out in reality in the UK look at this great graphic by Carbon Brief which shows the evolution of the UK energy system from dense pockets of mining and gas and nuclear power generation to massively distributed wind and solar farms.

Now play this out worldwide, and you can see how this all could develop.

And in this UK microcosm we see why the transition will be bumpy and hugely resisted.

The UK itself has seen regional shocks: coal-mining communities hollowed out, oil producing regions now with declining workforces.

But the new technologies are now providing effective employment eg in offshore wind and electricity services for EVs.

These technologies already provide as many or more jobs as the incumbent ones.

But they are distributed, unlike those of the fossil fuel industry which are concrete, tangible.

As an example  – a small new energy services start-up such as Octopus energy, founded in 2015, already employs (at 1,300) double the number of UK staff than that at the largest oil refinery in the UK at Grangemouth, originally built in 1924 (650).

This is not surprising.

A vast but distributed energy system based around the globe is bound to require a wide range of general to specific jobs to support it versus a densely-packed core of deep specialists in the complex, expensive fossil-fuel burning system.

In fact most oil and gas (and nuclear) companies pride themselves in low headcounts as a sign of efficiency (employees per barrel) and lower risk – lower people to machine ratios, given the dangerous nature of extracting and transporting highly flammable hydrocarbons at high pressures.

Thus the key metric for the energy transition will perhaps move from tonnes per kWh to employees per kWh, from giant combustion structures, to manufactured components and services.

And that will reflect the move from the energy “winners” of the past being lucky petro-states and international corporations, with massive balance sheets and (literally massive) resources, to the new energy leaders being manufacturing experts and clever new smaller people-intensive energy services providers.

Bluntly put – a move from from OPEC and Russia (and the US to an extent – “freedom molecules”), to China, Europe – and back to the US (“freedom energy services?”)

And yet.

It does not require a degree in political science to see how this will not go down well with a few well-funded incumbents and so not be smooth.

Take the UK again.

We can see the potential profile of how this will play out globally (not exactly but at least in broad terms).

The new energy system – wind, solar, batteries, storage – is large but distributed,  thus like any exponential growing system it can be ignored at first, ridiculed even, but then suddenly become large enough to create an impact.

As an example the removal of coal from the UK grid almost overnight – from 40% of UK power in 2012, and for decades before, to essentially zero today due mainly to wind and solar growth along with a switch to gas (we’ll come back to that “zero” in a minute).

During that sudden change oil and gas companies effectively teamed up with renewables as they could see a new wedge of new demand they could bridge, and assumed it was a wedge they could dominate.

But now that wind and solar have eaten into that new wedge, and UK gas as a percentage of power generation is lower than it was in 2016, the oil and gas firms are starting to put a significant effort into “discourses of delay”.

This ranges from promoting the new heavy post-modern fossil fuel technology of CCS and hydrogen (today almost exclusively produced via fossil gas), through to poster campaigns about investments in EV infrastructure and the like.

But Hornigold’s Law suggests these techniques are likely doomed, for all the lobbying power of the incumbents and their capital and expertise.


Largely because the new energy system is totally alien to the incumbents – and not the ersatz version of their own energy system that they are trying to create – adding costly CO2 extraction to an already inefficient process is hardly a winning strategy.

The new energy system is alien in many ways to the current one: geographically (it is everywhere, not focussed on a few endowed locations), technologically (efficient manufacturing versus complex extraction techniques), economically (returns to global scale and service provision versus singular remote mega-project investments), chemically (expertise in lithium and cobalt contracts versus oil reservoir negotiations) and supply chain and organisational (software, manufacturing plant and IT-platform development versus large-scale industrial engineering concepts involving huge on-site construction).

When oil and gas companies speak of the new energy system they still either use their in-house analogies of ‘CCS plant or hydrogen hubs” – ie heavy engineering plant suited to their expertise and skills and revenue – (to a hammer, all problems are a nail)

Or, they develop a veneer of investment in the “future” of EV infrastructure, or solar.

We say veneer not in a critical sense but a precise one, informed by the work of eg Clayton Christensen who noted how difficult it was for successful incumbents to transform quickly – even in lighter-weight industries such as photography (Kodak), and communications (Nokia, Blackberry) and software (Intel).

