Monthly Archives: November 2017

You are browsing the site archives by month.

Where are we on the Model 3 ramp?

The biggest question in the EV industry right now is where are we on the Tesla Model 3 ramp?

Prior to the initial handover event at the end of July, Elon Musk tweeted his prediction about the Model 3 ramp, indicating that in the first three months it would grow from 30 to 100 to “above 1500” in September.  He also went on to say that he expected to hit 5,000 per week by the end of the year.  Yes, that is exponential growth!

But like all things Elon says, you do need to take it with a grain of salt.  Clearly this is one optimistic fellow!

So how have things played out with the ramp?

Well July was in fact 30 (actually more were produced for validation purposes, but not sold).  But by the end of September, only a grand total of 260 were produced (220 of which were sold).  Well below the 1630 Elon had predicted.

Well how about October?  Did the expoential growth curve kick in last month?

Well, we can never really know for sure because Tesla does not release monthly sales figures (we will only get a fourth quarter total on or about January 2), but it seems like the answer is no.  InsideEVs predicted only 145 Model 3’s sold.  And the vast army of Tesla watchers that are in Fremont and other locales looking for VIN numbers of cars parked at the Tesla factory and car carriers being loaded (and unloaded) would tend to back that up.  We are not seeing hundreds of vehicles on the roads yet.

Do We Know Anything?

Well yes, there are a few data points we can look at.

Tesla had its third quarter earnings call on November 1, and you better believe the subject of the ramp came up.  Tesla management explained that the reason for the slower than expected ramp was due to bottlenecks in the battery pack production line at the Gigafactory in Nevada, and most specifically in “Zone 1 and Zone 2”, but particularly “Zone 2” of the pack assembly process.  Apparently there are 4 assembly zones that assemble Model 3 battery packs.  It’s a largely automated process and the packs awaiting assembly go through these 4 zones where robots work on them before proceeding to the next zone.  It turns out that zone 2 was not operating correctly, or at least not up to speed.  And since the assembly line is only as fast as its slowest step, it holds up the entire process.  To make matters worse, it’s completely automated so it’s not as simple as diverting more manpower to that step.

Now Tesla told us that they understand the issue.  They have basically had to redesign or at least rewrite the software that runs those zones.  They assure us that they would be able to bring these zones up to speed very soon, and once that happens the rest of the line is ready to ramp up quickly.  On this point I’m still a bit skeptical because I don’t believe the rest of the line has been adequately tested at full speed yet.  Fixing this issue will likely expose a new critical step that they will then have to address.  But things should continue to improve in steps.

As part of the earnings call, Tesla stated two important pieces of information:  they have revised their estimate of when they will reach the 5,000 vehicles per week from the end of 2017 to the end of first quarter 2018; and that they will postpone deciding when to add additional manufacturing capacity to bring their total run rate to 10,000 vehicles per week until they see where they are at on the existing line.  The ever optimistic Musk thinks that the existing line will be able to beat 5,000, perhaps by a significant amount, meaning they may not have to invest as much in expansion.  Let’s put that idea in our next bag of salt for now.

So management indicated what would appear to be a 3 month delay in the ramp.

Tesla reservation holders got a bit different update.  We received an email explaining that there were production issues that were delaying the expected ramp, and that while Tesla understood and was addressing the problem, our expected delivery dates would be updating and we could check our delivery estimators to get an update on our delivery.

Now my estimator previously estimated Nov ’17 – Jan ’18.  This is already an incredibly large window, and my thought was that if anything it would be at the very end of that window anyway.  I am not a Tesla, SpaceX or SolarCity employee, I am not a current Tesla owner, and I’m not on the US west coast, so I don’t get to skip any lines.  I did reserve pre-reveal, however, so that does help.  But a November estimate, even with Elon’s wildly aggressive initial prediction would have indicated I’d probably be in the first 10,000-15,000 and that deliveries would very quickly spread to the east coast.  I always thought January was a safer estimate.

But after the earnings call, my new delivery estimator only slid one month, to Dec ’17 – Feb ’18.  Now in all honesty, this could just as well been a slip from Nov ’17 to Feb ’18, or a four month slide.  But I do actually think that it represents only a single month slide in their estimate, perhaps two.

So was management just being overly cautious with their 3 month delay?  Could it be only a one month delay as indicated by the delivery estimator?

My Take

Well I am going to throw out the idea that it’s only a one month delay.  If it were only one month, we would have seen 1,000-1,500 cars produced in October.  We are simply not seeing evidence of that.

