Near-future quarterly financial projections

[EVNow]

It seems clear to me they are shipping a lot more cars - and haven't been selling that much in US. They will switch in December. Basically rewinding of the wave.

As to what this means for # of deliveries, we don't know. I think it is a fairly safe assumption that they will sell all that they can produce (and more from inventory). Whether that is 105k, 110k or 115k - we don't know. I don't think any of the clues has the accuracy needed.

Infact we face a new challenge this quarter. No InsideEVs estimates, either.

 

[Fact Checking]

No, I don't believe they had 7000 stockpiled battery packs and drive units on 9/30. Doesn't make sense to produce them so far in advance.

It makes sense, if:

• if they are produced opportunistically, like Carsonight suggested, when GF1 was running faster than Fremont could take the battery packs.
• Furthermore if they knew in advance that they won't have enough spare capacity at GF1 in Q4, because by that time the supply chain would be ramped up at Fremont.

We do know it from shipment manifests that Tesla is sending battery packs from GF1 to Shanghai, ~19 per container, per @kengchang.

What we don't know is the volume: just a handful for trial production, or the rumored 7,000?

 

[EVNow]

It makes sense, if:

• if they are produced opportunistically, like Carsonight suggested, when GF1 was running faster than Fremont could take the battery packs.

If Fremont was shutdown for maintenance at the end of Q3, it would have been a good time to make some packs and send them to China.

 

[kengchang]

It makes sense, if:

• if they are produced opportunistically, like Carsonight suggested, when GF1 was running faster than Fremont could take the battery packs.
• Furthermore if they knew in advance that they won't have enough spare capacity at GF1 in Q4, because by that time the supply chain would be ramped up at Fremont.

We do know it from shipment manifests that Tesla is sending battery packs from GF1 to Shanghai, ~19 per container, per @kengchang.

What we don't know is the volume: just a handful for trial production, or the rumored 7,000?

I started working on a parser last night and it looks promising...

 

[Doggydogworld]

It makes sense, if:

• if they are produced opportunistically, like Carsonight suggested, when GF1 was running faster than Fremont could take the battery packs.
• Furthermore if they knew in advance that they won't have enough spare capacity at GF1 in Q4, because by that time the supply chain would be ramped up at Fremont.

We do know it from shipment manifests that Tesla is sending battery packs from GF1 to Shanghai, ~19 per container, per @kengchang.

What we don't know is the volume: just a handful for trial production, or the rumored 7,000?

Alternate theory. Per the 10-Q, Tesla renegotiated supplier contracts in Q3. To get lower prices and a big Q3 COGS reduction, Tesla accepted delivery of higher volumes in Q3. They'll need those parts soon enough with GF3 & Model Y starting up, so why not?

Separate issue I need help with. The Q3 letter showed 1 billion of TTM FCF (technically OCF - Capex). But net debt only changed by ~800m over those 12 months, roughly the amount of cash raised in May's equity offering. So where did the billion go? Close to 100m went into leased solar systems - that investment is outside of the OCF-Capex equation. But I can't figure out where the rest went, and I've confused myself a few times trying. Can anyone who understands this help me out?

 

[ReflexFunds]

Alternate theory. Per the 10-Q, Tesla renegotiated supplier contracts in Q3. To get lower prices and a big Q3 COGS reduction, Tesla accepted delivery of higher volumes in Q3. They'll need those parts soon enough with GF3 & Model Y starting up, so why not?

Separate issue I need help with. The Q3 letter showed 1 billion of TTM FCF (technically OCF - Capex). But net debt only changed by ~800m over those 12 months, roughly the amount of cash raised in May's equity offering. So where did the billion go? Close to 100m went into leased solar systems - that investment is outside of the OCF-Capex equation. But I can't figure out where the rest went, and I've confused myself a few times trying. Can anyone who understands this help me out?

It is roughly:

Free cash flow: $972m
Solar purchases: -$98m
M&A: -$57m
Collateralised leases: -$518m
Capital Leases: -$297m
Stock options: +$243m
Equity raised: +$848m
Convert hedges purchased (net): -$302m.
Issuance fees: -$42m

Total debt increased $1.6bn to $12,475m while net debt decreased $792m to $7,137m. This is using real outstanding debt numbers rather than balance sheet debt numbers.

