Isn't Tesla stock highly undervalued?

I agree with what you're saying. My point is that public stocks rarely can be said to have a startup-like "chance to see those big numbers" (10x+ upside). When such a chance is say 20%+, then you don't need to worry too much about the 80% chance of a 0.5x or 1x downside.

I assume that all the major manufacturers have electric cars for sale by 2025, and they're much more able to ramp up production than Tesla is because of their existing factories.

It's not clear to me that it's easy to retool a factory from making ICE cars to making electric (w/ 2000 vs 20 moving parts in their drivetrains, respectively). Perhaps it's better than starting with nothing, but it's still going to be a huge cost.

Waymo (ie. Google) has a significant lead on Tesla, and would probably sell to all the other car manufacturers. Plus the last few years have shown that autonomous driving is very difficult to get to the truly valuable level.

I agree that Tesla does not have a clear lead over Waymo (and others). My rough impression (having not looked into it in detail recently) is that Waymo and Cruise are able to achieve a higher level of autonomy in more narrowly scoped areas, whereas Tesla achieves lower levels of autonomy, but across its whole fleet worldwide, w/ no geographic limitations.

However, one advantage Tesla has is that it's already booking revenue from selling its full self-driving package. When people choose the 7k full self-driving option [edit: announced today that it'll be 8k starting July 1st], that's cash that hits Tesla's accounts right away. And then they recognize the revenue over time, as they ship more pieces of the package. One might also expect the take rate of the package to increase over time, as "full self-driving" becomes less of a promise and more of an actuality.

And every car they sell (even those w/o the paid full self-driving option) sends data home to train their neural nets.

So, in two ways, Tesla has a continuous path from here to full autonomy: 1) financially they're able to incrementally profit more and more (w/ software-like margins) off of progress towards full autonomy, and 2) they're able to ship incremental features to their fleet, and iterate w/ huge (and by far market-leading) amounts of real-world data.

So there don't have to be any discontinuous leaps in progress for Tesla to get to full autonomy. They just have to keep hill climbing (or descending the gradient, if you prefer).

Their margins are currently riding on not having significant competition, which is changing quickly and the rate of change is increasing. Electric cars are currently a luxury choice, thus having a high margin. But once you are trying to sell to the entire US market, prices are going to have to drop, and margins will drop with them.

I'm skeptical that the competition will be too much of a challenge. It's the classic Innovator's Dilemma, as Liron mentioned. Tesla's batteries have the lowest cost per kWh of any manufacturer, because it's been one of their core competencies and they've focused on it obsessively. Other manufacturers will have to dump huge amounts of money into R&D and reworking their supply chains and factories to become competitive. All while losing money in the meantime because their costs are so much higher.

And other manufacturers (at least the US ones), are not in great financial positions to make that kind of investment right now.

It's hard to imagine other manufacturers truly competing with Tesla in the near term, unless they go all in w/ a bet-the-company approach. Presumably some will, once it's clear that that's their only hope. But I wouldn't be too surprised if many of those end up as the Blackberry or Nokia to Tesla's iPhone.

(One additional Innovator's Dilemma type factor preventing incumbents from switching to the new approach, is that, at least in the US, manufacturers have relationships w/ dealerships that legally require them to only sell through those dealerships. And dealerships have traditionally made most of their money through service. But electric cars need less service, due to having fewer moving parts. So dealers will be less inclined to sell the new EVs and will be a source of friction for manufacturers wanting to switch over from making ICE cars to making EVs.)

While I do think that Tesla can sell more than they currently are, the stock price already has that very factored in. Their market cap is $150B right now, which, assuming an insane $10k profit per car and a normal car manufacturer PE ratio of 10, shows sales of 1.5M cars/year, a 10x ramp up in car sales, again with an insane profit margin.

While I agree that big gains are priced in, do note that: 1) Tesla has consistently grown revenue at 50% per year since 2013 (first full year of Model S sales), and 2) The autonomy features they're already selling are a big boost to margins and will become a bigger boost over time (as they recognize more of the sales as revenue, and as the take rate increases).

So, in the medium term, one might expect Tesla to have margins somewhere in between those of a traditional auto company and those of a software company. (In the long term, their margins will depend on whether they capture a large share of the transportation-as-a-service market, and if so, whether the winners of that market end up with commodity-like margins or monopoly-like margins. My guess is the latter, but I'm not super confident about how that will play out.)

I agree the market is pricing in a normal kind of expectation of “significant gains” but it’s not pricing in a major industry-changing “surprise” breakthrough announcement, which on the meta level should not be a surprise IMO, but rather the expectation. This company has ingredients in place for a massive upside. I admit I don’t know the facts here but I’m feeling pretty skeptical that BMW will have caught up to Tesla’s vehicles in 5 years or even shrunk the size of the lead, even given their currently much higher scale of car production capability.

Re autonomy, I heard on the Third Row Tesla podcast that driving a Tesla in the last couple months has noticeably better autonomy than ever and the fraction of time/situations when you can let the car drive smoothly for you without touching anything keeps increasing. They are also collecting a lot more data than Waymo, the highest quality kind of data you can get: real driving on real roads, and times when the driver disengaged autopilot.

See also my Arbital claim from Dec, 2016, and my FB post from April, 2019.

Moved to answer, as it seems like clearly answering the question.

It is not very rational to have 40% of one's net worth in a single investment. You should use the https://en.wikipedia.org/wiki/Kelly_criterion to size your bets.

They just barely became profitable recently. I don't think P/E is super informative right now.

Like Amazon, they reinvest basically all of what they take in in growth, so don't end up with much in the way of profit. (Historically they've invested more than 100% of what they've taken in in growth, relying on selling equity. Recently it's started to be slightly less.)

So, I think you want to look at revenue, and project revenue growth and future gross margin to get the equivalent of a P/E ratio.

I think they are likely to grow to 100x their current revenue - reasoning backward from market potential the way a startup investor does, rather than anchoring from their current size.

luckily they also seem to be ahead of this trend by developing the driverless taxi fleet capability in their cars. it seems they're actively counting on and accelerating this trend.

cost of ownership will also go down commensurately since owners will be able to send out their cars to be used as autonomous taxis while not needed, thus earning money and offsetting the cost. potentially paying back the cost.

also given their position in grid level storage they have the ability to recycle the battery of the car for grid storage, thus lowering the long term battery cost even more.

this expertise will probably lead to greater functionality soon: i.e. using idle cars charging in their garages as powerwalls / distributed grid level storage, giving owners a piece of the profits.