Transition to EVs: set to climb the S-curve in the next decade

A new research brief from Bloomberg New Energy Finance describes how electric cars are destined to climb the s-curve of technology adoption in the next decade. The report is accessible to clients only, but there’s a write-up that draws on the research here.

Technologies follow remarkably similar paths to broad adoption. They putter along until they hit that magic sweet spot of value and convenience — and then they take off. Adoption rapidly increases until the market is close to saturated, when it trails off.

It looks like an S.

This great graph from asset management firm BlackRock shows a bunch of S curves, for different technologies:

s curves(Blackrock)

The exact shape is different for every technology, and recently the curves have beengetting steeper (for a variety of reasons we won’t get into). But you get the idea.

Right now, electric cars are still in the putter-along phase, as costs continue falling toward the sweet spot.

How fast are costs falling? BNEF did a bottom-up analysis of four variables: “regulatory support for EVs; the cost of battery packs; the total cost of ownership of EVs relative to internal combustion engine (ICE) vehicles; and EV consumer technology adoption forecasts.”

Long story short, here’s what they found: Assuming oil prices rise slowly back to $70 a barrel between now and 2040 (more on that later), battery electric vehicles (BEVs) will become cheaper than ICE vehicles, in terms of total cost of ownership, around 2022.

That, BNEF says, is when they’ll start climbing the S curve.

Electric vehicles will spread quickly

Here’s what BNEF projects for 2040:

bev adoption(BNEF)

There’s a lot to take in here, but notice three things.

First, sales of EVs (BEVs + plug-in hybrid electric vehicles, or PHEVs) will grow to about 41 million in 2040. That will put total EV ownership at about a 25 percent of the global fleet.

Second, by 2040, EVs will account for 35 percent of new vehicle sales (and rising).

Third, look what happens to poor plug-in hybrids. They never rise above niche sales and are rather quickly rendered irrelevant by BEVs. I wonder if we’ll even see much development in the PHEV space, with this kind of forecast floating around. Maybe for large commercial vehicles?

BNEF acknowledges that two things could substantially alter this forecast.

One is that oil prices could go even lower than expected. Note that BNEF’s forecast is based on the Energy Information Administration’s “low” reference case for oil prices, which has them rising to around $70 per barrel in 2040.

Theoretically, oil could keep falling, so BNEF modeled an extreme case in which oil falls to $20 a barrel and stays there. Under that scenario, EVs don’t start rising up the S curve until about 2030, and are only 25 percent of new vehicles sales in 2040 and 14 percent of the total global fleet.

tesla dealership(Shutterstock)
Will we all have Teslas someday?

The other is that EV adoption could be accelerated by widespread car-sharing and shared vehicle ownership enabled by self-driving cars. If adoption is faster, prices fall faster and EVs spread faster. BNEF modeled such a scenario and found that EVs would reach almost 50 percent of new vehicle sales in 2040.

So, between 25 and 50 percent of new vehicles sold in 2040 will be electric — that’s the range.

The great mystery of how EVs interact with oil markets

The Bloomberg analysis by Tom Randall (drawing on the BNEF forecast) goes on to get into some pretty speculative territory, saying that EVs could cause an oil crisis by the 2020s.

The logic works like this: From 2011 to 2013, US oil production surged almost 50 percent, thanks to fracking of shale oil. Typically in that situation, OPEC countries would dial back their production, to keep prices high. But OPEC declined to do that. So there was a supply glut, and oil prices plunged.


After a brief recovery in 2015, prices kept going down, eventually under $34 a barrel. (Much more on the baffling story of recent oil prices is in this Brad Plumer post.)

What precipitated all this? About 2 million barrels a day of surplus supply, over and above demand.

So when, according to the BNEF forecast, will EVs begin displacing an equivalent amount of oil?

It depends on how fast EVs grow. Currently they’re growing at 60 percent annually. That might slow, though. So here are three scenarios, showing 60, 45, and 30 percent growth:


Current growth in EVs would displace 2 million barrels a day by 2023, 45 percent growth by would displace that much by 2025, and 30 percent growth (BNEF’s forecast) would hit the target by 2028.

Now, there’s nothing magic about 2 million barrels a day. It’s certainly no automatic trigger for an oil crisis. (Would that oil markets were so mechanistic and predictable!) But it is an amount that has been shown to be highly disruptive in the recent past.

And of course it won’t stop there. As EVs head up the S curve, they’ll displace millions more barrels. This clip, from Bloomberg’s cool video, gives some sense of the scale. The little explosion is at 2 million barrels a day.


Maybe displacing 2 million barrels a day won’t precipitate another oil crisis. But how about 10 million a day? Or 15 million?

There’s a reckoning coming — all that remains to be worked out is the timing.

The details are hazy but the direction is clear

There are a million and one caveats about predictions like this. Above all, they rely on oil-price forecasts that have a dismal record.

The Energy Information Administration scenario this research is based on shows oil steadily recovering to $70 by 2040. But EIA itself stresses “heightened volatility and high uncertainty in the price outlook.” And oil doesn’t steadily do anything. Here are prices since 1980:

oil prices(MacroTrends)

Oil prices lurch and swing. Guess how often oil industry analysts accurately predict those lurches and swings? Hint: rarely.

Given that the price of oil is a key determinant of EVs’ competitiveness, making predictions about the exact trajectory of EV sales is just as futile as trying to predict the future price of oil.

What’s more, the influence doesn’t only run in one direction. When analysts forecast EV sales, they tend to treat oil prices as exogenous, i.e., an external variable. That makes things more computationally tractable, but it’s a bit of a fudge.

The fact is, by displacing demand for oil, EVs will exert downward pressure on oil prices. EV prices will affect oil prices, and vice versa. (If anyone is aware of EV forecasts that treat oil prices as endogenous, let me know.)

Anyway, as with most models, it is difficult to make local, short-term predictions with any great confidence; however, it is still possible to make long-term, directional predictions.

the garage of the future(Shutterstock)
The future, yo.

EV costs are declining rapidly. In the reasonably near future — maybe within a decade, maybe two — EVs are going to become cheaper and better than comparable ICE vehicles. When that happens, adoption is likely to accelerate, driving economies of scale that accelerate it even further.

When that happens, a whole bunch of oil is going to go unused. And long before global oil use declines in large absolute terms, it will decline enough to substantially affect prices.

Remember that a lot of today’s more extreme oil extraction projects — shale, tar sands, deepwater — are only economical with relatively high oil prices. If EVs suppress the price of oil in decades to come, they could spark serious retrenchment in the industry.

The exact timing can’t be predicted, but the transition of EVs from curiosity to destabilizing force is coming. The oil industry would do well to wake up to it.

Nerdy addendum:

Two other results from the BNEF research are worth highlighting.

First, the rise in EVs will also boost electricity demand: “Electricity demand from EVs will rise to 1,900TWh by 2040, from 2.5TWh in 2015, contributing the equivalent of almost 10% of global electricity demand in 2015.” That’s a lot!

electric car charging(Shutterstock)

As I’ve written many times, the increasing penetration of battery storage will help stabilize electricity grids and enable the penetration of much more wind and solar power.

Second, the oft-voiced worry that EV batteries will use up all the earth’s rare minerals is apparently misplaced.

BNEF’s forecast does mean that by 2030, production of lithium and cobalt, in particular, will increase substantially. But even in 2030, EV batteries will require less than 1 percent of known lithium, nickel, manganese, and copper reserves (4 percent of known cobalt reserves).

And after 2030, BNEF says, new battery chemistries are likely to take over, obviating the need for these minerals.