It's also worth noting that during the period warehouse lines + ABS outstanding increased by $341m. So excluding this build of the Model 3 lease fleet, corporate net debt reduction was more like $1.1bn.

 

[Fact Checking]

Alternate theory. Per the 10-Q, Tesla renegotiated supplier contracts in Q3. To get lower prices and a big Q3 COGS reduction, Tesla accepted delivery of higher volumes in Q3. They'll need those parts soon enough with GF3 & Model Y starting up, so why not?

This certainly seems plausible, as apparently they recently completed their automated warehouse at Fremont, according to @kengchang:

... so they could have stockpiled parts with lower storage footprint by the tens of thousands - which could explain some of the Q3 increase in raw materials inventory.

Beyond increasing scale this would also reduce tariff risks - while adding some FX risk due to higher outstanding inventory.

Yet this can only be part of the Q4 story, according to RO-RO ship tracking (h/t @elasalle):

Tesla Carriers​

Esitimating for end of business day on November 20, Fremont outgoing shipments are trending at about:

• ~38 "loading days" on the 51th day of the quarter (74.5% port capacity utilization),
• which best compares to ~23 loading days on the 54th day of Q3 (42.6% port capacity utilization).

So assuming that "loading days" are QoQ comparable (they might not be, and we don't know whether they changed port logistics), and that they are accurately tracked (which they might not be - but my confidence in that is higher), that's ~+75% QoQ growth in outgoing Fremont RO-RO shipments over Q3, with already 76% into Q4's ~67 days time window of west coast originated RO-RO shipments.

Even assuming that 20-30% of that growth is due to the unprecedented starvation of U.S. Model 3 markets by dedicating 100% of Fremont production to international markets, there's another +46%-56% QoQ Model 3 production growth, to the extent we can trust this metric.

As @KarenRei, @Lycanthrope, @Right_Said_Fred and others have documented, the EU side observable Model 3 deliveries rate has ticked up significantly as well - although it's too early to attempt to extrapolate Q4 EU deliveries from those numbers.

In Q3 they made 80k Model 3s, 6,087/week without downtime, in Q4 this extrapolates to a production level of 8,900-9,500/week, 117k-125k Q4 Model 3s in Fremont alone ...

Assuming the ship loading days methodology does not break down in Q4 big time, and GF3 can make a few thousand Model 3s as well by December 31, they might be shooting for the magical 500k/year Model 3 production rate that was the subject of ridicule and the subject of a SEC lawsuit as well, which would be 125k units in Q4.

~30k more units delivered in Q4 over Q3 would be a bombshell, it would also explode GAAP profits and cash flow, due to continued CoGs improvements and various fixed cost absorption effects.

And, I never thought I'd calculate this, but the Q4 GAAP profit threshold for S&P 500 inclusion is $968m ... Just outside the models even with a +30k units deliveries QoQ growth, but maybe doable with a S&X deliveries and margins surprise and a bit of ZEV and FSD revenue recognition.

Certainly a pie in the sky stretch goal at the moment, with less than 10% probability, but those mid January $700 lottery tickets are still trading at 3 cents, which looks like sloppy, somewhat complacent MM risk management to me.

Anyway, if even a more modest and more realistic +20k units growth over Q3 materializes (which is a big if and +25% QoQ growth), then $420 after the January ~3 production and deliveries report seems like a real possibility.

 

[Thekiwi]

Seem like a bit of a stretch - how could production increase that much without the addition of any new GA limes or stamping machinery etc?

Perhaps the per day loading capacity has changed with the increased number of shipments - it is possible last quarter they had more cars arriving at port from the factory on days when no ship was present (meaning they had a stockpile of cars and consequently more to load each day when ship was berthed), whereas this quarter they could be loading into ships at a lower rate per day and essentially taking each days full factory production.

I spent a decade working for a company that had RoRo ferries that transported some cars (nowhere near the same level as a dedicated RoRo car carrier) and it was always more efficient to unload a car transport truck directly onto a waiting ship, than it was to unload the truck onto the wharf, and then have to load the car onboard at a later time (obviously requires two periods of driving, rather than one).

