Elon Musk's pivot to robots isn't just shifting Tesla's business model. It is changing its DNA.

The electric-car maker was founded almost 22 years ago on a simple, if not radical, idea: Electric vehicles were possible not with some sort of breakthrough in batteries but rather through integrating proven technologies in a new way.

Today, Tesla is taking an alternate route. It is betting a still-nascent artificial-intelligence technology can remake the company, enabling, perhaps, one million driverless cars by the end of next year.

Musk's gamble that Tesla can be a robot company rather than one dependent on cars driven by people has helped keep the company's market value at a level befitting a tech titan rather than an automaker.

After many false peaks, Musk's deployment of robot cars for the public is supposed to begin this month, possibly on June 22, in Austin, Texas. That said, Musk this past week raised the possibility that this timeline could be further delayed.

Musk has been saying Tesla's autonomous cars were near for roughly a decade. The delays underscore the challenge of offering such technology safely and efficiently. In that time, Waymo, General Motors, Zoox and others have demonstrated vehicles on public roadways without drivers behind the wheels.

The genesis of Tesla Motors, as it was originally called, was in the idea of using a common technology, essentially laptop batteries, to power a vehicle. Until that point, the automotive industry had grown frustrated with efforts to develop EVs as engineers looked for the perfect battery -- something that could overcome cost and range hurdles that plagued their mostly abandoned efforts.

"As justification for quitting the EV business, the auto makers pointed out that battery technology had stagnated -- that the fundamental problems of weight, range, and battery lifespan could not be overcome," an early business plan for Tesla read in 2004.

The business plan noted that the industry's favored technology, heavy lead-acid batteries, provided an EV with a range under 100 miles. It also had to be replaced as soon as within 25,000 miles of use. "The auto makers are right: these limitations render even the best performing EV unappealing," Tesla's founders concluded.

But lithium-ion batteries held great promise. So, Tesla was created around the idea that the fat-finger-size lithium-ion cells made popular in consumer electronics years earlier could be wired together by the thousands to create a battery pack to power a sports car. "It is not only possible but surprisingly advantageous to use many small commodity Li-ion batteries to power an electric car," the business plan said. "This is not simple, but the technology has been developed and works exceptionally well."

It was a pitch that clearly spoke to Musk, who became the company's biggest initial investor and then chairman.

The real breakthrough for Tesla came when it used a combination of savvy software and mechanical engineering to keep the battery pack from becoming a fire hazard. This gave the startup a huge advantage when traditional car companies eventually realized the potential that lithium-ion cells held.

Later, as CEO, Musk would spend years working first to prove that there was demand for electric cars and then racing to scale production. More recently, he has become narrowly focused on the robotics vision for the business -- driverless cars and humanoid robots.

Some might see what Musk is doing today with autonomous vehicles as simply an extension of the original integration play. He is using Tesla's manufacturing muscles to integrate robot technology into cars.

Yet autonomous-vehicle technology is still in its relatively early days. This time around, Tesla isn't just pulling off-the-shelf cameras and wiring up them to work. The company has spent years developing its own AI to be the brains of the car.

Tesla fans point to the success of the company's advanced driver-assistance systems, known as Autopilot and FSD, as proof of how far along Musk's team is toward fully autonomous vehicles. They conveniently leave out that what Tesla is selling, by the company's own admission, isn't, in fact, fully autonomous. The driver remains legally responsible, no matter how much the experience might feel as though the car is driving itself. In other words, FSD is just a glorified cruise control.

What Musk is promising next is something different -- a car that is responsible for making the driving decisions. He argues that Tesla has a leg up on rivals given the fleet of vehicles it has sold, providing his team with real-world data to build out his system.

Still, robot cars, even for industry leader Waymo, are something just a step or two beyond an R&D project. GM gave up on its costly robotaxi ambitions. For Waymo-parent Alphabet, the robotaxi service is nowhere near replicating its Google advertising revenue.

With all of its success, Waymo has just over 1,500 vehicles on the road. Or a little more than the number of EV1 electric cars built by GM more than 25 years ago in California.

That failed car experiment by GM helped inspire Tesla's creation. The EV1, which had an initial range of less than 100 miles, was a great example of an automaker failing to commercialize a vision.

Ahead of the robotaxi deployment, Musk fans have mocked the small number of Waymos compared with Tesla's vehicle-making capacity.

"These are unmodified Tesla cars coming straight from the factory, meaning that every Tesla coming out of our factories is capable of unsupervised self-driving!" Musk tweeted this past week. He was talking about a video showing a Tesla Model Y driving without anyone in the front seat, presumably part of the company's testing in Austin.

Despite his bluster, Musk has cautioned that Tesla's early efforts will look small. Maybe 10 vehicles in the first week before scaling to a thousand in a few months and hundreds of thousands, if not more than one million, by the end of 2026.

"We are being super paranoid about safety," Musk posted this past week, "so the date could shift."

After all, what he is promising isn't easy -- even if many believe it will be.