EV VS GAS: WHY THE ROAD RUNS 10 YEARS BEHIND THE SHOWROOM

2026-10-20 · for founders & curious readers · 12 min read · endothermal systems

Aknazar Arysbek, founder — biomedical engineer (MSc, Tufts University). LinkedIn ↗

A quarter of the world's new cars in 2025 were plug-ins. How soon do electric cars replace gas cars on the road? About 10 years after they do in the showroom. Plug-ins may pass half of new-car sales in the early 2030s. Yet in 2046, about 4 in 10 cars on the world's roads still run on fuel alone.

The starting numbers

We follow one example through the whole note: the world's light vehicles, cars and light trucks, from 2025 to 2046. A plug-in is a car you charge from the grid: a battery-electric car, a plug-in hybrid, or a range extender. A regular hybrid charges its small battery from its own engine. All its energy comes from fuel. In this note it counts as a fuel car.

A lot of 100 cars

One point three billion vehicles is about 1 for every 6 people alive. Shrink the fleet to a parking lot of 100 cars you can count. Each year about 6 or 7 new cars drive in through one gate. About 4 or 5 old ones leave through the other, for the scrapyard. The lot grows by about 2 cars a year.

Now let 1 in 4 of the new cars plug in, as in 2025. That is 6.5 × 0.25 = 1.6 plug-ins a year into a lot of 100. The showroom reads 25%, but the lot gains only 1.6 points a year. At that rate, you wait about 30 years for a lot that is half plug-in.

An all-electric showroom from 2027

Let's push the showroom as far as it goes. From January 2027, every new car sold anywhere plugs in. How long until the road follows?

The simplest model is a conveyor belt. Each new car pushes the oldest car off the end. The world sells about 83 million cars a year into a fleet of about 1,300 million. The belt turns over once every 1,300 ÷ 83 = 15.7 years, and the last fuel car leaves around 2042.

But cars do not leave in order of age. Think of light bulbs. Put 100 bulbs rated for 1,000 hours into service on the same day. A few fail in the first week, most fail near 1,000 hours, and a few burn for 2,000. Cars retire the same way. One is written off in a crash at 3 years; another runs past 30.

So we add the missing part: a survival curve. For each year's sales, it gives the share still on the road at each age. In our curve, half of the cars are gone by age 19. Eight in 10 retire between ages 10 and 28.

Now each year's sales enter the fleet and thin out along the curve. We run the all-electric showroom through it:

In 2035, eight years after you can last buy a new car without a plug, 44 of every 100 cars on the road still have none. In 2046, 9 of every 100 still have none. We built the fleet one sales year at a time. Each sales year is a cohort. The method is a cohort survival model.

Now feed four methods the same all-electric showroom and perfect data. The two rules of thumb miss the survival curve in opposite directions:

The central path

The real showroom climbs instead of jumping. So let's feed the model the sales path we think most likely and read the road off it.

Read the table along its diagonal. The showroom reads 42% in 2030. The road reaches 42% in 2040. The showroom passes 57% around 2035. The road reaches it in 2046. The lag is about 10 years.

Ten years is also the average age of a car on the world's roads in our model. While sales climb at a steady rate, the road share in any year equals the sales share of the year its average car was built. Mature markets run older: the US fleet averages 12.8 years (S&P Global Mobility, 2025), and the EU's cars more than 12 (ACEA).

In 2046 the road holds about 1.7 billion light vehicles: 47% battery-electric, 10% plug-in hybrid, 13% regular hybrid, and 30% petrol or diesel only. The last two add up to 43%, about 720 million vehicles that run on fuel alone.

Two speeds

Our world line is the sum of four regional lines. Each region has its own lag.

You can see it in Europe in 2035. Its showroom passes China's, at 92%, yet its road sits 21 points behind. Europe's showroom climbs late and steeply, from about 32% in 2026 to 92% in 2035. Its cars also live longest.

The US federal purchase credit for electric cars ended in September 2025. Battery-electric sales fell to about 6% of the market, and hybrids reached a record of about 16% by mid-2026 (Cox Automotive). Hybrids count as fuel cars here. The US road is still about 64% fuel in 2046.

The road already built

Of the cars on the world's roads in 2030, about 3 in 4 already exist at the end of 2026. Of the 2035 fleet, about half do. Of the 2046 fleet, about 1 in 9 does.