It is a veneer to the outside world who looks at the basic statistics, versus perhaps a deeply-held strategy internally to those (invested executives for example) who actually believe there is a new direction underway.

Consider this – until 2015 or so maybe a few hundred million dollars of oil and gas company investment in solar or wind was impressive (even though still only a 1-2 % of their capital base).

But now the solar and wind industry commands over $350bn investment per annum – the time when dabbling in the renewables industry is now either over for oil and gas companies, or it is at the point when they have to get serious and either buy a major renewables player, or admit they are now, forever, oil and gas companies, and stop pretending they are a hip and cool energy player.

Simply put either be Exxon, Chevron, Statoil (Equinor) and so on, and take the societal and economic risks of the energy transition, and you stick to your guns – or keep kidding yourselves – perhaps the Steve Buschemi strategy.

Yet both are doomed by Hornigold’s Law – and they probably end in the same way.

Over the past 5 years an investment in the S & P index would have yielded a 125% return: an investment in the major oil companies – progressive or staunchly conservative  – would have resulted in zero or negative returns of over -30%.

And so to that point of “zero coal” in UK power generation. It is actually about 1-2% of UK power supply now, but headlines still abound about a new proposed regional (Cumbrian) coal-mine, and excite debate.

And likewise, the oil and gas industry in the north-east of Scotland, 1% of UK employment and 1% of GDP, still creates enormous UK national political angst and polarisation.

This will be the same globally – West Virginian mines are a political hotbed in the US, China’s coal dependence a global emissions issue, Australia’s coal fixation undermining progress to the new energy system, and incidentally development of their non-fossil fuel energy system, blessed with space, wind and sunlight.

This is all to say that the transition is not simple – actors good and bad will make moves and mistakes, with various intentions – see oil and gas price spikes in Europe and elsewhere this winter for evidence.


A new substantial energy system is now available to us, one that follows Hornigold’s Law – it is large, mature, effective, deployable, relatively carbon-free, more efficient and can scale allowing more equitable energy use globally.

We have covered how it will meet geo-political and incumbent-driven headwinds – but overall its growth path is likely now too strong to be severely diverted

Gradually then suddenly as they say.

Just do the simple math – reserves to production ratios for oil are about 40 years – for sure we seem to be able to keep that steady (ish) but soon the accounting methods will get more stringent and firms will run out of ways to unearth balance sheet resources and large oil fields will yield to their geology and then that 40 years then soon starts to look threadbare – try selling a house with a 40 year leasehold on it.

And ok – double that number – triple it for fun – so we have 120 years of oil reserves left – imagine being told that as you enter the world in 1900 that by 2020 all (economic) oil will be gone. To recall – the best fields already exploited, the learning curves flat – no shiny tech will soon be along to fix this, the fracking uptick in the US is a unique moment and largely unprofitable, and even now waning.

To end  – if a law is empirical it is useful, summarising observations.

If it is also predictive, then it is more powerful, allowing us to see the future.

We think Hornigold’s Law is the latter, and so more useful than other energy “laws” such as Vaclav Smil’s “energy transitions take a very long time”.

Whilst that may, or may not, depending on timescale, be empirically correct, it does not really say very much because there have only been a few and that is statistically irrelevant.

Any in any event – does that mean global or regional ? Because as we have shown in the UK the move from coal-dominated power energy to zero-coal energy took only 6 years after a century of dominance. Smil also adds so many of these caveats to his “Law” that it seems to be only observed in the breach.

To conclude – energy is pervasive and fundamental and the in many ways the basis of the universe, so that earth-bound dominance of fossil fuels may ultimately look like an energy side-note in the grand history of energy development on planet earth – and indeed it hopefully should assuming we want to live for another 50 generations, with finite (economic) fossil fuels to rely on for at best just a couple more.

Therein lies the key issue: comforted by the physicality of the existing, accessible, fossil fuel system, which has allowed us to expand as an energy-acquiring species by many orders of magnitude, we assume a) we can carry on in this way. b) we can manage the downsides of carbon emissions and impact, and c) we can continue to exploit the finite resource forever.