 

[Fact Checking]

New Seem like a bit of a stretch - how could production increase that much without the addition of any new GA limes or stamping machinery etc?

Here's a CleanTechnica article from last year that sheds some light on this question, reporting about Fremont production capacity constraints, when the factory was inspected by independent car industry experts:

Tesla Model 3 Production Has 25-33% The Number Of Steps "Found In Traditional Auto":​

• "So, what do Galliers and Ellinghorst think about the big question: can Tesla sustain, and hopefully increase, its production rate? “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."

That report was I think forgotten about and maybe even discarded by many investors due to Q1'2019 - but I think the audit by Galliers and Ellinghorst can still be trusted.

I presume their conclusions strongly implies that the press lines already had the capacity back then - and they had more than a year since then, and probably some of the Model Y capacity in shared facilities might already be online:

• stamp lines,
• paint shop expansion (underutilized due to S/X weakness as well),
• seat factory,
• plastics workshop,
• powertrain factory,
• glass factory,
• battery pack factory, etc.

One GF1 leak reported that Tesla planners expressed an "aspirational" goal of GF1 pack production rate of 15k/week - but then settled on "only" 10k/week by the end of next month...

Even with GF3 that's a lot of battery packs, which must be going somewhere?

Model Y production requires a new body line and a new general assembly line - but much of it is in the expansion of the above shared facilities. (Maybe @Krugerrand wants to correct any mistakes in my thinking, or add to this line of thought.)

What better way to stress test much of the Model Y production capacity and ramping up the shared part of the supply chain by actually using it for Model 3 production for a "peak quarter" - and then in Q1 and Q2 gradually shift those supplies to GF3 and Model Y production, without having to do an intrusive, time consuming ramp-up of the supply chain, which process is also stock full of operational risks due to the many moving parts largely out of Tesla's control?

(Obviously take this with a big grain of salt, not advice, your options will most likely expire worthless, etc.)

 

[kengchang]

This certainly seems plausible, as apparently they recently completed their automated warehouse at Fremont, according to @kengchang:

That building was completed long time ago..

 

[Thekiwi]

Here's a CleanTechnica article from last year that sheds some light on this question, reporting about Fremont production capacity constraints, when the factory was inspected by independent car industry experts:

Tesla Model 3 Production Has 25-33% The Number Of Steps "Found In Traditional Auto":​

• "So, what do Galliers and Ellinghorst think about the big question: can Tesla sustain, and hopefully increase, its production rate? “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."

That report was I think forgotten about and maybe even discarded by many investors due to Q1'2019 - but I think the audit by Galliers and Ellinghorst can still be trusted.

I presume their conclusions strongly implies that the press lines already had the capacity back then - and they had more than a year since then, and probably some of the Model Y capacity in shared facilities might already be online:

• stamp lines,
• paint shop expansion (underutilized due to S/X weakness as well),
• seat factory,
• plastics workshop,
• powertrain factory,
• glass factory,
• battery pack factory, etc.

One GF1 leak reported that Tesla planners expressed an "aspirational" goal of GF1 pack production rate of 15k/week - but then settled on "only" 10k/week by the end of next month...

Even with GF3 that's a lot of battery packs, which must be going somewhere?

Model Y production requires a new body line and a new general assembly line - but much of it is in the expansion of the above shared facilities. (Maybe @Krugerrand wants to correct any mistakes in my thinking, or add to this line of thought.)

What better way to stress test much of the Model Y production capacity and ramping up the shared part of the supply chain by actually using it for Model 3 production for a "peak quarter" - and then in Q1 and Q2 gradually shift those supplies to GF3 and Model Y production, without having to do an intrusive, time consuming ramp-up of the supply chain, which process is also stock full of operational risks due to the many moving parts largely out of Tesla's control?

(Obviously take this with a big grain of salt, not advice, your options will most likely expire worthless, etc.)

Some excellent points there! Totally agree it makes sense to have the shared facilities up to speed before Y production begins.

 

[Doggydogworld]

It is roughly:

Free cash flow: $972m
Solar purchases: -$98m
M&A: -$57m
Collateralised leases: -$518m
Capital Leases: -$297m
Stock options: +$243m
Equity raised: +$848m
Convert hedges purchased (net): -$302m.
Issuance fees: -$42m

Total debt increased $1.6bn to $12,475m while net debt decreased $792m to $7,137m. This is using real outstanding debt numbers rather than balance sheet debt numbers.

