GM's cost and boldness advantage

I've just done some reading (some 1/2 year old - pretty ancient) on Autonomous Vehicles and it prompted me to reaffirm a few points, I already thought are true, but perhaps have not stated clearly here yet.

Integration

Waymo had and maybe still has a lead in software, but this is far from enough to succeed. My favorite GM lacks some of Waymo's advantages (Google Maps, both for data and for demand generation), but compensates for this and then some by being fully integrated.

This post from GM Cruise CEO explains clearly why being able to retrofit hundreds of cars with self driving capability is so much different than building millions of cars on a proper assembly line. The former is nice, but only with the latter one can build a sustainable business.

What it also shows and I did not appreciate enough earlier, is that GM will have a huge cost advantage in building AVs compared to virtually everybody else. Waymo's deal with Chrysler for 62k cars surely will require similar assembly-plant integration, but fear of IP ownership and haggling on R&D and Capex costs will surely slow them down and make for an inferior process.

I think alleged Larry Page's decision to not front costs for Ford to do deep integration will be one of the biggest lost opportunities out there. I admire him in general, but being slow in bringing Waymo to market will probably make a difference between winning and loosing this huge market. Hell, Google should have bought Ford, GM or even Magna Steyr to enable rapid commercialization.

I don't believe working with auto OEMs at arm's length (Android model) will work for what's ultimately a complex integration between: base auto, self-driving hardware and self-driving software.

Development and deployment choice

Furthermore, by launching first in Arizona, Waymo seems to optimize for safety, rather than for business scale. Another post from Kyle Vogt shows GM's advantage in pace of learning, and it's hard for me to imagine that Cruise will not catch or overtake Waymo in software capability soon, or it hasn't already.

Waymo being first to launch will be a pyrrhic victory, as it will just pave the way for more scaled competition, probably GM. They will join Tesla and Uber in being 'first' (in having 'autopilot' and by having their self driving car causing first death - respectively). But it will not ultimately matter.

GM stock, catalysts and option pricing

It's another story when the market will catch-up with the opportunity and what will be the catalyst of realizing GM's value. I can see a few potential moments for that, with varying degree of uncertainty:
1) Waymo launching commercial service this year
2) Alphabet reporting a jump in 'Other' revenues in some quarter, attributable to Waymo
3) GM launching a beta test with the public, or another major partnership or actual vehicle production number.
4) GM launching actual service to the public sometime in 2019

Based on that, I'm inclined to buy a substantial amount of far out-of-the-money options for GM. The relative weakness and low volatility of the stock makes for a great risk-reward ratio, and while I do have some OOM options already, the time is ripe to buy much more to get leverage with little downside. I want to get the timing right, though, as time decay is simply costly, and GM has fallen after their AV announcement late last year, and can easily fall between latest SoftBank announcement and the time they actually deploy the fleet. This would provide an even better entry point.

I believe that while widely accepted, the Black-Scholes option pricing model can really fall down when dealing with real world, and Warren Buffett noted in his 2008 letter.
In this case, I think judging range of potential future prices of GM stock based purely on it's past volatility and not taking into account optionality (sic!) provided by Cruise is just wrong. It will allow me to profit at the cost of option sellers relying blindly on the model.
Of course one will argue that this optionality is already priced into GM stock. But a rational value analysis disproves that notion - GM is dirt cheap, even if it had no Cruise.

By the way - there's another interesting angle to GM - according to this site, it will likely surpass 200k US deliveries of plug-in vehicles later this year. This kicks-off a phase out period (which might be removed). In a few quarters of the phase out, GM could potentially deliver to Cruise tens or hundreds of thousands of vehicle, while still ripping off (partial) tax credit for them.

62'000, but when?

And just a few minutes after posting about GM, I'm reading that Waymo stroke a deal with FCA to buy 62'000 for the self-driving service.

As Alphabet is my second largest holding, I'm happy to see them move forward, but I worry this might be too little too late if the 62k number will only be deployed over a few years.
The encouraging part is naming San Francisco Bay Area as the next place they'll deploy their fleet.

I think the economics are going to be much better there then in Phoenix.

It's just great to invest in two reasonably priced businesses, which such big optionality for free.