So a miss in the sales forecast moves the near road by only a few points. The ranges below are 80% ranges: the true value should land inside about 8 times in 10.

In 2030, a 17-point spread in the showroom becomes a 3-point spread on the road.

Sales forecasts do move. In its 2022 outlook, the IEA expected plug-ins to reach 21% of new-car sales in 2030. One year later it said 35% (Sustainability by Numbers). Our central path reads 58% in 2035. That is about 8 points above the roughly 50% that the IEA's exploratory scenarios reach that year (Global EV Outlook 2026).

The rule: for the next 5 years, trust the road forecast over the sales forecast.

Signposts

You can check five readings once a year to see which way the sales path bends. Each one sets the road about 10 years later. For scale, the average battery pack cost $108 per kWh in 2025, and $99 for battery-electric car packs (BloombergNEF).

Limits

• Trucks, buses, and two- and three-wheelers.
• Used cars that cross borders. Exports from Japan, Europe, and the US keep fuel cars running longer in Africa, Central Asia, and parts of Latin America.
• Any difference in life between electric and fuel cars. The model gives them the same survival curve.
• Why sales move. Battery cost, policy, oil prices, and trade set the sales path; the model only turns that path into a road.
• Fuel use. A hybrid burns less fuel than a petrol-only car, and the model counts cars, not liters.
• Lessons from past technology switches, such as adoption curves.

FAQ

Does a faster showroom shorten the lag? No. While sales climb at a steady rate, the lag equals the average age of the cars on the road, whatever the slope. A faster climb lifts both lines; the gap stays near 10 years.

Is a hybrid an electric car? No. A regular hybrid has no plug, and all its energy comes from fuel. Plug-in hybrids count as plug-ins.

Will hydrogen cars change the picture? No. About 16,000 hydrogen vehicles were sold worldwide in 2025, Chinese commercial vehicles included, against more than 20 million plug-in cars (electrive, citing SNE Research). We put fuel-cell cars under 0.5% of new-car sales through 2046.

Has the EU banned new fuel cars from 2035? No. In December 2025 the European Commission proposed a 90% cut in new-car CO2 emissions for 2035 in place of the full 100% cut. Plug-in hybrids would stay on sale. The European Parliament's transport committee postponed its vote on the package in October 2026 (electrive).

Do electric cars last as long as gas cars? Unknown. Too few electric cars have reached 15 years to measure. A shorter electric life would leave the road with a smaller plug-in share.

Fuel cars in the 2040s

The lag reaches your own street. Start with your driveway. A gas car sold new in 2027 has an even chance of still running in 2046. Most gas cars sold this decade are still on the road in 2040.

Now look down the street in 2035. Worldwide, 28 of every 100 cars there plug in, while the showroom sells 58. Where you live sets the count: 54 in China, 33 in Europe, 15 in the rest of the world, and 11 in the US.

The fuel station on the corner keeps its customers. In our central case, about 720 million vehicles run on fuel alone worldwide in 2046. That is more than every car and van on the roads of the US and the EU in 2025, about 575 million.

When the IEA publishes the 2026 sales share in mid-2027, you are reading the road of about 2036. Of the cars on the road in 2046, about 8 in 9 are not built yet. People will buy them over the next 20 years.

Sources: IEA, Global EV Outlook 2026, executive summary · IEA, Global EV Outlook 2026, trends in electric cars · IEA, Global EV Outlook 2025, trends in electric car markets · CnEVPost, China NEV retail, August 2026 (CPCA) · CnEVPost, China NEV retail, September 2026 (CPCA) · ACEA, new car registrations, January to August 2026 · Cox Automotive, Q2 2026 EV sales · S&P Global Mobility, US vehicle age 2025 · ACEA, vehicles on European roads 2026 · ACEA, average age of the EU fleet · Hannah Ritchie, IEA projections against actual EV sales · BloombergNEF, battery pack prices 2025 · electrive, hydrogen vehicle sales 2025 · electrive, EU transport committee postpones the Auto Package vote · UN, World Population Prospects 2024. Fleet model: our own cohort survival model (Weibull survival, shape 3, median life calibrated per region). Related: how to turn an idea into a product · how much a hardware prototype costs.