So what about November?  Can we reach any conclusions there?  Certainly until this past week there was not an elevated level of activity.  But that appears to have changed.  First there were reports of car carriers loaded with Model 3’s headed to Los Angeles.  And within the past two days there have been reports from a parking deck in Playa Vista, CA of what appears to be a staging area of Model 3s, first with “several”, then with 20, then 30, and now 50.  Again, I’ll probably take these reports with a small grain of salt simply because the person  reporting 50 stated it was “guess” and “gave up trying to count”, despite the fact that there is incredibly intense interest in the subject, and in particular collecting VIN numbers, take away some of the credibility of the report.

Nonetheless, there have been SOME pictures of VINs, one of which is #1096, and picture of a large number of Model 3s.  Now all this could simply be that they staged a big shipment to Hawthorne for SpaceX employees, explaining the recent lack of activity, and may not be an indication of an increased production rate.  However, it does seem encouraging.  And while there is evidence that VINs are not being delivered close to sequentially, the appearance of a large number of higher numbered VINs does indicate progress.

I feel this development, in addition to sightings in areas outside CA (such as Wisconsin, Atlanta, etc.), in only the middle of the month does seem to indicate the production rate has increased.  Do I feel that November will be the September that Elon envisioned?  Well, I think Tesla will have produced 1,000 cars by the end of the month.  They will probably not deliver them all.  These cars in particular are probably being staged for delivery to a sales/service center where they will be cleaned and prepped for delivery to their eventual owners.  It will probably take them some time to get through this relatively small backlog of deliveries.  So I expect sales to be in the 500-750 range.

I think that puts Tesla just over 2 months behind their ramp.  Big picture wise, hardly significant.  But I also think there will be some additional bumps in the road.  Hopefully Tesla used some of the time whereby these packs were being assembled slowly to really check out the speed of the remaining steps in the line, but at the end of the day there is no substitution for actually running the line at full speed in production.  And I expect they will move from this bottleneck limiting production to 40 cars/week to another one that may limit to 100, and then another that may limit to 400, and so on.  They will get that ramp eventually, but it’s going to be quite as steep as originally predicted.

I do think that eventually management guidance is going to prove correct in the end with a total 3 month delay.  Delivery estimator windows that already had 3 month slack in them and have since pushed out another month could still be met, but I wouldn’t be surprised to see some move out another month.

And while at the end of the month we won’t have official numbers from Tesla to go on, I do feel that we will have some better news than we did at the end of October.  I will also say that I think within another month (mid-December) non-employees will finally start to get their invitations to configure their vehicles (which for the time being will be limited to selecting their wheels and exterior color).

What are your thoughts?

October 2017 US EV Sales Update

So we come into November with the news that President Trump’s new tax plan (at least the version put forth by the House) seeks to eliminate the US Federal EV Tax Credit, possibly rendering this entire discussion moot.  The Senate version, on the other hand, retains the tax credit.  We can only hope that either the bill is reconciled favoring the Senate version, or that the whole proposal gets caught in the usual Washington quagmire and simply goes nowhere before the end of the year.  But I digress…

The October sales figures are in and there are a few high (and low)-lights to note.

The Chevy BoltEV set a second consecutive monthly sales record with 2781 Bolts hitting the road in October.  Even with this result, it doesn’t break GM’s record of 4290 EVs sold in December of 2016, thanks in large part to the 3691 Volts sold that month.  But watch this space!

On the other side of the coin was the fact that Nissan only sold 213 LEAFs, it’s lowest month since February of 2011, in only the third month of its being available in the US when 67 copies were sold.  While one may be tempted to say that this is simply due to the fact that everyone is waiting for the 2018 LEAF 2.0 to appear, it’s probably more true that even if you wanted one you probably couldn’t find one.  The outgoing fleet is being sold off and there were fewer than 300 to be had nationwide at the end of October, and fewer than 200 as of this writing.

The Tesla Model 3 was also a disappointment as well.  Production bottlenecks continue to push back the anticipated ramp of Model 3 production.  While actual figures are not known, InsideEVs estimates that only 145 Model 3s were sold in October (although I believe this is poised to change in November–stay tuned).

Back to the highlights side, the Chrysler Pacifica PHEV sold 1175 units this month, establishing October as a record for FCA of 1485 vehicles.

Outlook

So what do things look like in the final 2 months of the year and heading forward?

Tesla

I’m still estimating 2400 Model S and 1860-1880 Model X for the remainder of the year.  Based on October’s estimate of 1120 and 850 respectively this may seem optimistic, but it will probably come out in the wash since December will likely be a strong month for Tesla (unless they decide to start diverting even more of those vehicles overseas).  In fact, that is the stated reason for the lack of sales–they are apparently stockpiling vehicles to send abroad in December in hopes of finishing the quarter off strong.  For the Model 3 it is starting to appear that the Gigafactory bottleneck that has been the cause of the slow start is finally starting to ease.  Recent VIN spottings and sightings of car carriers loaded with Model 3s indicate that there may be around 1000 Model 3s on the road as of mid-November.  Given the lag in these reports, this would indicate to me that we are probably going to see in the neighborhood of 1000 Model 3 sales in November, which basically puts us 2 months behind Elon’s initial ramp estimate.  The company themselves indicated that it’s perhaps a 3 month delay.  Either way this will certainly seem like an insignificant blip by next year.  I am going to maintain by estimate of 1000 Model 3s in November and 2000 in December, although this could be pessimistic.

