Even Elon Musk seems a little skeptical of his promise to get 1 million completely self-driving Tesla “robotaxis” on the road next year.

The South African-born entrepreneur at Tesla’s Autonomy Investor Day on Monday made a number of bold claims about what’s ahead for the electric car company. Musk unveiled a new microchip that he claimed would be “objectively” the best in the world, predicted that fully autonomous Teslas capable of operating without a human driver would be on the road by mid-2020, and said the company would have a fleet of “robotaxis” out by next year as well.

This isn’t the first time Musk has predicted that a fully autonomous car was on the one-year horizon for Tesla — the company made a pretty similar guarantee in October 2016 when Musk said Tesla would have a vehicle that could make a handsfree, self-driving trip from Los Angeles to New York by the end of 2017. He didn’t hit that deadline, and there’s a lot of skepticism that he’ll hit his new one, including from Musk himself.

“Sometimes I am not on time,” he said on Monday, “but I get it done.”

Plenty of observers agreed that 1 million Tesla robotaxis by 2020 is probably not a thing that’s going to happen.

“It’s totally off-base. There’s just no way that that is going to happen the way it was presented. They’re not even close,” Steven Shladover, a retired research engineer at the University of California Berkeley, told me.

“To be polite, I think we saw their vision of the best-case scenario. I think that’s what they rolled out. And their track record would suggest that that best-case scenario doesn’t always pan out the way they hope,” said Lou Whiteman, a contributor at the investment site Motley Fool and a veteran transportation reporter.

“The Tesla Network robotaxi plans seemed half baked, with the company appearing to either not have answers to or not even considered pretty basic question on the pricing, insurance liability, or regulatory and legal requirements,” Cowen analyst Jeffrey Osborne said in a note to clients.

That’s not to say that a fully autonomous vehicle, including a Tesla, may not be a reality eventually. (That would be level 5 autonomy, meaning the car could operate completely driverless anywhere, under any circumstances.) But the scenario is a long way away — Shladover told me he believes decades. Right now, Tesla’s automation system, similar to those from Mercedes, General Motors, and Volvo, is a partial one, or level 2. That basically means the car will control steering and speed on a well-marked highway, but it still requires continuous supervision by a driver.

In other words, it’s highly unlikely you’ll see a bunch of driverless Tesla taxis on the road en masse anytime soon.

Musk says your Tesla can be a taxi when you’re not in it

There are two prongs to Musk’s new plan: one, that Tesla will have fully autonomous vehicles on the road by 2020; and second, that those vehicles will be able to act as driverless taxis. (Tesla isn’t quite there yet on the technology, but Musk says that once the software is finally set for the vehicles, the concept should be ready to go.)

Once this robotaxi system is in place, basically what would happen is that Tesla owners would be able to turn their vehicles into a taxi when they don’t need them. Then strangers, using the Tesla ride-sharing app, would pay to take rides — similar to Uber or Lyft — but in your fancy electric car. Think about it: You drive to work, flip a switch to put your vehicle into taxi service, and then off it goes until you need it back.

Musk says the prospect could be a lucrative one. He predicted robotaxis could earn owners $30,000 a year, even after Tesla takes 25 percent to 35 percent of each rider’s fee.

Not all of the cars will be owned by individual customers. Tesla recently rolled out a new leasing structure for its Model 3 where customers can lease a car — but they won’t have the option to buy it once the lease is up. Tesla said it plans to use those vehicles in its ride-hailing network.

“Even if they were to achieve it, there are questions regarding the revenue and margins, if it would even be material to the overall company,” Garrett Nelson, a senior equity analyst at research investment firm CFRA Research, told me. “You have to look at the financials of Uber and Lyft, neither of which are profitable.”

Musk on Monday said that, given the potential for Tesla’s vehicles to be a moneymaker for customers who loan out their cars as taxis, it would be “financially insane” to buy anything else. “It’ll be like owning a horse in three years. I mean, fine if you want to own a horse, but you should go into it with that expectation,” he said.

If Tesla were able to deliver customers $30,000 a year on a roughly $40,000 vehicle, of course it would be amazing. “If it wasn’t hard to believe, everyone should go out and buy five of these,” Whiteman said. “Forget buying the stock, buy the cars.”

To be sure, growing timelines in automated vehicles isn’t a problem that’s exclusive to Tesla, Greg McGuire, the director of the MCity autonomous vehicle testing lab at the University of Michigan, said. “If you look at what has been announced in the industry to date and the timelines, you can see that there’s a lot of exuberance, and then reality comes in,” he said. “Some companies have announced very aggressive time scales that have lengthened over the last few years and shown that it’s a really hard problem.”

A Tesla spokesperson did not return a request for comment for this story.

The timing of this is interesting

There are multiple hurdles Tesla and Musk have to get over to make this robotaxi dream a reality.

They have to perfect the technology, and both they and their competitors seem to still have some significant hurdles to cross. As Tim Higgins at the Wall Street Journal notes, Waymo, the self-driving car unit owned by Google parent company Alphabet, has been working on autonomous technology for a decade. It has deployed a robot taxi service in Phoenix, but a human operator still has to be in the car. And the safety of autonomous vehicles is still in question — Tesla’s vehicles have been involved in crashes, and so has one operated by Uber.

The PAVE Campaign, an automated vehicle industry group, in a series of tweets after Musk’s comments on Monday, warned that promises of “fully autonomous” and “driverless” vehicles can make people think their cars have more automatic capabilities than they actually do, thus putting drivers at risk.

McGuire also pointed out that the industry hasn’t yet defined what counts as safe. “How many millions of miles should we drive then before we’re comfortable that a machine is at least as safe as a human? Or does it need to be safer? Does it need to be 10 times as safe? What’s our threshold?” he said.

There is also the regulatory question. Musk said on Monday that he doesn’t think his robotaxis would get regulatory approval everywhere. He said he thinks he’ll get it somewhere, but he didn’t go into specifics. While it’s certainly possible that Tesla might win local approval for robotaxis, state and potentially federal regulators might be less game.

The timing of Tesla’s investor autonomy day itself is “curious,” Nelson said, because it lands right before the company’s earnings report, set for Wednesday, in which it is expected to report a loss. “This may be a way to kind of boost interest in the story, given the difficult results that we expect them to report on Wednesday,” he said. “There’s no reason why they couldn’t have held this until after earnings.”

Tesla may have been hoping to distract from its other problems — pressures on its margins, its cash burn, and its governance shake-up, etc. — with the announcement that fully autonomous cars are just a year away. But it appears Musk, who often promises things he ultimately cannot deliver, at least in the timeline he promises, didn’t nail the pitch. Case in point: Tesla’s stock price fell by nearly 4 percent on Monday.

“We’re at the point in this story that people are not believing him anymore,” Nelson said.

Privacy Policy and European users agree to the data transfer policy.