After several years of testing, California-based Google has lifted the veil off of an experimental self-driving car designed and built in-house.

The yet-unnamed prototype takes the form of a small two-door vehicle shaped like a 21st century bubble car. Google explains it intentionally built a very basic prototype so its team of engineers can learn from it and quickly make changes to the overall design if needed.

Inside, the prototype offers a Spartan interior with two seats, a pair of cup holders and a small storage bin. It does not have a steering wheel, pedals or stalks, and the equipment is largely limited to a small button used to start the car and a large screen that displays the route.

The centerpiece of the self-driving system is a powerful roof-mounted laser used to generate a detailed map of the road ahead. The car analyzes the radar's data in real time and compares it with pre-programmed maps in order to detect obstacles such as other cars, light poles and pedestrians off in the distance. Radars mounted all around the car help it obey traffic laws and bring it to a stop if an obstacle darts in its path at the last minute. Similar radar-based driver aids are already available on select high-end vehicles.

Technical details were not published but Google says its prototype is powered by an all-electric drivetrain. Top speed is electronically limited to 25 mph for safety reasons.

Google will build approximately 100 examples of its first self-driving prototype over the next few months. Designed to participate in a pilot program, the cars will be equipped with a steering wheel and pedals so the driver can take over if something goes wrong.

The tech giant has not announced what the future holds for its driverless car program. Google could eventually market a mass-produced self-driving car, or it could bring its autonomous technology to the market by forming a partnership with a major automaker.

Google's self-driving car isn't a car, it's the future

Where we're going, we don't need steering wheels

Speaking about self-driving cars last September, Elon Musk preached caution. The man who wants to send us all to space and shuttle us between cities at outrageous speeds told the FT that "my opinion is it's a bridge too far to go to fully autonomous cars."

Somewhere deep inside the secret labs at Google X, Sergey Brin must have read that and smiled. And then climbed into his tiny car — the one with a strange smiley face for a front and a noticeably missing steering wheel — and with a single button press instructed his car to drive him wherever billionaires go to cackle at the short-sightedness of other billionaires.

On Tuesday night, onstage at the Code Conference in California, Brin revealed an entirely new take on a self-driving car, one decidedly more ambitious than anything we've seen before. Google's as-yet-unnamed car isn't a modified Lexus. It doesn't just park itself. It's an entirely autonomous vehicle, with no need for steering wheels or gas pedals or human intervention of any kind. You can't drive it even if you want to.

The Google Car is fully electric, big enough for two passengers. It'll only go 25 miles per hour. Your involvement with the car consists of four things: get in, put on your seatbelt, press the Start button, and wait. While you're waiting, maybe check out the large screen in the center console, which shows the temperature and the time remaining in your journey but could easily display just about anything else. Like, say, a Chrome browser for catching up on your Gmail or watching YouTube while you ride.

Google's had trouble finding willing partners for its ambitious automotive project — it's always just hacked sensors and cameras onto existing vehicles, which creates a bunch of visibility and sensory limitations — which may be because what it was asking for was only just shy of insane. Nothing about this car is traditional: it has a front made of compressible foam, a flexible plastic windshield, and a dual-motor system that keeps the car running even if part of its engine fails. It's easy to imagine executives at GM balking at quite literally reinventing the wheel to help Google X with its latest moonshot.

But Google evidently found a willing supplier or two (Brin wouldn't specify) and it made a car that works. One that it will test 100 or so of in California, that it's already seated journalists inside and placed its employees in front of as it puttered around parking lots in Mountain View.

Self-driving cars are coming. That's essentially a given: the technology already mostly works, and nearly all automakers believe autonomous vehicles are both a good and feasible idea. They disagree only on the timing, though "by 2020" has become an increasingly popular refrain. The biggest remaining challenges appear to be regulatory rather than technological, as governments start to answer questions like who's responsible when a self-driving car gets in an accident.

Rules in a few states (including California) now permit self-driving cars so long as there's a driver behind the wheel in case something happens, and other lawmakers around the world are quickly warming to the idea. There are important privacy issues at play here, too, and the extent to which Google might use self-driving cars for its own advertising and data-collection ends remains a big and possibly frightening unknown.

