Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register

Tesla, TSLA & the Investment World: the Perpetual Investors' Roundtable

This site may earn commission on affiliate links.
They didn't say that. There's a wide spectrum of technological solutions that are not "100% autonomous all the time", which are still economically superior to "taxi driver" based service:
  • A fleet dispatch center with "remote drivers" who are trained to take over when there's an FSD disengagement event.
  • Speed limits for when the car is driving unattended.
  • Geofencing and geomapping based on fleet feedback.
  • A quick response fleet technicians who can recover malfunctioning or stuck cars.
Tesla Network doesn't have to be all or nothing, it has to generate money and can grow constantly.

Returned cars from leases could be a Turo like service owned by Tesla as well. Pick up and drop off can be done independently by the renter via an app giving them a temporary key to the car they've rented. I'd rent a 3 year old Model 3 over a newer ICE rental car any day. Enabling FSD on all such rentals would sell many cars. EAP is decisively the feature that most impresses me about my dual motor 3 (well to be fair, the repeatedly improving EAP heading to FSD, so the OTA updates are a big part of this).
 
Hmmmm.... that is interesting timing. So you're suggesting they were cell-limited during the crucial months of January and February when they were shipping to Europe and Asia, and stopped being cell-limited just in time to switch back to US production (which was obviously suffering a hangover from the tax credit expiration)?

I think they were cell constrained during January and February, and they might not have increased production in March even if they could have, to keep U.S. inventory at a reasonable ~2 weeks level. Do we have good estimates about precisely how many Model 3's were shipped to Europe and China? Since the last ship left early March that would give us a good idea about March U.S. production.

They might also have been stockpiling LR battery packs from Q1 into Q2 though, the 10-Q could tell us if there's a run-up in 'parts' and 'materials' inventory compared to Q4 which had similar production levels. A LR pack would increase inventory by less than $10k, so it's a lot easier to stockpile say 5k battery packs (a ~$50m cash expense) than increase finished goods inventory by 5k cars (a ~$175m cash expense).

So there's a number of scenarios that are consistent with their press release from yesterday that they are cell supply limited.
 
My only question is production. And no, it's not good news that they are killing two variants; it implies that they need to do so in order to get production rates up. (LR RWD was popular. There are only two reasons to kill it: one, it has worse margin than AWD, but in that case, why not raise the price; and two, having more variants is a problem. This tells me that having more variants is a problem.)

There's a third, fifth and sixth possible reason as well:
  • If SR+ generates more margin per cell, then it makes more sense to sell SR+ than LR RWD, if they are cell constrained. The LR battery pack consumes ~40% more cells than SR+.
  • An additional factor is that to Tesla it makes more sense for people to spend money on a $5k software option like FSD (100% margin), than on "more cells" (not only lower margin but is also crowding out other sales). In Europe the FSD take-rate appeared to be very high in March already: 40%. This will further improve as FSD functionality improves.
  • Increasing units to Europe increases the value of the FCA "pooling agreement" - which could either increase income dynamically in 2019 already, or could improve Tesla's negotiation position in 2020. Every new Tesla unit shipped reduces FCA's fines by about $8k-$10k... This I believe is one of the factors why the SR+ was unlocked for Europe today.
I.e. the basic business logic for Tesla is now to increase unit sales, while keeping per cell margins constant.

BTW., I'm pretty sure that the SR+'s margins are pretty good and the per cell margin is higher, and it's the primary reason why they discontinued the Medium Range without much fanfare: it consumed too many cells.
 
Last edited:
How much could an Uber driver earn if they worked 22hrs a day, 360 days a year?

Multiply by .75, to ensure driverless is cheaper than driven.

That's the earnings potential of a driverless car, in the first year of service.

It gives a ballpark for how much the value of these cars appreciates when nation one legalises autonomous Teslas. Subtract shipping cost to said nation.
 
Returned cars from leases could be a Turo like service owned by Tesla as well. Pick up and drop off can be done independently by the renter via an app giving them a temporary key to the car they've rented. I'd rent a 3 year old Model 3 over a newer ICE rental car any day. Enabling FSD on all such rentals would sell many cars. EAP is decisively the feature that most impresses me about my dual motor 3 (well to be fair, the repeatedly improving EAP heading to FSD, so the OTA updates are a big part of this).
I’d prefer being able to rent a Tesla when I’m on holiday (or anywhere I don’t have access to my own car), and certainly if it is a near-zero effort to do so. Right now car sharing is too fragmented to bother signing up to any of the local car sharing services.
I’d only use Uber/Lyft in dense city situations where parking is an issue.
FSD is not a prerequisite to make Tesla Network a success, just ubiquitous access to a car is a prerequisite. Which is something Tesla can add to their cars with zero marginal cost, i.e. purely in software. Other car sharing services need a solution for sharing and limiting access to keys.
FSD will only make the market accessible to Tesla Network much bigger.
 