Altogether this brings Tesla to 160,222 by year end.  I think it’s very clear at this point that they will not cross 200,000 anytime soon in 1Q18, and therefore I figure they will position themselves to cross that line in April of next year.

However, based on my assumptions for the ramp and expected Model S and X sales in 1Q18 (2000 per week in January, 2500 climbing to 3500 per week in February and 4000 climbing to 5000 per week in March), and a continuation of Model S and X sales, I estimate that Tesla would cross 200,000 just before March 15.  They will therefore have to take some kind of action to cross the line in 2Q18.  With the assumption that they divert an additional 5% Model S & X overseas (above and beyond what they normally do), and either stockpile 1000 Model 3s a week, or basically stop selling for the final 2.5 weeks of March, effectively creating a stockpile of 12,000 Model 3s ready to go on April 1.  With these actions, I estimate they will be at 199,248 on March 31, right where they want to be.

GM

I have revised my BoltEV assumptions to 3000 per month for November and December.  Uncertainty around the fate of the US federal tax credit, and a typically strong December for the car industry and EVs in particular will likely translate to record sales for the Bolt.  With these assumptions, I have GM leading Tesla at year end with 167,486 US sales.  Furthermore, this puts them on a run rate of around 4500 vehicles per month.

Extrapolating that into next year (with a few tweaks), I estimate that GM will exit 2Q18 at about 192,400 vehicles.  If they want to game the system, they too could delay hitting 200,000 until 4Q18, but at this point it’s probably too soon to tell just how far BoltEV sales will go.  If they continue their upward trend, then GM may find themselves beginning the phaseout period next July, only one quarter behind Tesla.  For now I will consider this the most likely scenario.

Nissan

I project Nissan will exit 2017 with 115,690 US sales.  The thought is that LEAF 2.0 will revitalize Nissan’s sales figures, although it remains to be seen how the market will accept the slightly shorter range, but significantly cheaper vehicle.  Even if we assume 5,000 vehicles per month, they will easily cruise into 2019 with their tax credit intact (assuming the politicians don’t get in the way).

Ford

I estimate Ford to be at 104,731 US sales at the end of the year, which is a surprising figure given their sole BEV, the Ford Focus Electric, has only had lukewarm sales, tallying a total of only 8780 sales between its introduction in 2013 and my 2017 year end estimate.  Most of Ford’s sales come in the form of strong PHEV sales.  But still there is a long way to go and it will be well into 2019 before we can start to seriously look at Ford’s 200,000 aspirations.

Others

Despite a lot of talk and on-again off-again plans, the European automakers–BMW, VW, Daimler–sit at 55,203, 26,510 and 13,176 respectively.  Supposedly 2020 will be the year when we really start to see massive numbers of EVs from those makes, but until then, there is no danger of them getting near 200,000.

Toyota has a respectable 67,326 (estimated) and is showing good strength with their plug-in Prius.  But until they start to get really serious about battery electrics (again, supposedly in 2020), we’re still a few years away from seeing them cross the line.

Fiat-Chrysler is actually estimated to be at 29,155 by year-end, thanks to the success of the Fiat 500e and more recently the Chrysler Pacifica Hybrid.  While obviously a long way off, it will be interesting to see how that minivan does in the market.  It may just turn out to be a winner if they market it.

Hyundai/Kia is estimted at 12,903, but with they are poised to come on strong with their Ioniq lineup.

 

September 2017 Sales Update

It’s November 1 today and here I am posting a September sales update!

Well the reason is that I want to get the September sales baseline on the record before updating my numbers to reflect October sales (which are not finalized yet).

The purpose of these updates is to track progress towards, and my estimates of, when various manufacturers will cross the 200,000 US vehicle sold mark and start the clock ticking for the phaseout of the US $7500 federal tax credit for EVs.  The phaseout works like this:

Up until a manfuacturer sells its 200,000th car in the US, the buyer of that manufacturer’s car is eligible to receive a $7500 federal tax credit, subject to them having sufficient tax liability.  Once a manufacturer sells it’s 200,000th car, cars sold in that current calendar quarter and the following quarter are still eligible for the full tax credit.  For the following two quarters, the credit is cut in half to $3750, and for the two quarters following that, the credit is again cut in half to $1875.  After those two quarters are up, the tax credit for that manufacturer expires.  Note that each manufacturer has its own timetable unrelated to other manufacturers.  That means that Tesla may lose its credit while Chrysler may continue to get the full credit for several years!  (Note, I am strongly opposed to this structure as I think it motivates manufacturers to sit back and let others do the hard work and burn through their credits).

