There's a rationale for this. Google can either become an auto manufacturer, or become a supplier to auto manufacturers. Becoming an automaker is very expensive, very complicated, far from Google's expertise, and not highly profitable even if successful. Becoming a supplier of self-driving technology is likely to be profitable, but won't require huge amounts of capital or the risk of competing in a crowded auto industry.

So Google/Alphabet has apparently decided to become an automotive supplier. That means establishing a presence near auto companies. That's not a bad thing. The Google people need to be talking to auto stylists and engineers. Google has good self-driving technology, but their stylistic integration of the sensors with the vehicle is terrible. The sensor people and the stylists need to spend lots of time together arguing and fighting until the sensors are either invisible or attractive.

Then the Google people need to work with the hardcore auto engineers about dealing with the real world of automobiles, where there's ice, snow, slush, potholes, rust, dirt, and inept or nonexistent maintenance. They'll need self-defrosting sensors, washing sprays, automatic realignment at power-on, dirt detection, and all those real-world problems Google hasn't addressed yet. They'll need to drive over the bump road and skid on the skid pan, then make sure everything still works.

That's what needs to happen in Novi. That's the real non-fun engineering that needs to be done.