It's also worth noting that during the period warehouse lines + ABS outstanding increased by $341m. So excluding this build of the Model 3 lease fleet, corporate net debt reduction was more like $1.1bn.

Thanks, this is helpful. I just used the balance sheet line items for "long-term debt and finance leases" to calculate net debt. I was hoping that would simplify things (e.g. avoid backing non-cash interest out of OCF, etc.). I'll work it back through using your approach.

The numbers add up, of course, but I'm trying to get a better feel for what they mean. For example, Tesla gets cash every month from solar panel lessees. That cash flow is almost entirely additive to OCF and FCF. But that cash must go right back out the door to repay project-level lenders and tax equity partners. Since Tesla can't use that cash to build factories or whatever, it's not "real" FCF the way we normally think of it. It's not a huge deal at ~300m per year, it's not "fwaud" and Tesla should get to keep cash flow from those systems eventually. But it requires an adjustment that few people touting Tesla's FCF bother to make.

Since a lot of that 300m of outflow reduces (non-recourse) debt, it comes out in the wash when I compare FCF to the change in net debt. But when I see stuff like 518m of collateralized lease repayments, which is part of the dozen moving parts related to RVG accounting, I don't understand it well enough to make similar adjustments. And it makes me wonder if there's a similar "inflows count toward FCF, outflows don't count against it" effect in there somewhere.

 

[ReflexFunds]

Thanks, this is helpful. I just used the balance sheet line items for "long-term debt and finance leases" to calculate net debt. I was hoping that would simplify things (e.g. avoid backing non-cash interest out of OCF, etc.). I'll work it back through using your approach.

The numbers add up, of course, but I'm trying to get a better feel for what they mean. For example, Tesla gets cash every month from solar panel lessees. That cash flow is almost entirely additive to OCF and FCF. But that cash must go right back out the door to repay project-level lenders and tax equity partners. Since Tesla can't use that cash to build factories or whatever, it's not "real" FCF the way we normally think of it. It's not a huge deal at ~300m per year, it's not "fwaud" and Tesla should get to keep cash flow from those systems eventually. But it requires an adjustment that few people touting Tesla's FCF bother to make.

Since a lot of that 300m of outflow reduces (non-recourse) debt, it comes out in the wash when I compare FCF to the change in net debt. But when I see stuff like 518m of collateralized lease repayments, which is part of the dozen moving parts related to RVG accounting, I don't understand it well enough to make similar adjustments. And it makes me wonder if there's a similar "inflows count toward FCF, outflows don't count against it" effect in there somewhere.

Actually I missed the VIEs impact from the above breakdown, but it is not a large cash flow driver (aside from non recourse debt repayment which doesn't impact change in net debt).
VIEs cash flow impact is the sum of "Proceeds from investments by noncontrolling interests in subsidiaries", "Distributions paid to noncontrolling interests in subsidiaries" and "Payments for buy-outs of noncontrolling interests in subsidiaries".
This was -$46m in the LTM. Due to the stupid accounting rules for VIEs, this net cash flow change is what feeds through the P&L in the "Net income (loss) attributable to noncontrolling interests and redeemable noncontrolling interests in subsidiaries " line.
Most of the third party equity in VIEs is actually repaid using tax credit proceeds which never hit Tesla's consolidated P&L or cash flow, either +ve or -ve. Tesla only books revenue from actual customer lease payments as solar revenue - the tax equity is all sold separately off balance sheet. Only 20-30% of cash flows from actual customer leases are used to repay VIE partners - VIE partners are mostly paid back with the tax equity proceeds. I discussed this in a bit more detail below, but it's still confusing.

The VIEs were structured primarily to monetize the tax credits which Solarcity/Tesla couldn't monetise directly - it was mostly the tax credits that were monetized, and only a much smaller amount of future customer revenue.
Future customer revenue was monetized more via non recourse debt financing - so a majority of solar revenue will be used to repay non recourse solar debt, but this doesn't impact revenue, profit or free cash flow, just net cash flow.

