Monthly Archives: May 2018

You are browsing the site archives by month.

IMO, it is all but assured that Tesla will defer 200K until 3Q

As we are now in the second half the 2Q, the consensus opinion is very quickly becoming that Tesla is attempting to push delivery of car #200K out until 3Q, thus prolonging the full US tax credit until December 31, 2018.  It’s not a unanimous opinion for sure, but every day it’s looking more and more likely.  The most contrarian opinion is that of TMC and Model 3 Owner’s Club poster Troy who thinks it’s likely that Tesla has already crossed 200K (or will any day now):  https://model3ownersclub.com/threads/tesla%E2%80%99s-progress-to-200k-us-deliveries-federal-tax-credit-phaseout-op-updated-5-7-18.5437/page-21#post-104512

In the quoted post he believes it is 90% certain that they will cross (or have already) 200K in May, and 99% certain they will cross by the end of June.

Now on the one hand it’s hard to argue numbers with Troy.  His Model 3 Delivery estimator is renowned for being “spot on” and he tracks every last tweet and tidbit of data he can get about Tesla and crunches the numbers.  He probably has a better handle than most on Tesla’s numbers.  But in my opinion, he sometimes gets lost in the details and can’t see the forest for the trees.  And his delivery estimator accuracy is nothing more than a parlor trick:  what he is able to correctly infer, based on your reservation time, your place in line as compared to all other reservation holders, and when you are at the beginning of the queue for those that haven’t received an invite yet, your turn is likely up next, and lo and behold, once the invite comes in, it turns out his estimate is deadly accurate.  But so is your file download dialog accurate when it says you have 1 second to go.  The estimate it gave you 3 minutes ago wasn’t quite as accurate, and the same is true with the estimator.  For those further out dates, every time there is a new piece of data revealed, his model jumps around significantly, sometimes months at a time.

Unless you are truly in the know at Tesla and have actual production numbers, trying to create a super-precise model using data that is inherently imprecise is going to result in predictions that have a lot of precision, but aren’t terribly accurate (unless you get lucky every now and then).  Therefore I prefer a simpler model that is periodically calibrated to fit the data that comes in, but never try to fit a model that incorporates ALL data that is flooding in.

At any rate, I digress.  For some reason when Troy adds up his numbers for Tesla’s sales to date, he estimates it between 198K-202K as of May 24.  I really don’t know where he is getting that figure.  I suspect he may be confusing production and delivery.  Now I admit that I put a lot (let’s even say 100%) of faith in InsideEVs estimates.  But let’s face it, they have access to better data than others do, and an excellent track record.  And they have gone back to adjust their numbers when needed.  Now when I add up InsideEVs estimates, along with my own estimates for the month of May (and predictions for June), I have Tesla at right around 192K as of today (I’ll explain my numbers shortly).  And actually I recently revised my figures down by 1,900 after re-reading the federal tax credit notice that states that it only applies to vehicles built after December 31, 2009, so the 1,900 Roadsters I had taken into account don’t even count!

So where do I get my figures?  Well as I mentioned, I take InsideEVs at their word through the end of April.  And that brings Tesla to a total of 183,795 at the end of April.

Now looking forward into May and June:

First of all, I assume a fairly typical Model S & X delivery schedule of 1836 & 2754 Model S deliveries and 1326 and 1989 Model X deliveries respectively.  Now, there have been some indications that Tesla would shift some S & X deliveries abroad.  This would be a great strategy to help defer 200K to 3Q.  But for now, let’s go with these numbers.

As for Model 3, I am trying to separate production from delivery.  For purposes of the tax credit, only delivery counts.  So when we read reports of the Model 3 line producing 2,250 vehicles in a week, it’s important to note that those vehicles still need to be shipped, prepped and delivered.  In the case of remote east cost points of delivery, that might be up to 5 weeks later due to the various delays in shipping cars across the country.  So I have built into my model a lag factor that follows the typical Tesla pattern of designating vehicles for east coast points earlier in the quarter, and west coast points later in the quarter (although this is typically done to boost end of quarter deliveries, which is certainly not the goal if they want to defer deliveries, but it doesn’t really change the numbers).

So my delivery numbers for the week ending May 5, for example, really represents what was produced at the factory the week ending Mar 31, and so on down to the deliveries for week ending Jun 30, representing cars built week ending Jun 16 (and delivered close to the factory).

Also keep in mind that we are dealing with two factory shutdowns:  one during the week ending April 21, and one week ending June 2.  I assumed 500 vehicles produced in each of those weeks.  I used the real numbers we had for the last week of March and the first two weeks of April, and the rest I assumed a ramp towards 6000 (probably conservative: Elon only promised 5000) for the last week of June.  Interestingly, Elon recently said in an e-mail that the factory achieved a daily rate that would extrapolate to 3,500 a week, and sure enough, my estimate for the week ending May 26 is 3,500!

