A research note from Barclay’s Bank last week summed up what the boom in electric vehicles, along with gains in fuel efficiency, might mean for oil demand—a reduction, by 2025, almost as large as Iran’s total production. And if electric vehicles seize a third of the car market by 2040, the drop in demand would be nearly as much as Saudi Arabia produces.

That kind of jaw-dropping outlook has become increasingly common in recent months amid signs that a tipping point is coming for electric vehicles.

The technology breakthroughs, market forces and government policies might also augur a peak in oil demand, and that would be a big step toward wiping out emissions of greenhouse gases from the automotive tailpipe.

From Europe to Asia, and in parts of the United States, policymakers are talking about how to make it happen.

France and Britain committed in July to ban the sales of all gasoline- and diesel-powered cars by 2040, motivated largely by health concerns about air pollution. Then China, the world’s largest auto market, announced last month that it will set a deadline for automakers to stop selling internal combustion engine vehicles and set emissions targets for automakers. California officials said they want to follow suit.

Major automakers have been falling over each other in recent weeks to announce plans for electric vehicle fleets and the phase out of internal combustion cars. General Motors said it would launch 20 new all-electric models by 2023, including two within the next 18 months. Its global products chief, Mark Reuss, declared that “General Motors believes in an all-electric future.” It seems the question is when, not whether.

Electric vehicles still face challenges, such as range limitations, battery weight and cost, and sparse charging spots. But the industry has been steadily chipping away at them, and confidence is growing.

The IEA documented a global 60 percent growth in the EV market from 2015 to 2016. While EVs make up just 0.2 percent of passenger vehicles worldwide, in  six countries, their market share has surpassed 1 percent: Norway (29%), the Netherlands (6.4%), Sweden (3.4%), France, the United Kingdom and China (1.5%).

“Norway is leading the way,” said Luke Sussmas, a senior researcher at Carbon Tracker, the British think tank. Norway introduced far-sighted EV incentives in the 1990s, and this year, around 40 percent of all vehicles sold there were electric. California introduced a modest zero-emissions vehicle policy in 1990 and has followed that with other incentives for hybrids and EVs. In the first quarter of 2017, 5 percent of new car sales in California were zero-emissions vehicles.

But policy incentives will only take EVs so far. The real tipping point will come when they are cost-competitive with internal combustion engines.

How Soon Could EVs Be Competitive?

Gregor Macdonald, editor of the Terrajoule industry newsletter, lays out his criteria for a “competitive” electric vehicle: it must go 200 miles without recharging and be priced, without subsidies like tax credits, within $3,000 of a comparable gasoline model. “The buyer won’t even have to consider the additional savings of the lower lifetime running costs” for EVs, he said, noting that they are cheaper both to fuel and to tune up.

The Chevy Bolt and the Nissan Leaf are getting close to his criteria already, and more choices are coming from across the industry.

Volvo has promised that all its models from 2019 on will be fully electric or hybrids. Jaguar Land Rover made a similar promise starting in 2020, and BMW and Volkswagen announced that they were planning more electric models. Ford is creating a new team to accelerate its development of electric vehicles.

Macdonald thinks some EVs will be competitive worldwide by 2020. Other analysts see a longer trajectory. Kevin Book, managing director of ClearView Energy Partners, thinks it will be seven to 15 years.

Earlier this year, Sussmas co-authored a study for Carbon Tracker titled “Expect the Unexpected” that predicted EVs would be cost-competitive with internal combustion engine vehicles by 2020 and have a 35 percent market share by 2035.

Oil industry forecasts tend to be much more conservative. The oil company BP, for instance, forecasts 6 percent of the worldwide fleet will be electric by 2035.

Many oil companies say the peak is decades away. They cite the International Energy Agency’s “current policies” scenario, which finds slowing growth but no peak for oil beyond 2040—but assumes no changes in policy or technology. But some, including Shell, say it could come as soon as 2021, and are changing their business models accordingly.

The IEA itself sees a range of possibilities. A chart of deployment scenarios in its 2017 Global EV Outlook shows some curves rising more sharply than the current pace of technology might suggest—and some meeting or exceeding IEA’s own scenario for what’s needed to achieve the goals of the international Paris climate agreement.

