Comparing the Amazon AWS Tipping Point With Tesla FSD

A recent Morgan Stanley 66 page report on Tesla makes the case that Tesla AI, Dojo training computer and the FSD (full self driving system) can be as critical to Tesla share price as AWS (amazon web services cloud computing) was to Amazon. AWS has become 70% of the Amazon’s profit (EBIT). This means the Amazon share price is about four times more with AWS than it would have been without. The profit alone is nearly four times higher but the growth is faster with cloud computing than without it.

In 2015, Amazon’s core North American and international eCommerce businesses were a combined 93 percent of Amazon revenue. AWS is growing much faster at 49 percent. Compare that to the North America business, which grew 24 percent in the first quarter year over year, and the international business, which shrank 2 percent. AWS in 2015 was a $5 billion a year business. In Q2 2023, revenue for AWS increased 12% year-on-year in the second quarter to $21.4 billion. This is about sixteen times the quarterly revenue of AWS in 2015.

At the end of 2022, Tesla reported about 285,000 people in North America had adopted either the monthly subscription or purchased FSD. This was about 19% adoption on 1.5 million cars. The adoption rate is lower in the rest of the world. There will be about 700,000-900,000 new sales of Tesla cars in North America in 2023. This should be about 160,000 more sales of FSD. The prices are $10,000 to $15,000 and now back to $12000. The number of sales of autopilot is far higher and the prices are $4000 to 6000. The software deferred revenue is about $2.4 billion. Half of the FSD revenue is recognized and all of the Autopilot revenue is recognized. This would mean about $6-8 billion of autopilot and FSD software sales globally (with about $2.4 billion deferred).

30-40% of the lifetime software (autopilot and FSD) sales are happening this year. This is because of the constantly growing new car sales. There were 3.4 million Tesla cars sold at the end of 2022 globally and there will be 1.8 to 2.0 million Tesla cars sold in 2023. This should mean about $2-3 billion in autopilot and FSD sales in 2023.

– Continued to ramp vehicle production at Giga Texas & Giga Berlin, with both building 3k Model Y/week in mid-December
– Reached 500k+ Solar Panel & Solar Roof installations 🏘⚡️
– Enabled anyone in North America who has purchased FSD to request FSD Beta, reaching 285k cars total

— Tesla (@Tesla) December 29, 2022

The Morgan Stanley report means that most of the large institutional investors are prepared to track the financial performance and growth rate of FSD and autopilot. they will reward or punish the stock based upon quarterly changes with the FSD and autopilot outlook. The 10% jump in Tesla share price in reaction to the Morgan Stanley observations means that significant FSD and autopilot revenue was not priced into the share price.

Morgan Stanley projected a $500 billion increase to Tesla share value within about 12-36 months from autopilot and FSD. If the Tesla autopilot FSD business was being awarded a 50X price earnings multiple, then if Tesla makes $10 billion per year in profit from FSD and autopilot then this would be fully pricing in that business.

If we use the price of $12000 for FSD then using only FSD would take 833k annual sales if the sales were counted completely as profit. If the autopilot sales were double the FSD sales, then FSD would only need to be about 60% of the software revenue. This would be about 500,000 annual sales of FSD.

If Tesla sells 3 million cars globally in 2024 then the global FSD adoption rates would need to be 16.7% to reach the 500,000 annual sales of FSD.

If Tesla has fully revenue recognition by taking FSD out of beta, then this would help analysts to not discount the FSD sales for deferred revenue.

If Tesla is selling 10 million new cars per year in 2027 and the FSD adoption rate increases to 50%, then this would mean 5 million annual FSD sales in 2027. The autopilot share of software revenue would drop to about 20-30%. This would be about $80 billion per year in software (FSD and autopilot revenue). Tesla would be matching the absolute revenue of AWS today but with four times the income.