General Motors (NYSE:GM) is running its self-driving car program like a start-up, it'll spend about $600 million this year on the effort, and the first self-driving car GM brings to market will be a version of the electric Chevrolet Bolt EV. 

That's all according to GM's chief financial officer, Chuck Stevens. In a presentation to Wall Street analysts on Thursday, Stevens shared some additional information on GM's fast-moving effort to develop a fully self-driving vehicle. 

Barra is standing next to a white Chevy Bolt with self-driving sensors.

GM CEO Mary Barra with a self-driving Chevrolet Bolt EV test vehicle. GM plans to produce "thousands" of self-driving Bolts for a large-scale test soon, according to reports. Image source: General Motors.

Cruise Automation is the center of GM's self-driving effort

Here are the key points from Stevens's presentation that touch on GM's autonomous-vehicle program:

  • Cruise Automation, the San Francisco self-driving start-up that GM acquired early last year, is the center of GM's effort. Cruise is now a GM subsidiary; it hasn't been integrated into GM's big-company structure.
  • Cruise's CEO and co-founder, Kyle Vogt, has "full responsibility for the operational and financial performance of our autonomous vehicle business," Stevens said. 
  • Cruise will be the business center of GM's autonomous-vehicle business long-term. In time, GM may report its earnings results as a separate business segment. 
  • GM expects to spend about $150 million per quarter on its self-driving effort in 2017, or about $600 million this year versus only about $150 million last year. Spending may continue at that pace into 2018. The spending covers engineering and research and development costs. 
  • The spending does not cover the capital costs of developing an all-new vehicle for self-driving, because GM is already producing the vehicle in question: It's the Chevrolet Bolt EV.
  • GM's first self-driving cars will be built for car-sharing and ride-hailing businesses, not for retail buyers. 

Stevens explained that Cruise has been kept a bit separate from GM for a reason. But at the same time, it's still very much part of GM:

The autonomous program is being managed within the company like a Silicon Valley start-up. The business -- Cruise Automation -- has a budget. They have short- and long-term financial targets, and it's monitored at the highest levels of the organization by Mary, Dan and myself on a very frequent basis. 

("Mary" and "Dan" are GM's CEO, Mary Barra, and its president, Dan Ammann.)

Why would GM leave Cruise as a semi-separate organization? The answer probably has a lot to do with maximizing GM/Cruise's ability to recruit top engineering talent. Engineers with the skills and experience needed to develop self-driving software are scarce and in exceptionally high demand right now, particularly in northern California. Salaries have soared, and the competition to hire and retain the best engineers has become fierce.

For GM, the decision to leave Cruise's casual start-up culture intact -- and to leave Cruise as a subsidiary that could offer new employees some of the upside of a stand-alone company -- almost certainly makes it easier for GM/Cruise to recruit engineers who might be tempted by the huge potential upside of an early stage start-up, but wary of the risks. Ford Motor Company recently set up a similar arrangement, taking a stake in Pittsburgh-based self-driving start-up Argo AI. 

The upshot: What we now know

Here are the key takeaways: GM's self-driving effort is centered on Cruise Automation; GM will spend about $600 million on it this year; its first self-driving car will be the Chevy Bolt; and at least at first those self-driving Bolts will be only for car-sharing and ride-hailing businesses (likely GM's Maven subsidiary and Lyft, respectively).

What we don't know for sure is when those self-driving Bolts will start to appear. But we do know that GM and Lyft have both hinted that a very big public test of self-driving Bolts in Lyft service is in the works, and will begin in the near future -- possibly later this year. 

Long story short: GM's self-driving effort is getting close to leaving the lab and becoming a business reality -- and GM is aiming to show that it's not about to be "disrupted" by new technology companies.