The company’s new Palo Alto, California, office will be called Level 5. That’s the stage of autonomy where the human race can die off, and the cars drive themselves. It’s a big reversal for Lyft, which had smartly struck partnerships with Alphabet Inc.’s Waymo, General Motors Co., Nutonomy and Jaguar Land Rover.

In an industry that loves to lose money, Lyft had a cheaper option. A week or so ago, an Uber investor even expressed envy over Lyft’s less costly partnership-based strategy. Uber Technologies Inc. is sinking hundreds of millions, if not billions, of dollars into its autonomous vehicle work. Now Lyft is veering more toward the Uber route. Lyft said its focus remains on partnerships.

If Lyft wants to continue to hew to its partnership strategy while delving into self-driving car research of its own, it could potentially take money from, say, less-tech-savvy automakers and let them in on the research effort, helping to mitigate the costs. And even though Lyft only operates in the U.S., autonomous cars would have value globally, giving the company a potential licensing model down the road.

[Related: What are the 5 Levels of Autonomous Driving?]

There’s a lot we don’t know: Does Lyft have the right talent to build a viable offering? How much is Lyft investing? How easy is it to catch up to companies that have been at it for years?

One thing is clear: Building its own autonomous vehicle technology gives Lyft more leverage in negotiations with other companies. Less apparent is whether it will be worth the money Lyft spends. And even less obvious than that is when self-driving cars will actually be cheaper than Lyft’s human-based driver model. (That model has the added benefit of employing human beings, rather than cutting big checks to other companies to borrow their robot cars.)

It’s easy to see why going without a self-driving car program made Lyft nervous. The nightmare situation for ride-hailing companies is that a single player has a sizable lead over everyone else. (Silicon Valley conventional wisdom would put Waymo as the most likely candidate.) That company could presumably find money to build its own ride-hailing network or take a big share of the profits in any deal with a ride-hailing firm.

On the other hand, the optimal situation for Lyft or most other ride-hailing companies would be that autonomous technology is largely commoditized—no more differentiated than buying a Prius or a Civic. That would give software networks all the leverage. Lyft, Uber, Grab, Didi, Ola, Yandex, etc. (some now geographic monopolies) would each be able to force Nutonomy, Waymo, GM’s Cruise, BMW, etc. to compete.

Lyft’s strategy seems to be to nudge along that commoditization—assemble off-the-shelf technology into a complete package that it could sell to or use as a bargaining chip with automakers. You have to believe that based on what Lyft has seen working with automakers so far, it must be operating under either (or both) of these assumptions:

1) Lyft can catch up to at least some of the autonomous vehicle projects out there, or 2) Lyft is more capable than automakers in putting together a viable offering of buyable parts. 

Lyft doesn’t need to be the first, or best, autonomous project out there—right now, it’s neither, and it’s hard to believe it expects to be. Lyft just needs to ensure that it has leverage and that autonomous vehicle companies, automakers and everyone else is as close to an interchangeable commodity as possible.

Sign up to receive the Fully Charged newsletter in your inbox, and follow Bloomberg Technology on Twitter and Facebook for more.

And here’s what you need to know in global technology newsElon Musk's tweet was preposterous. Musk claims he got “verbal” government approval for an underground hyperloop from New York to Washington, D.C. As Bloomberg's story points out, “It’s not clear what Musk is doing with these announcements on Twitter. Such an ambitious project would require billions of dollars in funding and extensive approvals from federal, state and local authorities.”

Masa has so much money to spend. With $93 billion committed to the Vision Fund, SoftBank's Masayoshi Son tries to explain what the plan is exactly. I'm still not sure.

Microsoft's Azure cloud services revenues grew 97 percent in a quarter. Satya Nadella is crushing it. Microsoft's market capitalization is $573 billion, up from $288 billion in fiscal year 2013, back when Steve Ballmer was still in charge.