Bittersweet news from GM

The big news for today (as I'm writing this from Newark airport, still on May 31st) is the Softbank investment in GM Cruise.

Given that GM was my largest holding, I'm naturally happy to see it go up, as I did not plan to buy more of it in the near future (at current price).

But I think the implied valuation of $8.2B now and $15.5B at the point of commercialization is well below the true value of Cruise at this point. I estimated that segment to be worth ~$200B as of today, by discounting future cashflows.
It is true that it's still early stage and there are still non-trivial risks (we have discount rate for that), but the chances of any other firm catching on with Cruise are narrow and chances of Waymo getting to scale fast enough are slim.
We're talking about multi-trillion market with relatively high profit margin, first-mover advantages and big network effects and valuing a clear leader to capture it at only $15.5B.

I completely understand why SoftBank did this and would do it on their side any minute.
GM was very shareholder oriented all the time and I think it continues to be. I see two potential motivations on GM side:
1) Frustration on permanently depressed valuation and lack of patience to wait until it catches with intrinsic value (which I'm estimating at close to $200, vs. today's ~$40).
2) Being "blackmailed" into this deal by SoftBank. Maybe they effectively threatened that they will buy other AV startups and subsidize the human-operated ride sharing competitors of GM's future network in the markets GM will deploy. That way, GM wouldn't be able to reach scale on the demand side, until SoftBank-funded AV technologies catch up and become good-enough for commercial deployment through existing ride-sharing networks. Few companies have capability to outspend SoftBank and GM's shareholders might not have a stomach for that. Alphabet of course has the resources, but I doubt they will be able to deploy at the same pace that GM will.


But it's good that GM keeps the option of having their own network open, per Seeking Alpha transcript:

John Murphy

But Dan the concern I guess is like what we’ve seen in the past with other fleet sales and obviously is a very different thing than what we’ve seen in the past with these. But there is some concern that you would be supplying vehicles to a fleet and then be diminished in value in value capture. So I mean is there way that even if you partner with somebody else and selling to a fleet or partnering with a fleet that you can really protect yourself in economics. Do you think even that has been in the past in the value chain?

Dan Ammann

We see this in a very different way. We think the -- bringing this technology to scale with the right level of safety and really operationalizing and then commercializing and it’s been an extraordinarily unique challenge, and we don’t see a number of people being successful with that in the near term. It’s also important as we talked about before to understand that the rate of improvement and rate of iteration that we’ve been talking about, that continues just as much the day after commercial deployment as it does between now and then. This is a product and a service that will continue to improve rapidly over time. That’s why we think rate of development is so important. That’s we think first mover advantage is so important. And so we think this is a completely different dynamic versus some of the other business models that you alluded to.

It is true that selling AV fleets will not be the same as selling dumb car fleets, but I doubt there is much value to be captured beyond the value of the TaaS service itself. This means, that as soon as a few players can deploy cars at scale, a race towards basic economics is almost inevitable and will put a tough ceiling on profits of even the most efficient and successful company.


There's also an interesting theory in the comment:
What I gather from this transcript is:
1. Most Analysts are still crueless.
2. No spin-off is planned.
3, GM will be valued as Sum-of-Parts: OEM and Taas.
4. Masayoshi Son is a great partner for global TaaS market.
5. There’re two segments of TaaS:
1). Moving people.
2). moving goods.
6. On moving goods, GM’partners are:
1). US-Amazon
2). China-Alibaba.
7. On Moving people, GM partners are:
1). US-Uber.
2). China- ?

I agree with first 3 points. I also think that GM will likely partner with others on retail part of 'moving goods' segment. For "Uber Rush" equivalent, I can see them potentially going alone too.

For moving people, which I expect to be the core of TaaS business, I think anything except Cruise deploying their own network would be a mistake.
Given that economics of AV win hands down with human-operated ones, especially in mature markets, I think it won't be hard for GM to quickly take over customers from Uber and Lyft. The very high Customer Lifetime Value of ride-sharing users, can easily justify Customer Acquisition Cost even into the thousands and the product is going to be very sticky.

To sum up, I see this as a validation of my thesis, that reduces risk of my GM stake, but also reduces the potential reward, as I just 'lost' 20% of it to SoftBank in exchange for them being presumably less aggressive in this space and some pocket change.