Two Billion Machines

May 5, 2023
Two Billion Machines
Two Billion Machines

One of Voyager’s favorite essays of the last five years is Saul Griffith’s “One Billion Machines”. In it, Griffith lays out, clearly and compellingly, how the US can get to net zero emissions by electrifying 1.063 billion devices in the next two decades, ranging from clothes dryers to breaker boxes. It is outstanding analysis and outstanding writing, too.

We thought it worth extending Griffith’s thought experiment horizontally and vertically: outside the US, and to other processes that use serious amounts of energy. There are a few more devices to address, and several new frameworks to consider. Decarbonizing not just transport and home cooking and heating, but aviation, shipping, industrial heat and chemical transformation is not a matter of a mere billion machines. It requires addressing twice that number: two billion machines. 

These machines have a few characteristics in common. Almost all of them feature rotating machinery, whether as a key step to transforming energy into useful work, or for the prime purpose of providing motive power. Almost all of them feature heat as well; they either produce it as the primary product or as a byproduct. Managing heat is either a primary priority, in terms of producing it consistently at specific temperatures and pressures, or a secondary but significant priority, in terms of disposing of waste heat in a safe fashion.  And, they have long asset lives. The average age of a US automobile is now 12 years.  Aircraft lifetimes can be 35 years or more, and power plant lives can be 60 years, or longer. Finally, all these devices represent a significant asset allocation decision. Cars are the biggest consumer purchase besides a home. Airplanes cost millions; power plants and steel plants, billions. 

Now for the numbers. Most of those two billion machines are road transport vehicles. There are just over 1.5 billion cars, trucks, and buses on the world’s roads today.1 There are more than 27,000 commercial planes and more than 100,000 merchant ships; as of 2017, there were more than 100,000 locomotives as well. 

There are also tens of thousands of very large energy and industrial projects that emit carbon dioxide and other greenhouse gases from industrial processes. There are more than 13,600 coal-fired power units in existence today, nearly 9,900 natural gas plants, more than 3,600 cement plants, nearly 1,000 steel plants, and several hundred aluminum smelters. 

We can add in 500,000 coal- and gas-fired industrial boilers, as well as large furnaces, district heating systems, refineries and other large, heat- and energy-intensive systems. There are also the many tens of millions of yard tools, many of which can already be electrified (a major US retailer has almost 700 in inventory). And then, including home and commercial appliances – all those fossil-powered boilers and stoves and water-heaters - we are close to two billion machines that must be decarbonized.

The Auto Exemplar

Automobiles are a useful case study in how changing two billion machines can start. Global passenger car sales peaked in 2017, around the same time that the electric vehicle market was achieving modest global scale. Since then, internal combustion engine sales have fallen – fewer are now sold than in 2010, during the recovery from the global financial crisis – while electric vehicle sales exceeded 10 million in 2022.  The entirety of the car industry’s recovery from Covid and supply chain stress has been electrics; to put it another way, there is no auto industry growth without EVs. 

Chart, bar chartDescription automatically generated

Last year, one in seven new cars sold was either fully electric or a plug-in hybrid (the split is 3:1 in favor of pure electrics). The distribution of that electrification may surprise those outside the auto industry, too. SUVs are electrifying faster than sedans, coupes, and wagons. Emerging and developing economies, largely thanks to China, are electrifying sales faster than developed countries. 

In all geographies, accelerating electric vehicle sales illustrate the power of inflection points: buyers will favor the technologies that are best-performing and best-priced, accelerating investment into innovation and production of EVs that are now lower-cost and lower-maintenance than fossil-fueled vehicles. Decarbonizing two billion machines requires the decarbonized entrant to simply outperform the incumbent technology.

ChartDescription automatically generated

Road transport sectors have very different electrification rates. Nearly half of all bus sales are electric, but less than four percent of commercial vehicles sales, light or heavy, are electric. 

Together, the electric vehicle fleet is averting nearly two million barrels per day of oil demand, a figure that will only increase. In context, this is nearly the same as Germany’s daily oil consumption. Road transport is far from solved, in terms of CO2 emissions, but the solution set becomes more apparent by the year.

Road transport is also a market-tested laboratory for deploying technology and innovation at scale. With so many units shipped per year, and with regular updates to vehicle designs and manufacturing processes, companies regularly improve products along cost vectors, efficiency vectors, or both. 

Electric vehicle battery chemistry is an instructive example for other electrifying sectors that will produce hundreds of thousands or millions of units of annual production. Electric vehicle makers have started to substitute lithium iron phosphate (LFP) batteries for the incumbent nickel manganese cobalt (NMC) lithium-ion chemistry, with LFP claiming nearly a third of the EV market in the third quarter of 2022. LFP batteries are lower-efficiency than NMC, but they also eschew scarce and expensive cobalt, and have much lower fire risk than NMC batteries. CATL, the world’s biggest lithium-ion battery manufacturer, also intends to produce sodium-ion batteries at scale, which would further reduce the global battery industry’s exposure to commodity inputs. Two automakers, JAC and Renault’s Chinese joint venture, have announced plans to produce vehicles using sodium-ion batteries. Tesla continues to innovate as well, announcing at its recent investor day that it will develop motors that use no rare earth elements, obviating another cost and procurement challenge. 