In classic Google fashion, though, Brin talked less about what the Google car could mean for Google and more about how it might change the world. What if we all sold our cars? What if every time we needed a car, we unlocked our smartphones and called for one with a single tap, and as soon as it dropped us off it went off to its next job? We'd need fewer parking lots, reduce our emissions, stop driving drunk, and get in fewer accidents. Those who couldn’t or shouldn’t drive – the blind, the elderly — could still get around. This is the future Brin imagines, one with huge ramifications on everything from the environment to the economy. And the cute little car he's been developing at Google X is the closest thing we've ever seen to making that idea real.

This is only the beginning, of course. Google's not shy in admitting its cars have trouble in rain and snow; they'll work nicely in a consistent and comfortable climate like Mountain View's, but the mountains of Lake Tahoe might prove another story. And they sure as hell can’t drift. And for Google's car to be the future of cars and not of golf carts, the company will need to solve for those and countless other problems around the world.

Eventually it's going to work, though, even if by the time autonomous vehicles hit the mainstream they'll more likely have a Ford or Nissan logo than a Google Doodle. (Brin himself mentioned taking a "partnership approach" for the tech.) Google doesn't have the scale, the infrastructure, or likely the desire to enter the car market in a real way. But Google's hardware moves, from the Chromebook Pixel to Google Fiber to Project Loon, have never been about sales. They're about proving what can be done, about pushing the limits, about making us think bigger and differently about what's possible.

Even if only 100 ever see the road, Google's car will force lawmakers to finally figure out what happens when cars stop helping us drive and starts truly driving us. It will force automakers to think two steps further down the self-driven road than they had before. It will force customers to get used to the idea of not owning a car, and the notion that it's actually more convenient doing things the Uber and Zipcar way. It’ll teach us to think of cars as public transportation, a service provided for us. Even if we're years away from the wide availability of the technology it's now clearer than ever that's what a "self-driving car" really means.

Google’s not just trying to make cars that drive us around. It’s trying to reimagine what a car is for in the first place, to teach even Elon Musk to think a little bigger. And it might be just rich and audacious enough to pull it off.

Google is spinning out its self-driving car unit as a separate company called Waymo, the company announced today. The name is derived from its mission of finding “a new way forward in mobility.”

“We’re now an independent company within the Alphabet umbrella,” Waymo CEO Jon Krafcik told an audience at a press event in California today. Krafcik also noted that they had the first fully driverless ride on public roads in Austin last year, using a car with no steering wheels and no pedals in “everyday traffic” on real city streets.

This first ride put Steve Mahan, a friend of Waymo principal engineer Nathaniel Fairfield who also happens to be legally blind, in the self-driving car solo for its debut fully driverless ride on public roads. Mahan had ridden in Google test vehicles previously, but always accompanied and escorted by police. This time, he rode with neither, and the car negotiated four-way stops, pedestrians, narrow streets and more in public in Austin.

Recent changes to Google’s self-driving business unit include the appointment of ex-Hyundai North America executive Krafcik as the project’s CEO. That, combined with the later addition of ex-Airbnb and TripAdvisor exec Shaun Stewart later this year suggested a move away from pure research and tech among the top ranks at the self-driving unit, and towards experienced business executives with a history of building commercially viable companies.



“We’ve talked a lot about the two million miles we’ve driven on public roads,” Krafcik said at the event. “Now we’ve driven another million miles on public roads. We don’t talk as much about miles we put on in simulation. We’ve done over one billion miles in simuliatio[…] And we have taken over 10,000 trips with Googlers and guests in places like Mountain View, Austin and Phoenix.”

Google X – and now Waymo – has accomplished a lot in all that time spend driving and testing, but Waymo’s head of self-driving tech Dmitri Dolgov says that there’s still a lot left to work on, including building out better maps, making the ride smoother overall, and improving navigation in inclement conditions like heavy, driving rain and snow.

As for where the business is headed under the new Waymo brand, Krafcik spoke of a range of potential opportunities.

The new Waymo logo.