My understanding broadly is that c.50-75% of cash raised from VIE partners was repaid in year one with tax credit proceeds. The rest is repaid to VIE partners out of solar lease customer cash flows over 5-30 years. Solarcity used a variety of different contract structures, but for most, after the initial tax credit proceeds was paid out to VIE partners, then only 20-30% of future cash flows (from actual customer revenues rather than tax credit proceeds) were paid out to VIE partners - with the bulk of cash flow paid by VIEs to Solarcity/Tesla. Many people claim that the bulk of cash flows generated by VIEs for the first 5/10 years goes to VIE partners rather than to Solarcity - this is strictly true, but only because the tax credit proceeds are paid directly back to VIE partners - the cash flow generated by actual customer revenue mostly goes to Solarcity, even for the initial years.
This VIE structure doesn't impact revenue though - just cash flow and net profit.

The Collateralised Leases line is all from vehicle leases. This is from cars previously sold to leasing partners with a "Resale Value Guarantee and a Buyback Option" or "where Exercise is Probable". No new leases qualify for this accounting treatment any more, so no new cars are sold under this scheme. This line wasn't always negative in the cash flow - when the lease fleet (qualifying for this accounting treatment) was growing, this cash flow item was positive as it reflected the cash received upfront from leasing partners, but which will only be registered as revenue over several years. Now this portfolio is fully in runoff, so the number is just cancelling out revenue booked from these cars in this period for which cash was received upfront several years ago. This fleet is rapidly running off, so should soon become less significant in the cash flow. This line item changes the resale value guarantee liability, but it doesn't impact real debt.

The most confusing line item is capital leases/finance leases. PP&E under capital leases stood at $688m at 2017, $1,520m at 2018 and $1,980m at 9M19. I think it most likely this is nearly all Panasonic and other supplier's GF1 equipment. Panasonic equipment within PP&E was $473m in 2017, $1.24bn at 2018 and $1.65bn at 9M19. And it sounds like this equipment is accounted as a capital/finance leases: "As these terms convey a finance lease, as defined in ASC 842, Leases, their production equipment, we consider them to be leased assets when production commences. This results in us recording the cost of their production equipment within property, plant and equipment, net, on the consolidated balance sheets with a corresponding liability recorded to long-term debt and finance leases."

So when Panasonic installs equipment at GF1 Tesla adds this assets to its PP&E and also adds a finance lease liability to its balance sheet. This hits the P&L through interest on the total PP&E finance lease balance (maybe around $25m per Q currently if at 5% on the $2bn balance) and also as depreciation ($174m depreciation on capital lease assets in in 9M19, so around $60m per Q). However this is all non cash, so will get added back somewhere within the operating cash flow adjustments (mostly depreciation addback). The real actual lease payment should then hit the cash flow separately through the finance/capital lease line. However I don't understand why Tesla should actually pay cash lease payments to Panasonic for its battery equipment. Surely this should be paid within COGs for battery purchases? So its possible these capital leases in the cash flow do not actually relate to Panasonic, or else maybe they are cancelled out somewhere else in the cash flow. I've never seen the full details of the Panasonic contract so its hard to know.

 

[Doggydogworld]

This certainly seems plausible, as apparently they recently completed their automated warehouse at Fremont, according to @kengchang:

View attachment 479295​

... so they could have stockpiled parts with lower storage footprint by the tens of thousands - which could explain some of the Q3 increase in raw materials inventory.

Beyond increasing scale this would also reduce tariff risks - while adding some FX risk due to higher outstanding inventory.

Yet this can only be part of the Q4 story, according to RO-RO ship tracking (h/t @elasalle):

Tesla Carriers​

Esitimating for end of business day on November 20, Fremont outgoing shipments are trending at about:

• ~38 "loading days" on the 51th day of the quarter (74.5% port capacity utilization),
• which best compares to ~23 loading days on the 54th day of Q3 (42.6% port capacity utilization).

So assuming that "loading days" are QoQ comparable (they might not be, and we don't know whether they changed port logistics), and that they are accurately tracked (which they might not be - but my confidence in that is higher), that's ~+75% QoQ growth in outgoing Fremont RO-RO shipments over Q3, with already 76% into Q4's ~67 days time window of west coast originated RO-RO shipments.