The transition underway in vehicle production illustrates how we will change the nature of two billion machines and change our emissions trajectory in the process. That said, given their regular turnover, vehicles are far easier than the other machines that need reinventing. Addressing these other machines means thinking about how we change fundamental demand for fossil fuels.  

Replacements, Alterations, Inputs, Outputs

How will we tackle the rest of the machines that produce emissions? Voyager categorizes them in three-part framework:

  • Machines or processes that can be replaced 
  • Machines or processes that can be altered
  • Machines or processes that leverage new inputs and produce new outputs 

Replacement is already underway in two major emitting sectors: transport and power generation. Electric vehicles replace internal combustion engines, and renewable power generation replaces power generated using coal or gas. This replacement is both nascent and predictable over the course of the next few decades. To put it another way, replacing the billion-plus machines in road transport and much of power generation means that these sectors, if not yet solved from an emissions perspective, will be largely solvable given time, capital, and rational permitting and regulation. 

“Largely” is the important word here. Baseload power generation, capable of operating without interruption for weeks at a time, is outside the scope of variable generators like wind and photovoltaics, and long-duration energy storage has yet to scale. Decarbonizing the last aspects of power generation will require significantly scaling up baseload technologies such as geothermal and potentially, new modular nuclear tech. Bringing those to market will involve some technological innovation, but will rely more on policy and regulatory support, and power market mechanisms to pay for always-on generation and delivery. 

Alteration has begun as well, though much of it is still nascent. With the proper retrofits, gas turbines can accept a blend of hydrogen in their combustion fuel, as GE and others have demonstrated. That alteration comes at a technical cost, and it comes at a financial cost too. Still, alteration will be constructive in places that are resource-constrained (in particular, places where wind and solar generation will be physically difficult to deploy). It can also be fruitful if the volume of already-deployed capital — physical and financial — is sufficiently large that re-purposing it is a net positive societally. 

Blending a zero-carbon fuel into a carbon-rich fuel mix does not fully decarbonize an incumbent technology. Going beyond whatever technical fuel blending limits exist will mean completely rebuilding technologies for a 100% hydrogen fuel mix, or de-commissioning existing energy assets and using their physical sites and interconnections to host a new zero-carbon technology such as small modular nuclear reactors. 

Inputs and Outputs
New inputs, and new outputs, are more complex still. The ‘hard-to-abate’ sectors, where emissions reductions have proven very difficult to date, are hard to abate for very real reasons. Liquid hydrocarbons have excellent energy density and produce intense, high-quality heat for processes that need both, such as aviation and cement manufacturing. There, we expect to see new fuels that replace what we have today, such as sustainable aviation fuel instead of jet aviation kerosene. We may also see electrified aviation, given the early success of prototype small electric jets on regional and inter-island routes. 

And then there is heat, which between homes, industry, and other applications is the largest end use of primary and accounts for about half of total energy consumption. Electrification of industrial processes can serve the same role as combustion, in terms of providing heat (and will often work in tandem with alteration of processes). Heat pumps, which can supply temperatures of up to 160 degrees C (320 degrees F) have become a practical substitute for residential and commercial heat applications. They are also popular, with sales in Europe up almost 40% in 2022 to about 3 million units. In the US last year, consumers purchased more heat pumps than gas furnaces.

Chart, bar chart, histogramDescription automatically generated

One new output that two billion machines can expect to engage is an output that has been with us since the industrial revolution: carbon dioxide. CO2, managed through point-source capture or direct air capture, will transform in these cases from an atmospheric pollutant to a new industrial machine output. 

It is also, simultaneously, a new input if captured carbon is put to productive use replacing fossil hydrocarbons in petrochemical processes. And if direct air capture does scale to become a global industry, then it too will become a new subset of industrial machines with its own questions – in particular, its own energy consumption. 

Measure, Optimize, Aggregate, and Vault Ahead

Transforming our two billion machines will require more than just swapping electric vehicles for internal combustion engines or dropping sustainable aviation fuels into our fleet of commercial aircraft. Efficient transformation at global scale will require measurement, optimization, and aggregation. 

Industrial process emissions will need to be measured with precision and rigor as a foundation for accurate progress. Continual optimization, at the level of individual plants and fleets, will be essential for eking out increasing gains through process innovation, and for integrating new concepts that are potential step-change improvements. Aggregation of assets, and the data they produce, will allow the creation of fleets, or entire economies, of industrial machines that are decarbonizing the world. 

Ultimately, addressing the two billion machines that move us, heat us, and produce our industrial goods is about more than just replacing the system we have. It is also about building a better system, in many dimensions. Electrifying processes will mean declining primary energy use, without the inefficient conversion of fossil fuel combustion into final energy. It will mean reduced point-source emissions of other particulates. It can mean less reliance on fossil fuel imports, when renewable electricity is coupled to electrified industrial processes. And it will offer inventors and entrepreneurs a planet-sized power, transport, and industrial opportunity surface on which to build.

1 There are also nearly a billion two- and three-wheeled vehicles, but this analysis does not include them. They have shorter lives, different manufacturing processes, different sales channels, and are electrifying quickly already.