I pick up a piece I published on LinkedIn some time ago. I argue that the argument that Tesla should be valued like a software company as softwarization of cars is the future lacks a sound framework.

Why valuing Tesla as a software company is not sound

First one needs to understand the part the software plays in the product bundle. Software is part of the new product, the various Model X/3/ cars, for sure. But being part of the product, it is not the game winner. In the end it is a car transporting from A to B. An equivalent in the software space is e.g. Zoom. They master videoconference technology. This is difficult, requires deep engineering knowledge and has hard problems to solve. But others a just a click away, so limited network effect if at all.
We at Iceventure within our technology impact valuation and technology scouting service call this a secondary order of digitalization effect.
What is the definition of it: New offers use current technology, but it is rather state-of-the-art process than USP. This is often mixed up in valuation arguments. This is as technology propositions and valuation propositions come as bundles which needs unbundling when analyzed. 

What criteria of a software company do not apply? 

Back to Tesla and software valuation: Unbundling the software argument, one needs to apply the full criteria of stellar SaaS (software) valuations to Tesla. These criteria are:

Cars do not have network effects. It could, maybe with AI and autonomous driving, but this is another topic, another technology battle of the future. It needs another analysis.

It is not sticky, not like software. An important fact in the SaaS space is to get mission critical information in the system. This is a barrier to exit. It is soft, there are APIs or export, but there are switching costs. But it works rather well. Sure, car brands are able to create loyalty, but there are limits to this.

The economics of software do not apply. You buy a car once, so it would be on-premise, not SaaS. This makes a difference in valuation. But even if it would be SaaS, it does not match.  Yes we can model CAC and LTV for tesla, even churn. But what is about all the economics of scaling up? There is another important argument, I leave out as it is reserved to our customers related to the last point. Will you know what it is?