Even assuming that 20-30% of that growth is due to the unprecedented starvation of U.S. Model 3 markets by dedicating 100% of Fremont production to international markets, there's another +46%-56% QoQ Model 3 production growth, to the extent we can trust this metric.

As @KarenRei, @Lycanthrope, @Right_Said_Fred and others have documented, the EU side observable Model 3 deliveries rate has ticked up significantly as well - although it's too early to attempt to extrapolate Q4 EU deliveries from those numbers.

In Q3 they made 80k Model 3s, 6,087/week without downtime, in Q4 this extrapolates to a production level of 8,900-9,500/week, 117k-125k Q4 Model 3s in Fremont alone ...

Assuming the ship loading days methodology does not break down in Q4 big time, and GF3 can make a few thousand Model 3s as well by December 31, they might be shooting for the magical 500k/year Model 3 production rate that was the subject of ridicule and the subject of a SEC lawsuit as well, which would be 125k units in Q4.

~30k more units delivered in Q4 over Q3 would be a bombshell, it would also explode GAAP profits and cash flow, due to continued CoGs improvements and various fixed cost absorption effects.

And, I never thought I'd calculate this, but the Q4 GAAP profit threshold for S&P 500 inclusion is $968m ... Just outside the models even with a +30k units deliveries QoQ growth, but maybe doable with a S&X deliveries and margins surprise and a bit of ZEV and FSD revenue recognition.

Certainly a pie in the sky stretch goal at the moment, with less than 10% probability, but those mid January $700 lottery tickets are still trading at 3 cents, which looks like sloppy, somewhat complacent MM risk management to me.

Anyway, if even a more modest and more realistic +20k units growth over Q3 materializes (which is a big if and +25% QoQ growth), then $420 after the January ~3 production and deliveries report seems like a real possibility.

Q1-Q3 had 94 loading days for ~90k Model 3s. Q4 has 37 loading days through day 50 (11/19), so figure 35k Model 3s. Overseas production started a few days before Q3 ended, so we're talking about 7.5 weeks. By this rough math they're loading a little less than 5k/week.

The last EU ship sails from Port 80 by day 61. China ships sail later, but I don't expect that in Q4 since missing the potential 12/15 tariff hike would be dumb. Basically we've got 10 more days, and I'm confident the dockworker union negotiated a nice Thanksgiving break, so figure 7 more loading days. That's 44 days total, a new record, and ~42k overseas Model 3s. Maybe add 1-2k for a higher rate of container shipments to AU/NZ than in Q1-Q3.

This leaves 1k+/week to fill US orders. Despite the official 4-6 week wait time, my guess is sales people have access to "hidden" inventory for buyers who need a car immediately and will go elsewhere if not fulfilled.

 

[EVNow]

In Q3 they made 80k Model 3s, 6,087/week without downtime, in Q4 this extrapolates to a production level of 8,900-9,500/week, 117k-125k Q4 Model 3s in Fremont alone ...

What are you extrapolating from ?

85k looks like a decent extrapolation.

Q1 '19 62,975
Q2 '19 72,531
Q3 '19 79,837
Q4 '19 est 85,000

 

[Fact Checking]

Q1-Q3 had 94 loading days for ~90k Model 3s. Q4 has 37 loading days through day 50 (11/19), so figure 35k Model 3s. Overseas production started a few days before Q3 ended, so we're talking about 7.5 weeks. By this rough math they're loading a little less than 5k/week.

The last EU ship sails from Port 80 by day 61. China ships sail later, but I don't expect that in Q4 since missing the potential 12/15 tariff hike would be dumb. Basically we've got 10 more days, and I'm confident the dockworker union negotiated a nice Thanksgiving break, so figure 7 more loading days. That's 44 days total, a new record, and ~42k overseas Model 3s. Maybe add 1-2k for a higher rate of container shipments to AU/NZ than in Q1-Q3.

This leaves 1k+/week to fill US orders. Despite the official 4-6 week wait time, my guess is sales people have access to "hidden" inventory for buyers who need a car immediately and will go elsewhere if not fulfilled.

What are you extrapolating from ?

85k looks like a decent extrapolation.