Even if we cared not for CO2 emissions the finite resource of fossil fuels will run out.

Then what?

The good news is –  if we are near “peak fossil fuel production” as we likely are, Hornigold’s Law has an inverse: we are likely at peak infrastructure, as the new (lighter) system replaces the old one.

Hereon in we will see the dematerialisation of the energy hardware as incumbent mines and offshore platforms and power plants reduce quicker than the transplants of panels and turbines – but this will be a wrench for specific (politically-supported) communities.

We therefore need to remind everyone that a new, just-as-big in terms of scale and jobs energy system is coming in  – so start to embrace it and recognise it is important not due to CO2, but for long-term generational survival, and short-term economic wealth.

We therefore need to make sure we do not outgrow our energy system, and we will if we stick to fossil fuels: if we stick with them long-term we end up in that Mad Max world feuding over minor stocks of petrol and diesel.

Equally, if we mis-time the energy transition we will have a big mis-step between the existing and the incoming systems leading to price spikes and political angst.

So here is the paradox.

The incumbent fossil fuel industry is now too large for our near-term energy needs, and yet far too small for our long-term energy requirements.

We have too much coal and gas and oil that cannot be burnt safely over the next 100 years, yet if we rely on it we will use it all up in the same time-frame, leaving us without long-term global energy.

Example – environmental issues aside, OPEC exists to regulate the supply of oil and has shut in over 10% of its capacity for the last 5 years to prop up prices – hardly a sign of demand strength or indeed a healthy market structure.

So –  we use too much now and we not only cause generational emissions issues, we also use up a large part of our declining resource and divert capital and talent from alternatives.

This is key – Hornigold’s Law notes that this alternative energy system will require substantial investment – so diversion of funding could be a major problem. This is why the “all of the above” strategy advocated by many new energy supporters could be self-defeating.

If the finite amount of annual capital energy investment is diverted over the next few years into technology dead-ends – either innocently or by incumbent diversion –  such as nuclear fusion or small-scale fission or blue hydrogen or CCS  – then it diverts from the real “billion tonne” global scale technologies of wind and solar plus battery storage.

OK – these scalable technologies likely win out in the long-run. But we don’t have the “long-run” to play with.

Unless we just want to focus on ourselves and not the future generations we will create.

Energy is personal – this is our legacy

As Hornigold notes in his podcast, there is a sort of compact we have with the world: we inherit decisions from previous generations, and pass our choices on to future ones.

If previous generations make giant mistakes or grand breakthroughs  – that then becomes our starting fate.

David Deutsch trenchantly put it when discussing the destruction of progress or optimism in medieval times (just 10 generations ago) in The Beginning of Infinity – noting the ability of incumbents or powerful individuals to hold back progress – “ We should take it personally. For if any of those earlier experiments in optimism had succeeded, our species would be exploring the stars by now, and you and I would be immortal”.

We stick with fossil fuels – we consign the next 5 or 10 generations to a planet grappling for energy resources and in constant feud.

Or, if we embrace the new large-scale technology alternative, we give ourselves a chance to release the yoke of energy scarcity, and become energy immortals.

Overblown? Perhaps. Perhaps not.

Thankfully wind, solar, batteries and storage are achieving major funding, and learning curves mean that even though investment has been flat over the past few years, the actual energy capacity installed has been rising quicky – by about 10% pa.

Take heart.

The installed capacity of wind and solar electricity capacity by 2025 (36 months away) will likely be over 3 TW – that’s about a third of the global total from a standing start of practically nothing n 2010. By 2035 at this rate it will be enough to cover all power consumption globally.

We can create energy immortality via this route.

This is our possible legacy.


We can double down on the fossil fuels beneath us, extract them all, and leave no trace other than more bones and more fossils for no creature to ever see.

The Earth left rotating with a wild climate, undiscovered, giving no history to the universe, like a silent screaming version of Mercury the planet in the image above, an unEarth if you will.

And despite all our dreams and ideas it will be as if we had never been.

That is what is at stake.

————-  ————-