For its part, Carbon Tracker assumes that governments will achieve their Paris commitments and that the costs for low-carbon technologies will continue to fall rather than just stay static. The notion that costs and policies wouldn’t change “is obviously a completely unrealistic proposition,” Sussmas said.

Electric vehicles are just one factor influencing when global oil demand and emissions will decline; others include broader transportation demand, fuel efficiency standards and urbanization.

Carbon Tracker predicts that oil demand will peak as soon as 2020, remain somewhat flat until 2030, then drop off. It bases that analysis in part on its growth projections for EVs, which it says would displace 2 million barrels of oil use by 2035—equivalent to the supply glut that caused oil prices to crash in 2014. By 2050, EVs would take 66 percent of the market, displacing 25 million barrels a day, with “grave consequence for the industry,” Sussmas said.

Samantha Gross of the Brookings Institution said the spate of government deadlines for the banning of traditional cars “seem a little early.”

“They might be better off letting the market do that,” she said.

But Macdonald counters that the ambitious policies are “affirming a trend that’s already in place” and the distant targets set leave plenty of time for them to play out.

“In venture capitalism, startups and the history of new products, there’s an old saying: the first 1 percent is the hardest,” he said. For EVs, in the United States, “we weren’t quite at 1 percent last year. We’ll be at 2 percent next year,” he says. “That’s very dangerous for the auto industry.”

If the automakers’ recent announcements are any indication, they’ve gotten that memo.

In China and India, Air Pollution Drives Change

Just a few years ago, conventional wisdom held that China and India’s billion-plus populations would enter the middle class, buy cars and lead U.S.-style lives, rocketing through the global carbon budget. “We just haven’t seen that yet,” Macdonald said. “It’s not happening.”

Instead, in China, toxic air pollution from coal plants and car emissions are causing human illnesses and crop failures that have sparked public protests. The government responded to the pollution problem with strong subsidies for EVs, Sussmas said.

In India, oil demand grew at 8.3 percent last year, and Macdonald predicts ongoing growth there for another five to seven years. But it is also reaching a crisis point with air pollution, Sussmas said. This summer, India’s power minister, Piyush Goyal, said that by 2030, India is aiming for 100 percent electric vehicles.

“I think most people think that’s unrealistic and almost impossible to deliver,” Sussmas acknowledged. However, such targets “are still significant to the degree they illustrate discussions behind the scenes.”

Oil: Not Just for Cars

Most of the people who are saying peak oil demand is coming are focused on these shifts in private car and truck use, especially EVs and culture change, because personal vehicles use 26 percent of oil worldwide and 45 percent in the United States.

But other uses for oil are not so easily replaced: shipping, aviation, trucking, plastics. And that is part of the reason many oil companies feel confident in ongoing demand for their product.

“It’s not that ExxonMobil is insisting that we all use them. We’re still using this stuff because it works really well,” said Gross, who has worked for the oil industry.  Regardless, she added, “you don’t need to get rid of everything to see global oil demand start peaking and declining.”

What does this all mean in terms of holding the world to less than 2 degrees Celsius of warming, as nations worldwide have agreed—possibly even 1.5 degrees?

The Energy Information Agency (EIA), which is the data arm of the U.S. energy department, recently projected that worldwide emissions of carbon dioxide from the burning of all fossil fuels—oil, coal and gas—would grow 16 percent by 2040 from the levels of 2015, the year that the nations of the world agreed to the landmark Paris Agreement on climate change that is intended to reverse the trend.

The EIA’s current scenario shows a slowing, but no decline, in global petroleum use. Liquid fuels, mostly petroleum based, remain the largest source of energy consumption, and with energy demand growing around the world, so do emissions.

For EVs to make the most of their emissions-reducing promise, the electricity sector will also have to continue its rapid shift to renewable energy sources.

And while that is beginning to happen, even the relatively optimistic Carbon Tracker doesn’t see an end to warming.

The group says that if oil demand peaks in 2020, and we see declines in other fossil fuel emissions as well, the temperature rise would still exceed the Paris targets, reaching between 2.4°C (50 percent probability) and 2.7°C (66 percent probability) by 2100.

However, that’s significantly less than the business-as-usual trajectory toward 4°C and beyond used by many fossil fuel companies.