Elon Musk got one thing right about selling electric cars — but it’s not enough to save Tesla from his mistakes (TSLA)

Autopilot, Tesla’s semi-autonomous technology, has always had a bit of a Hail Mary quality. Tesla’s entire narrative was centered in sexy all-electric cars until a couple of years ago when the big story shifted to autonomy.

Musk wisely pushed Tesla in that direction, but in typical fashion, he needed a quickie fix that wouldn’t overwhelm Tesla’s R&D resources. So Tesla went with a camera-based self-driving setup because it couldn’t realistically catch up to the likes of Google (now Waymo), which has developed an expensive laser-radar system and spent years testing it.

The move was smart and it leveraged what Tesla had: a lot of vehicles driving on the road that could create a massive visual data stream. With enough computing power, all that information could be crunched and enable Teslas to drive themselves theoretically anywhere, like a human operator (lidar functions best in well-mapped, geographically defined areas).

But Musk also made a huge mistake. He didn’t realize that self-driving should be a business rather than a feature. Tesla’s cars were the future, so when the future became autonomous, of course they had to drive themselves. Owners would demand it.

But the Waymos of the world were thinking in terms of different demands. Demands for mobility coming from people who didn’t want to own and who thought of autonomy as part of a service.

General Motors figured this out by looking at what Waymo might do and bought a self-driving startup in 2016, Cruise Automation. That division, after a combined new investment from SoftBank and GM of $3.35 billion, is now valued at $11.5 billion — about a fifth of Tesla’s market cap.

Waymo itself could be worth anywhere from $50 billion to $70 billion, according to some analysts.

Autopilot might ultimately be a major selling point for Tesla, but Musk currently doesn’t have any way to meaningfully leverage it as a business, apart from telling Tesla owners that they could loan out their $80,000 car to strangers (the sometimes discussed but as-yet-unfounded “Tesla Network”).

Tesla has always had the luxury electric car market more or less to itself and hasn’t had to worry about catching up. The opposite is the case with autonomy and suddenly more clear business models in the marketplace. Tesla might never catch up on this front, unless it acquires a self-driving service, and the price tags on those are now going through the roof.