One nice thing about the just-revealed brouhaha between the SEC and Tesla (NASDAQ:TSLA) is that the correspondence flying back and forth includes financial data not disclosed before. (Unlike on other occasions, it seems this time Tesla is going to add these disclosures to the regular 10-Q and 10-K forms).
In particular, Tesla has for the first time disclosed its full-year maintenance revenues. It had previously done so only for 2015Q1 and Q2. The company says:
Service and other revenue during the years ended December 31, 2015, 2014, and 2013 were $305.0 million, $191.3 million, and $91.6 million, which includes maintenance service revenue of $111.8 million, $68.7 million and $22.5 million for the periods presented.
(Bold in the original - that's the part that wasn't disclosed before).
Although Tesla gets much (or perhaps most) of its service revenue in yearly increments, when customers pay for the service plan, and many actually buy multi-year plans, that's not how revenue is recognized. Rather, Tesla recognizes the revenue when the service is provided, which is to say as time goes by. So if you pay let's say $2,400 for a four-year plan, the first month that you own the car, Tesla will only recognize revenues of $2,400/48 = $50.
Thus, we can estimate the cost per year by estimating the number of car-months that went by in any particular year. As in other articles, I do this by taking the number of cumulative deliveries at the start of a quarter, and assuming that the cars delivered during that quarter spend on the road on average one month. Now, if deliveries were distributed uniformly throughout the quarter then they'd spend on average 45 days on the road, but we know Tesla usually makes an end-of-the-quarter push for deliveries.
I assume no attrition, i.e. no cars that have been lost to accidents or for other reasons; this will make the per-car cost lower but the Tesla fleet is reasonably young (under two years old on average) so there cannot have been much attrition.
I get the following numbers:
The total number of car-months in 2015 was 926,064, which implies revenues of $120.73 per car per month, or $1,450 per car, per year.
We see that, far from being a fluke, the numbers disclosed for the first half of 2015 were lower than for 2015 as a whole. For the first half, Tesla's disclosed revenues ($20 million and $23 million in Q1 and Q2, respectively) imply $110 per car-month, while for the latter half it's close to $130. I wouldn't place too much significance on this increase as on the other hand the figures are lower than for 2013 and 2014. But it's safe to say revenues per car-year are well, well over $1,000.
Curiously, the period during which Tesla disclosed service revenues is the one with the lowest figures. Even then they seemed to me ridiculously expensive, but I guess you get what you pay for (or not).
Incidentally, the SEC's letter also confirms that there are no other big sources of revenue inside "Services and Other." For example, in 2015, the section had $305 million in revenue. After removing CPO, maintenance and powertrain sales, only $18.5 million are left. Doing the same operation for 2014, it turns out only $7.3 million are left, out of $191.3 million in Services and Other revenue. That's why my estimate of $160 per car-month during 2014 turned out not to be far off - Tesla's new data implies $150.
As for my assertion that Tesla's service business loses money
The last article included some tables and calculations that in retrospect seem a bit superfluous (see the last link). You can simply check its latest quarterly report. I choose that one because it's the most recent, but you will reach the same conclusion for any other quarter of 2015 or 2016.
- For "Services and Other," which includes maintenance revenue, total revenues are $149.7 million while gross costs are $144.6 million.
- Therefore, gross profit of "Services and Other," including Tesla Energy and the CPO business, is only $5.1 million. Tesla has repeatedly stated its Energy business has a positive gross margin, and while I cannot find a similarly unambiguous statement for its CPO business, it would make no sense to keep that part of the company going for two years without gross profit.
- It follows that gross profit of the service business (excluding Energy and CPO) can only be a couple million. Tesla's selling, general and administrative (SG&A) expenses for the quarter, by contrast, totaled $336.8 million.
- Of course, net profit is gross profit minus operating expenses, and service's operating expenses are (by definition) included within those $336.8 million in SG&A costs.
There you have it: If you believe Tesla's service is turning a profit, you're really believing that less than 1% of its SG&A expenses go towards service.
So where is the other 99% going?
Disclosure: I/we have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours.
I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.