I don’t think people should count on a blowout quarter and I definitely don’t recommend people betting more than they can lose on short term stock price gains from Q4 deliveries/earnings, however it’s also stupid to ignore that there is at least a possibility of a large increase in production this Q, particularly in light of accumulating evidence. You are never going to make money on the stock market if you discard new evidence just because it doesn't match current consensus - consensus is priced in. You make money when you are right and consensus is wrong.
I’m sticking with a base case deliveries estimate of 110k (hitting minimum FY guidance requires towards 105k), but I think it is possible production could be anywhere from 104k to 120k and deliveries up to 130k given the 15k Model 3 inventory at the end of q3.
We have many data points which alone I would largely ignore, but together are painting a consistent picture: Jerome’s July email, CFTC supplier comments, Q3 “commercial negotiations” COGs reduction, Elon’s Q2 FY guidance, Galliers and Ellinghorst due diligence last year, Panasonic public comments, Carsonight leak, International shipments tracking, US delivery data in Troy’s spreadsheet.
Arguments for a large step up in Model 3 Fremont production:
- In July Jerome told employees they were preparing to raise output at Fremont. "The company is “making preparations” to raise output at its auto plant in Fremont, California, Tesla’s automotive president, Jerome Guillen, said in an email to employees on Tuesday. “While we can’t be too specific in this email, I know you will be delighted with the upcoming developments.”
- In July CFTC announced doubling of Model 3 components starting in August 2019. 'Precision stamping service provider China FineBlanking Technology (CFTC) will increase monthly shipments of components for relays used in Tesla Model 3 from about 20,000 units currently to 40,000 beginning August 2019, according to industry sources.” Presuming this is for both US and GF3, this could be towards 10k per week Model 3 production capacity.
- During Q3 Tesla achieved a very significant reduction in COGs which they partially attributed to "commercial negotiations with suppliers”. A significant reduction in supplier prices is consistent with a significant ramp in orders which delivers the supplier staff and depreciation cost leverage.
- In the Q2 call Elon said he expected Fremont Model 3 production to get to 8.3-8.6k per week by year end (but of course he may have been misspeaking and thinking about global production). Joseph Osha: And just as a follow-on then, could we see you manage to make 8,000, 7,500, 8,000 Model 3s in Fremont by the end of the year you think? Elon: Yes. I mean I feel confident it's -- let's just say that the trend is very clearly towards being able to get to 10,000 vehicles a week of which that would be -- there is rough numbers like 8,300 to 8,600 Model 3s and the balance in S and X. So, there's sort of 1,600 to 1,800 SX. In round numbers 8,500 3s, 1,500 SX per week, but probably a bit more than that.”
- Last year the Galliers and Ellinghorst report said Fremont could get Model 3 production to 8-10k per week with limited capex. “Based off our tour and what we saw, we see no reason why Stamping and General Assembly should not be able to handle [seven to eight thousand cars per week] today, and even potentially 10k units, with very little incremental Capex. We believe the same is also true for the Paint Shop when it comes to reaching 8k units a week, with some incremental capex potentially required to get to 10k units. For Body, our understanding is that incremental capex is required (our impression is in the tens and not hundreds of millions) in order to get to both 8k units and eventually 10k units.
- Panasonic published a graph showing a predicted 20% QoQ GF1 cell production increase in Q4 in a public presentation last week.
- Carsonight reported 20% QoQ GF1 Pack production increase to 8k per week. Cleantechnica reported 7k packs were sent to China in the 12 weeks to mid October before shipments were finished. If we assume this is true and there has been no further pack stockpile build since, then about 10 out of 12 weeks were produced in Q3 and 2 of 12 in Q4 (so 1.2k packs sent in Q4). This puts Q4 packs going to Fremont at 13*8k -1.2k = 104k -1.2k = 102.8k Model 3s produced at Fremont
- International ship loading days up 60% QoQ QTD and assuming we are now finished for the quarter, will finish up 49% QoQ. This suggests International Shipments of 50-55k Model 3s this quarter. This is real data by the way and not "tea leaves", some people are spending a huge amount of time tracking all the shipments.
- US delivery date entries in Troy's tracker are up 40% in the first 64 days of the quarter (up to December 3rd) vs the first 64 days of Q3. I have seen many people assume Tesla has been producing 100% international cars this quarter, but this doesn't look to be true at all. It seems US production has also increased QoQ so far this quarter. There are 149 delivery date entries in the spreadsheet in this period in Q4 vs 114 in the same period of Q3, 40% QoQ growth. This did include a large acceleration in the last 7 days. If we assume Tesla stopped International production after 57 days this Q we can instead take just the first 57 days of Q4. Here there are 107 entries or +4% growth vs 103 in Q3. InsideEVs estimated 44k US deliveries in Q3 (I don't trust their monthly breakdown) and in Troy's tracker 1/3rd of Q3 US delivery entries were in the first 57 days - suggesting c.15k US deliveries in the first 57 days of Q3. Taking 4% growth suggests 15.5k deliveries in this period of Q4. If we assume half of these were inventory and half new production it suggests Tesla potentially produced 58-63k Model 3s in the first 57 days - which extrapolates to 93k-101.5k total Q4 Model 3 production and largely reconciling with the datapoints we have from GF1.
Arguments against a large step up in Model 3 Fremont production:
- We haven’t heard stories of significant production shut down or new equipment installation at Fremont.
- We haven’t heard any specific leaks that Tesla is aiming for a huge step up in delivery targets this quarter. (But these emails are more common when Tesla is fighting to meet targets and may not be necessary if they are already almost sold out for the Q).
- Why did Tesla not ramp up to this production rate before if they had the capacity with limited capex? (Most likely because they needed to wait for Panasonic cell production rate to catch up and for word of mouth marketing plus new region rollouts to get quarterly demand to the next level.)
- Ramping up Model 3 production in the US shortly before ending exports to China and also shortly before potentially cannibalising demand with Model Y launch could lead to excess production for US and European markets next year. However it’s possible this ramp is only temporary and they will redirect components to GF3 and Model Y next year when the new production lines are ready for them.
- Towards 100k Model 3 production in Q4 just feels too good to be true.