Q1 '19 62,975
Q2 '19 72,531
Q3 '19 79,837
Q4 '19 est 85,000

Sigh, I messed up: I applied Q3 loading-day speed (1.1k/day) to estimate quarterly production - but that's invalid.

Here's how I estimated the per quarter loading day and international production figures:

• Q1: 40 days, ~38k Model 3s international
• Q2: 24.4 days, ~25k Model 3s international
• Q3: 30.1 days, ~35k Model 3s international
• Q4: ~40 days already on day #53 of ~60-67 days window (adding in the currently moored CN ship which arrived 3 days ago) - extrapolated to 45-50 loading days depending on the Thanksgiving disruption and China timing.

Note that my sum for Q1-Q3 international is 98k, @Doggydogworld's is 90k. I used the Insideevs estimates and guesses for Canada.

Note the faster apparent loading in Q3 - but my numbers could be off.

If we apply Q3 loading speed (1,163/loading-day) to the Q4 extrapolated range, we get 52k-58k production for the first 60-67 days, which if we extrapolate it to 92 days gives us a Q4 production of 80-89k.

Sorry.

 

[EVNow]

If we apply Q3 loading speed (1,163/loading-day) to the Q4 extrapolated range, we get 52k-58k production for the first 60-67 days, which if we extrapolate it to 92 days gives us a Q4 production of 80-89k.

Yes - ~85k looks like a decent estimate.

Through out this year, we have had big misses in EPS, even deliveries - but our production numbers have been quite accurate. It is obviously more difficult to move this wildly (since this is under Tesla's control) - than deliveries or EPS.

I just hope they are filling a lot of orders from Netherland - and leaving the rest for Q1. That would be a good way to balanceQ4 & Q1 deliveries.

 

[Thekiwi]

I do wonder if they are also pacing the potential China delivery start date based on how well Q4 is already setting up to be. Not that they would delay things if it was 100% ready for delivery start, but perhaps they wouldn’t feel the need to try and rush to get to that 100% level earlier.

 

[Fact Checking]

I started working on a parser last night and it looks promising...

Wondering whether you managed to get any battery pack container shipment estimates from the U.S. to China, or is that not possible to extract from the data?

 

[Fact Checking]

The last EU ship sails from Port 80 by day 61. China ships sail later, but I don't expect that in Q4 since missing the potential 12/15 tariff hike would be dumb. Basically we've got 10 more days, and I'm confident the dockworker union negotiated a nice Thanksgiving break, so figure 7 more loading days. That's 44 days total, a new record, and ~42k overseas Model 3s. Maybe add 1-2k for a higher rate of container shipments to AU/NZ than in Q1-Q3.

We are already at 45 days - and in Q3 they also loaded an EU ship in PHY which departed on September 9, and a CN ship in SFO that left on September 4. They might or might not do it this quarter - the window of opportunity closes at around December 10 or so.

Regarding a Thanksgiving break: I've read up the west coast longshoremen union contracts, and it says nothing about a Thanksgiving break: federal holidays including Thanksgiving are recognized as "overtime", with a 60% increase in pay for those days. San Francisco port dock workers are still working 24/7, in 3 shifts and on holidays as well. (I suspect there might be extra bonuses offered as well, should they be short on staff for any given day.)

The ship tracker appears to have confirmed this: "Rcc Europe" left a couple of hours ago, with 3.72 loading days (2nd highest in the quarter), and only 0.79 idle days, most of which idle time was accumulated on the 26th.

I.e. they were loading what appears to be the last EU ship from SFO in Q4 all through Thanksgiving.

Q4 has accumulated 45 loading days so far, which is +45% over Q3 loading days.

Further notes:

• There's some uncertainty about whether the last Q3 ship, the Lydden, headed for China, and which unloaded on the 24th of September, managed to deliver those units in the last ~6 days of Q3. If not then it's 30 days for Q3 and 46 days for Q4 so far, a +53% increase over Q3.
• Even if the last Q3 ship was delivered in Q3, there might be a similar China-headed ship in Q4 as well: the window of opportunity probably closes on the 5th of December, still a week until then. BTW., the departure date of that ship will possibly tell us whether they sent the Q3 Lydden too late.