On topic of Q4 cash flow forecasts, it's important to note the sustainable nature of Q3 cash flow. It was driven by genuine long term cash flow per model 3, and not short term movements in payables, inventory and receivables.
In Q3 I estimate average revenue booked by Tesla per model 3 was $57k, an additional $2.2k was booked as GHG credit revenue and a further c.$2k was recorded as deferred revenue (reserve for future internet, software update & Autopilot commitments). (So ASP was $59k).
Therefore total cash received per model 3 was $61.4k, excluding short term working capital.
COGs recorded per model 3 was $47.7k but of this $3.1k was non cash depreciation and $2.1k was non cash warranty reserve (this will only consume cash over several years). So total cash consumed per model 3 was just $42.5k.
So while gross profit recorded per model 3 was $11.4k (20% margin, 16.2% ex GHG sales) or a total $636m, cash generated was actually $18.7k (31.5% margin) or a total $1,041m.
Note this is excluding payables, inventory and receivables working capital balances, which will roughly scale as a % of model 3 sales in the short term, but are not long term cash flow. There is significant room for these shorter term working capital items to generate more cash in Q4 relative to Q3.
A few consequences of this analysis for the profitability of Tesla’s base $35k model 3:
I estimate the base model 3 would have cost $40.6k to produce in Q3 (including provision warranty costs, but not including deferred revenue related costs). This is down from c.$49k in Q2.
Elon disclosed a current base model 3 production cost of $38k in November. I presume this is on an accounting basis (which is how Elon normally talks about margin guidance) like my $40.6k, but it is possible Elon’s number includes deferred costs (in which case cost would be $36k on my basis) and also possible it doesn’t include warranty (in which case it would be $40.1k on my basis).
To achieve 25% gross margins on average across the model 3 product range (using my estimate for long term take rate and margin for each option), Tesla needs to reduce this base model 3 cost from $38k to $32.3k. To get to Tesla’s long term target 30% model 3 gross margins, the base cost needs to reduce to $30k (which happens to be the model 3 cost target Elon gave on the Q3 call).
Right now, at a $38k production cost for model 3, the base model would generate revenue of $36.9k per car ($35k + $1.2k delivery + $2.2k GHG credit sales -$1.5k deferred revenue). Reported gross profit would be -$1.5k and cash flow would be $4.8k.
The easiest way for Tesla to reduce costs is to increase production, leveraging staff costs and depreciation. Going from 4.5k to 7k per week is unlikely to increase depreciation materially, and staff cost increase is likely very small as well (production volume gain is mostly from increased productivity per employee per hour). This alone could possibly reduce cost per car by c.$3k. Further significant costs savings will come from the new module design and lines set to ramp up in Q1.
If Tesla is demand constrained in the US with the current expensive model 3 options, then the margin they will achieve on releasing and selling additional $35k base cars should use the marginal production cost and not the average. (Note this doesn’t make sense if the $35k model will significantly cannibalise higher margin sales, but if released just to remaining US reservation holders it is pretty clear by now they are not going to trade up). The marginal gross profit of selling 2.5k base model 3s excludes most of the D&A and staff costs and could actually generate marginal gross profit per base car of c.$6.5k right now, even at the $38k average production cost. Marginal cash flow would be c.$10k for each 35k car sold. While this would generate Tesla more gross profit in absolute terms, it would drag down the overall % margin.