“We can imagine this [technology] in ridesharing, in transportation, trucking, logistics even personal use vehicles and licensing with automakers, public transport and solving the last mile” he said. “Self driving technology is awesome in all these categories.”

Krafcik also emphasized that the new company is focused on technology, not necessarily cars. This fits with reports that the business will be looking to partner more with vehicle makers, rather than building its own cars.

“We are a self driving technology company,” he said pointedly. “We’ve been really clear that we’re not a car co. although there’s sometimes some confusion on that point. We’re not in the business of making better cars. We’re in the business of making better drivers.”



Krafcik said that Waymo is currently in the “build phase” of putting next generation sensor load outs in the Chrysler Pacifica, which it announced earlier this year it would be using in a 100-car pilot project in partnership with Fiat Chrysler. They’re currently readying these vehicles for road tests, he said.

Uber is also working with carmakers in deploying its own self-driving vehicles for its ride-hailing service, including Ford and Volvo. Other automakers, like Volkswagen and GM, have opted to build or acquire their own self-driving tech and on-demand mobility service offerings.

Bloomberg reported Tuesday that in fact, Alphabet’s new independent autonomous car company will be teaming up with Chrysler for a ride-sharing service deployment, which would see semi-self-driving Pacifica vans hit the roads to trey passengers as early as the end of 2017.

As mentioned, Google previously announced a plan to build 100 prototype autonomous vehicles based on the Pacifica platform in partnership with Fiat Chrysler, but this new plan will involve a much broader scope and higher vehicle requirement. Fiat also plans to unveil an all-electric Pacifica at this year’s upcoming CES show in Las Vegas, and that could be a key ingredient in its tie-up with Google, since EV fleets are much more practical option for the future of autonomous on-demand services.

As an independent company under the Alphabet umbrella, Waymo will likely be less insulated from scrutiny regarding its progress and performance as a business, so its next steps in terms of partnership and sales or licensing model will be very interesting to watch.

What really makes a vehicle autonomous is not having to count on human supervision to execute its tasks, regardless of the area in which it can do that. This is something Mahmood Hikmet already explained in a series of videos about autonomous driving technology. That said, we finally have robotaxis available. The California DMV (Department of Motor Vehicles) granted Waymo and Cruise deployment permits for them to offer autonomous services.
14 photos

These permits differ from the ones these companies previously had because they allow them to charge for autonomous trips. The autonomous testing permit they had until this point limited the compensation they could get from the public while validating their technology.

According to the California DMV press release, Cruise now can use “a fleet of light-duty autonomous vehicles” for commercial services on certain parts of San Francisco. They can operate on public roads from 10 PM to 6 AM at a top speed of 30 mph (48 kph) and can work under light rain and light fog.

Regarding Waymo, it can have its fleet of autonomous passenger cars in a few San Francisco and San Mateo parts. Its cars can run at speeds of up to 65 mph (105 kph) with the same limitations of light rain and light fog, but no time limitations as Cruise has. The difference has an interesting explanation.

Patrick McGee asked the California DMV about them. The Financial Times correspondent learned that Cruise would be able to pick up passengers without a safety driver. Waymo will have to operate with a safety driver behind the steering wheel.

Although that makes it seem that Cruise has a more advanced tech, that probably has to do with the speed and hours at which Waymo wants to operate. In other words, they have established different ODDs (Operational Design Domains), and opted for a different sort of authorization, with different goals.

Waymo will probably use the modified Chrysler Pacifica and Jaguar I-PACE for its service. Cruise is in a more delicate position: it uses modified Chevrolet Bolt EVs for the task, which makes us wonder if they have already received the new battery packs. Some Bolt EV units caught fire, and GM realized they were caused by defective batteries produced by LGES (LG Energy Solution).

As we mentioned before, regardless of the area covered by Waymo and Cruise, these are the first robotaxis in California. Waymo has had authorization for tests on public roads since 2014, and Cruise got one in 2015. They respectively obtained driverless testing permits in October 2018 and October 2020.

If the commercial use of these robotaxis goes well, we’ll probably see it expand in a few months. Other companies may also obtain deployment permits while some keep using their customers to test beta software on public roads with no authorization for that – business as usual.