So back to what this means for the Tesla total shipments.  If Tesla took no evasive action, my model predicts that they cross 200K on around May 31 (less than a week away), and 225K delivered by the end of the quarter.  So really I’m not too much in disagreement with Troy here.

BUT!  I think Troy’s mistake is that he gets too focused on the numbers and doesn’t consider what else is actually going on and the solid reasons why Tesla would want to push 200K into 3Q.

Consider the following actions and behavior of Tesla recently:

  1. Tesla opened up invites to Canadians in March.  Now it took awhile, but they have finally started receiving their cars.  It’s unclear exactly how many cars will be delivered to Canada (it is thought that many will wait for AWD), but from news reports of sightings of a thousand or more vehicles in Toronto, it does appear to be a fairly sizable number (Tesla in particular prioritized Ontario due to the pending removal of a significant tax credit).  Deliveries to Canada will provide a way for Tesla to realize sales without ticking up the counter in the US.  So I have built into my model a total deferral of 10K vehicles to Canada between this week and the end of the quarter.
  2. Next, and despite back-to-back batches of record invites on 4/13 and 4/18, there has not been a new batch of invites sent out by Tesla for over 5 weeks (since 4/25-26).  This is highly unusual and indicates to me anyway that they are flushing the delivery pipeline.  And any new invites received from here on out would not be delivered until July anyway (although there may not be a huge invite batch sent out even now due to the fact that they are opening up orders for AWD and P models).
  3. Most telling of all is the infamous “delivery delay” e-mails that were sent out.  Tellingly enough, orders placed on 4/18 or earlier stated their deliveries were delayed until June, and orders placed 4/18 or later were delayed until July.  The interesting split somewhere during the day of 4/18 tells me that Tesla believes that the 200K’th car was ordered sometime during that day and they were intentionally slipping out car #200K or later into July.  Now, some “July” orders have been getting VINs assigned and delivery dates scheduled, some as soon as May, so Tesla may be sharpening their pencils somewhat and/or finding ways to defer more vehicles abroad to give even the “July” folks a 2Q delivery, but I think that specific behavior indicates to me the deferral intent.

As I mentioned, diverting Model 3 deliveries to Canada (10K total) shifts my 200K crossover to June 8, and puts us at 212,527 by the end of the quarter.  So more action is needed.

Let’s assume that just an additional 20% of S&X deliveries are diverted abroad.  Well, at this point, S&X deliveries are only a fraction of Model 3 deliveries, so this doesn’t have a tremendous effect.  But it does help.  It brings us down to 211,609, and the crossover point to about June 10.

What this means is that Tesla would have to stockpile about 12,000 Model 3’s (give them a small safety margin) between now and then.  It doesn’t mean that they have to completely shut down deliveries on June 10, they just have to slow down deliveries to a relative “crawl”, which really is not even so much of a crawl…it just means they can deliver at the same rate they do today (around 2000-2500/week if my estimates are correct) with the surplus going to “stockpile”.

By stockpiling those 12,000 vehicles, they can defer crossing 200K until 3Q, deferring the revenue until then as well, which will certainly help their 3Q numbers, almost certainly pushing them to a profitable 3Q (writing off the resultant poor 2Q numbers).

Now 12,000 vehicles sounds like it would need a huge parking lot.  Well yes, if they were all stored in one location.  But why would they do that?  There would be no reason not to pre-ship these cars to be close to their eventual delivery centers.  Because Tesla has not sent out any invites since the end of April, they pretty much know exactly where these vehicles are heading (and perhaps some of the “July delay” folks happen to be getting their deliveries at under-represented delivery centers), so they could be distributing these vehicles to their target centers for delivery ASAP at the beginning of July.  There are around 100 delivery centers across the US, meaning on average, each delivery center would have to accommodate 120 vehicles.  Well that is still a pretty big number for some smaller delivery centers, but that’s only an average.  The Charlotte, NC delivery center, located in an old auto dealership, could likely hold 300 vehicles if they needed to.  And if need be, smaller delivery centers could find local storage options for the month of June and early July.

I really think it is all but assured that Tesla is making a play to defer 200K into July, which would have the tremendous benefit of maximizing the number of buyers eligible for the full credit, not only by getting an extra 12,000 deliveries early in July, but they can take advantage of the full ramp to 5,000 vehicles per week (or more) for the entire two quarters the credit is still at 100%.