Some Robotaxi Calculations Based On Elon's $100k FSD Tweet

I'm working on another big post at the moment, that's about 20-25% done. It'll probably take another 30-40 hours of writing to finish, but I haven't been super motivated to write as of late, so it might take some time before I get around to finishing it.

For now, this blog is a short TMC post I made in response to Elon's tweet from a few hours ago, which I thought was interesting enough to cross post here.

https://twitter.com/elonmusk/status/1262526115266035715

The Value of FSD

Currently an Uber ride is anywhere between $2 and $3 per mile. Elon talked at Autonomy Day about offering rides at $1 per mile, significantly undercutting the competition.

Let's assume the ratio of useful miles vs non-useful miles is 1:1, although some articles claim it is significantly higher.

That means a robotaxi will earn $0.50 per mile driven.

Elon talked about a cost of less than $0.18 per mile to operate at that point in time, so a robotaxi would generate $0.32 profit per mile.

Elon also talked about an annual mileage of 90,000 miles, but he based that on 16 mph x 16 hrs per day. The average Uber trip travels at over 25 mph, so I reckon this is a too conservative.

However, what really matters is the longevity of the vehicle, which we all know will soon be 1,000,000 miles. Whether the robotaxi travels these 1,000,000 miles in 9, 10, or 11 years doesn't matter that much, it will earn about 1,000,000 * $0.32 = $320,000 over its lifetime.

That'd be $180,000 in profit on a $140,000 investment, which would be a 12.9% yearly ROI and a 8.6% compounded annual return over a 10 year period.

Even if you make more conservative assumptions:

• Charge $0.80 per mile
• Cost $0.20 per mile
• Profit $0.20 per mile

A robotaxi could still earn $200,000 in approximately 10 years, which would be $60,000 in profit on a $140,000 (Vehicle + FSD) investment, a yearly ROI of 4.3%. If you drop the price of the FSD package to $80,000, that'd be a yearly ROI of 6.7%


The Value of TSLA

Now this is where I'm struggling, because if Tesla is really going to be able to sell the FSD package for $100,000, or even $50,000, TSLA is truly going to the moon beyond a point of what seems possible.

A $100,000 FSD software package is going to be near 100% margins. Maybe not 100%, but whether it's ~90% or ~95% or ~98% is not going to make a big difference in the following calculations. Keep in mind that it's not unusual for some software companies to have near 100% margins. Some have margins as high as 95%, and Adobe has 88% margins on their software.

Tesla's goal is to continue growing at >50% annually through 2030, which would be ~30M vehicles produced in 2030, but let's assume they manage to produce ~20M vehicles in 2030, give or take a few million.

It seems obvious that, just like standard autopilot, every single car is going to be sold with the FSD package. A lot of regular consumers might not be able to afford a $140,000 vehicle, but plenty of investors will buy up all the vehicles Tesla can produce, if they can make anywhere from 4.3% to 12.9% per year on this 'investment'. Tesla will also be happy to invest any of its own spare cash into this great investment.

As a result, Tesla would end up selling 20,000,000 * $100,000 = $2,000,000,000,000 ($2T) worth of FSD packages each year, at near 100% margins.

Although one would have to subtract taxes first, at a 10x earnings multiple, that'd mean a $20T valuation for JUST Tesla's FSD software business, excluding any vehicle sales, energy, and Tesla Network income.


What?!?

Yes... what indeed.

Once again, even if you want to be more conservative and assume:

• $50,000 FSD package price
• 90% gross margin
• 10M vehicles sold in 2030
• 30% taxes before 10x earnings

That'd lead to $400,000,000,000 ($400B) in profit, $280,000,000,000 ($280B) in earnings, and a $2.8T valuation JUST from FSD software sales alone.

Especially in the first $20T valuation calculation, something I struggle with is where the $2T is going to come from, because it's a lot of money. And of course there'd be a ton of money flowing through Tesla's AMaaS (Tesla Network) business, so IF it really gets to that point, I wonder if it's going to influence the world economy. Another poster on TMC mentioned a while back about how Tesla might use their own currency within their Tesla Network, kind of like WeChat does.

People who've done the math on FSD & Robotaxis before should've already realised this, but if Tesla develops software that is proven to drive safer than a human, it will be truly massive.