Tesla is on the verge of completely changing the way it does business

Ever since Tesla entered the Elon Musk era, around 2008, the electric automaker has practiced a counterintuitive strategy: overpromise and under deliver.

Musk and the team never live up to the ambitious expectations they establish, and for good reason. There is nothing to be gained from aiming low if you're trying to reinvent a 100-year-old technology like the automobile.

That pattern was already due to be modified if not changed outright, as Tesla moves toward being a mass-market mobility company with the debut of its $35,000 Model 3 in late 2017.

Customers at that price point - and there are already about 375,000 of them, potentially, based on $1,000 advance reservations - want what they need when they need it. For them, a car is less a statement of their embrace of the future and more a basic means of transportation.

Still, Musk could have gotten away with overpromising and underdelivering yet again, as Tesla tries to ramp up production and deliveries to 500,000 vehicles per year by 2018.

A brave new Autopilot

But then, on Thursday, Tesla announced that all its new cars would have the hardware and computing power they require to achieve Level 5 self-driving - the highest level, under which the driver is no longer a driver, but rather a passive passenger being conveyed from point A to point B by a car that does everything.

(Tesla hasn't said anything yet about jettisoning the steering wheel and accelerator pedal, neither of which are necessary with full autonomy - presumably because the lighting 0-60 mph times that its cars can clock will remain a selling point, and some old-school drivers will want to control their need for speed.)

If Tesla can get this massively upgraded version of Autopilot, currently a semi-self-driving technology, to work and work reliably, then we will be at the threshold of a long-anticipated and quite radical disruption in how we get around.

The term "disruption" is tossed about all the time in tech business circles, but a fleet of fully autonomous and full electric vehicles, selling for under $40,000, would be the beginning of the end for the immediately antiquated conventional automobile, running on gas and driven by a person.

Musk's timetable for this great leap forward is aggressive: he says that Tesla will send a driverless vehicle on a coast-to-coast self-driving road trip by 2017. That's well ahead of schedule for the auto industry. Ford, for example, doesn't plan on unleashing a similar technology until 2021, and General Motors intends to explore full autonomy in a ride-sharing collaboration with Lyft in big cities, taking its time with a consumer application.

Tesla's strength

Interestingly, getting self-driving tech into its cars, on the road, and functioning properly is something that Tesla is very good at - Autopilot is currently state-of-the-art for self-driving in a vehicle you can actually buy. New-new technologies are a strength for the automaker.

Unfortunately, building cars - the old-old technology - is something that Tesla struggles with. Even if the company has a blowout fourth quarter, it will only just make the low end of its 80,000-90,000 delivery guidance for the full year.

But again, overpromise and under deliver. The markets don't like this, but the penalty that they inflict on Tesla stock isn't intense. Shares might fall far from, say, $200, knocking billions off Tesla $30-billion market cap. But they won't collapse, and this is a company that IPO'd in 2010 at $17.

And here's the thing: full autonomy for a Tesla Model 3 is going to cost something like $8,000 - that's a hugely expensive option for a mass-market car buyer. It could obviously be an indispensable one, if the technology works as advertised. But on this score, underdelivering while continuing to maintain an elevated, futuristic vision isn't going to cut it. The self-driving rubber needs to precisely meet the road.

Tesla naysayers have pointed out that with electric vehicles, there's nothing all that special about Musk and his company. And they're right. Tesla's packaging is sexy, which has been a departure from the more virtuous school of EV design, and by stressing performance over sustainability, the carmaker has captured hearts and minds. Ludicrous-Mode acceleration at supercar velocities is pure automotive candy, the opposite of EV broccoli.

But four wheels and a couple of electric motors isn't much of an engineering challenge, in the grand scheme of things. General Motors has taken about two years from start to finish to get its Bolt EV, like the Model 3 priced under $40,000 but with superior range, into production.

Doubling down

Fully self-driving Autopilot is another story. Musk could have pulled back on Autopilot development after a fatal crash in May, but he chose instead to stay the course and remain focused on what he argued were Tesla's two main mission points: improve Autopilot and launch the Model 3 on time.

It was a brave move, based on his belief that 35,000 annual auto-related deaths in the US alone - most the result of driver error - is an intolerable statistic.

Electric cars, produced in large numbers, are part of his master plan to accelerate humanity's exit from the fossil-fuels era. Autopilot is intended to save lives immediately.

Signs are good that the radical upgrade in Tesla's self-driving tech will be successful. Musk understands that solving a problem as complicated as making a vehicle fully autonomous demands heavy-duty computing power, so Tesla's new Autopilot system - made up of multiple cameras and sensors - will be managed by a separate computer 40 times more effective that what the carmaker is now using.

Musk also knows that what Tesla is calling "Hardware 2" has to be integrated with Tesla's cars during the manufacturing process. A bolt-on technology, along the lines of what Uber is using on its experimental self-driving fleet in Pittsburgh, won't get the job done - it could work on Models S and X but fail on Model 3.

But the days of overpromising and underdelivering at Tesla are coming to an end. A car that can kinda-sorta drive itself will be epic and damaging compromise. Elon Musk has just made his most out-there pledge to date. Now Tesla needs to execute on it to perfection.