Ford plans to make self-driving cars for commercial ride-sharing or on-demand taxi services by 2021, a target the automaker says it will reach by expanding its Silicon Valley research lab as well as investing in or buying autonomous vehicle technology startups.
One of those investments—$75 million in Velodyne LiDAR, the leading supplier of technology that lets self-driving cars see and avoid what’s around them—was announced earlier Tuesday. Velodyne raised a total of $150 million. Ford also recently invested in 3D mapping startup Civil Maps, acquired Israel-based computer vision and artificial intelligence company SAIPS for an undisclosed sum, and signed an exclusive licensing agreement with machine vision company Nirenberg Neuroscience. All of these efforts, along with plans to double the size of its research lab in Palo Alto, Calif., are part of Ford’s strategy to develop a fully autonomous vehicle safe enough for commercial use in five years.
The Palo Alto research lab, which opened in January 2015 with just 15 employees, has been a critical piece of the company’s effort to innovate. The lab currently employs more than 130 people. That number will double by the end of next year, Ford CEO Mark Fields told Fortune.
Get Data Sheet, Fortune’s daily newsletter about the business of technology.
Fields and other top executives at Ford (F) want the company to do more than make and sell cars. Fields has mentioned to Fortune—and more than once—that while the company accounts for nearly 6% of the $2.3 trillion auto market involving the sales, financing, and repairing of cars, it hasn’t touched the $5.4 trillion market for car-sharing services like Zipcar and ride-hailing services like Uber that are unrelated to conventional car ownership.
And while Ford has experimented with various ride-sharing and hailing projects, its real interest is in combining on-demand transportation with self-driving cars.
“When you take the driver out of the equation that improves the cost of the ride service by five or tenfold,” Fields said. “And when that happens, people use that service more.” The business opportunity is two-fold, Fields said. Ford could produce and supply the autonomous vehicles and potentially share some of the revenues generated by the ride-sharing and ride-hailing businesses.
Whether it decides to operate its own ride-sharing service or partner with an existing company like Uber is unclear—a strategy that rival General Motors is pursuing in its partnership with Lyft. Fields would only say there are many options available.
For now, Ford is focused on developing a fully autonomous vehicle, or Level 4, a classification by the Society of Automotive Engineers that means the car can handle all aspects of driving in some conditions or driving modes. For example, the cars may only operate under certain environmental conditions or a specific area within a city. It would be one step short of the highest rating, Level 5, when a car can handle all driving tasks in all road and environmental conditions normally controlled by a human.
It’s important to note that these cars will be not be for personal ownership. Raj Nair, Ford’s chief technology officer, noted that self-driving cars will be too expensive, at least initially, to sell directly to consumers.
“Deploying these vehicles through ride-hailing, ride-sharing, or dynamic shuttles gets you the best availability and significantly improves the affordability of those services,” Nair said. “And as we get the volumes up, the economics of scale and the advancements of the technology will probably bring the cost down.”
Nair and Fields predict that personal ownership of self-driving cars will be more likely towards the end of the next decade.
Ford is open to partnerships in Silicon Valley:
While perhaps not as visible as other companies, Ford has been working autonomous vehicles for years. Self-driving cars are part of its so-called Ford Smart Mobility plan, an initiative introduced more than a year ago that involves building cars with additional Internet connectivity, experimenting with different forms of transportation such as car-sharing as well as using big data analytics like grabbing information from in-vehicle sensors to learn more about how people travel.
Fields and Nair wouldn’t say what this fully autonomous vehicle may look like. Fields did provide one hint: “It will have four wheels.” And the company says it will produce a high volume of these vehicles without providing specific numbers.
What’s perhaps more important is what it won’t have. According to the company, the vehicle will not have a steering wheel or gas and brake pedals. That puts them on a direct path to compete (or partner perhaps) with Google, a company that has developed its own self-driving car that has no steering wheel, throttle or brake pedals, but lots of sensors and software.
Right now, that design presents a challenge for companies like Google. Self-driving cars that still look and feel like traditional vehicles face relatively few regulatory and legal barriers. But companies like Google and Ford that envision a self-driving car without a steering wheel or foot pedals face significant restraints, according to a preliminary report released in March by the U.S. Department of Transportation.
Self-driving cars cannot be tested on public roadways without brakes, an accelerator, or a steering wheel that lets humans take over if necessary. Preliminary rules in California—where Google is headquartered and does the majority of its testing and where Ford also has a permit to test its self-driving tech—prohibit the use of fully autonomous driverless cars that lack a steering wheel or a brake pedal.
Fields doesn’t seem to be too worried about those regulatory headwinds, and instead is focused on the impact self-driving cars will have on the industry and society.
“This decade, for the auto industry, will be defined by the automation of the vehicle,” Fields said. “From our standpoint, we see autonomous vehicles as the potential to have the same societal impact that Ford’s moving assembly line did more than 100 years ago.”