The following illustrates how this would work:

US Federal Tax Credit Illustration. Example only, not real data.

Under the system, it would be most advantageous to the consumer if the manufacturer sold car #200,000 on the first day of a new quarter.  This would give the manufacturer practically two full quarters of the full tax credit before it starts to phase out (green shading).  If, on the other hand, car #200,000 was sold on the last day of the quarter, the full tax credit would only be available for another 3 months, vs. 6 months if they had waited one extra day.

So, as the third quarter ends, two manufacturers in particular are getting close to hitting car #200,000:  Tesla and GM.  So all eyes are on these two manufacturers and their US sales figures to determine exact when the phaseout clock starts ticking.  Particularly in the case of Tesla, which looks like it will cross that line first.  There are a few people that say that it will cross the line in December 2017 (unlikely if only for the reason that they will likely wait until the beginning of a new quarter before crossing that mark), and I would say that the consensus opinion is that they will cross the line in 1Q18.  A minority (including myself) think it will be 2Q18, meaning the full tax credit should be available until the end of September 2018, and not phase out completely until September 30, 2019.

Now I get most of my data from InsideEVs Monthly Plug-in Sales Scorecard.  It’s not the authoritative source of information, but the official source of information, the IRS IRC 30D – Plug-In Electric Drive Motor Vehicle Credit Quarterly Sales page, which incidentally is periodically updated, strangely includes only data for Ford, BMW and Mercedes, so we really don’t know about all the other manufacturers.  But InsideEVs tracks these numbers very closely based on actual sales reports from the manufacturers, so they are reasonably trustworthy.  Now I will point out that while InsideEVs and I update our numbers monthly as automakers release their monthly sales figures, Tesla only reports sales quarterly, so for the non-quarter-ending months, it’s a little bit of a guessing game, but InsideEVs does have a good track record and does go back and adjust the numbers to fit the reported data at the end of the quarter.

So without further ado, I present my figures as of September 2017:

US EV Sales summary as of Sep ’17

This shows that through the end of 2016, Tesla had sold 111,949 vehicles in the US.  Through September, they had sold 35,140, bringing the total sold to date (not shown) to 147.089.  I am projecting that by year-end 2017, they will have sold 62,717, bringing their total by year-end 2017 to 174,666, well short of 200,000.  And even though Elon Musk has “promised” us 5,000 Model 3’s per week by the end of December, that still puts us well into 1Q18 before they hit #200,000, which is why I think they will delay selling #200,000 until April 2018.  After all, Elon Musk tweeted that Tesla would do the right thing to maximize the number of people that got the credit, even if it meant taking a hit on quarterly sales.  This means to me that Tesla would divert Model S and X sales outside the US, and stockpile Model 3’s until the beginning of the new quarter.  Not only would this allow them to immediately sell a larger number of Model 3’s at the full credit amount, but it would extend the full tax credit to the timeframe when the dual-motor version of the Model 3 will be available (which will make a LOT of Model 3 reservation holders very happy).  It will also shorten the time period that the Model 3’s nearest competitor (the Chevy BoltEV) would have a tax credit advantage over the Model 3 to one quarter (possibly even zero).  And since those who hold Model 3 reservations are already given a three month delivery window, they could easily shift deliveries up to nearly a quarter without anyone being the wiser.

I will get into more stats and details in coming months, but as one closing comment for this post, let me describe my assumptions for Tesla and GM sales going forward.

For Tesla, I am assuming Model S growth of 15%, or between 2500 and 2600 sales per month (this tracks a 15.83% actual growth rate), and also 15% growth for Model X (although its growth rate has been larger), assuming between 1950 and 2000 sales per month.  For Model 3, it’s a bit trickier.  Elon Musk said 30 in July, 100 in August, 1500 in September, and climbing to 5000 a week by the end of December.  By now we know that they missed the mark by quite a bit in September, and we’re not really sure when they will recover, or even if it’s just a delay in the ramp up, or something they will make back up.  What we do know is that the numbers were more like 75 in August and only 115 in September.  I have revised my estimates for October-December to be 2000 in October, 4000 in November, and 8000 in December, but I feel those are very optimistic best-case scenario type numbers.  But for purposes of this kind of analysis, we do want to think conservatively (in hindsight, I DEFINITELY know those are optimistic)!

For GM I am assuming moderate 6% growth for the Volt (3% actual) and a flat 2500 units per month for the BoltEV.

That’s it for the time being.  I’ll be back shortly with the update for October!