No idea how people here a trying to spin this into a positive news?!

The optics of this whole back paddling again, killing SR, and enabling SR+ overseas without any fanfare is not just bad, it's effing catastrophic.

Good luck, my friends - you'll need it.

People do. The market and media do, too.

It's a broken promise and some former Tesla fan/sympathizers feel betrayed be that. On an human level, there's nothing good about that news. Absolutely nothing.

Sorry, but you are talking 100%, complete, absolute, breathtakingly unadulterated nonsense there:
  • You can order a Model 3 Standard Range Plus in Germany for €44.500 today.
  • $35,000 plus ~$3,000 shipping costs, plus 10% import tariffs plus 19% VAT in Germany at a 1.12 exchange range is €43,600.
For this €44.500, which is €900 on top of the 2016 price (with 0% inflation, before incentives) you get the following features over the 2016 Standard Range Model 3:
  • :) A SPECTACULAR glass roof, (metal roof promised in 2016),
  • :p ~10% higher range (220 miles promised in 2016, 240 miles delivered today),
  • ;) ~200 kW peak charge rate, (twice of the charge rate of 2016),
  • :D Autopilot, (was a $5,000 option in 2016),
  • :rolleyes: better acceleration, (5.6s promised, 5.3sec delivered),
  • :cool: upgraded interior, (basic, all-plastic interior promised in 2016),
  • :D 12-way power adjustable heated front seats, premium seat material and trim, custom driver profiles (only basic seats promised in 2016),
  • :p better audio, (~twice the number of speakers, "Immersive Sound"),
  • ;) auto-dimming, power-folding, heated side mirrors, (simple mirrors planned in 2016),
  • :rolleyes: better online services,
  • :cool: center console with storage, 4 USB ports and docking for 2 smartphones. (Simpler trim planned in 2016),
  • :) Dog Mode, Sentry Mode, Romance Mode and whatever future OTA upgrade Tesla can think of,
  • :D and a 10x faster FSD computer, which includes two of Tesla's revolutionary AI chips that might or might not fall under U.S. weapons export controls due to the danger they pose on humanity should SkyNet break free.
There's the promises Tesla kept, and in fact Tesla over-delivered fantastically:

giphy.gif


You are welcome! :D
 
Last edited:
I disagree. Some convolution neural nets CNNs now exceed humans in pure classification performance (ImageNet). With additional context and training through “shadow mode” it is entirely conceivable that FSD will be running in 3 years.

Forget about human timescales. Think of performance gains due to Moore’s Law. Read up on AlexNet and the ImageNet Challenge:
The data that transformed AI research—and possibly the world

Worth noting that Alex Krizhevsky and Ilya Sutskever were behind Alexnet and the paper that kicked off the deep learning revolution in 2012.
Ilya is Chief Scientist at OpenAi and worked with Andrej Karpathy before Andrej left to lead AI for Tesla. I also presume Elon's discussions with Ilya were influential in Tesla's AI strategy.
Alex worked for Google for several years, including work on Waymo. Who does Alex think is leading in self driving?:
“I think Tesla has the unique advantage of being able to collect data from a very wide variety of environments because there are Tesla owners with self-driving hardware all over the world,” he tells me. “This is very important for machine learning algorithms to generalize. So I would guess that at least from the data side, if not the algorithmic side, Tesla might be ahead.”
 
Last edited:
That's nonsense you are talking there:
  • You can order a Model 3 Standard Range Plus in Germany for €44.500 today.
  • $35,000 plus ~$3,000 shipping costs, plus 10% import tariffs plus 19% VAT in Germany at the 1.12 exchange range is €43,600.
There's the promise Tesla kept, and you got upgraded interior, heated seats, better audio, better services, Autopilot and an FSD computer for around €900, with 0% inflation.

You are welcome!
Tesla might have decided to flood EU market knowing that they can take pooling credit from other manufacturers and keep production line busy to lower fixed costs
 
Tesla might have decided to flood EU market knowing that they can take pooling credit from other manufacturers and keep production line busy to lower fixed costs

That, plus:
  • FSD take-rate in Europe was 40% back in March - which would add another ~$2k margin to sales.
  • If they are fundamentally cell supply limited as Tesla's press release disclosed yesterday, then it makes sense to sell SR+ units with ~55 kWh packs vs. LR units with ~78 kWh packs, which consume ~42% more cells.
I.e. if we start thinking about Tesla cars hardware platforms that enable the selling of software products then it makes sense for Tesla to increase the number of units sold.
 
FRA is probably not a great indicator anyway.

Remember last years (July?) "5k per week" announcement? +8% in FRA, US close under -5%.

Tesla effectively increased prices - Wall Street generally (but not always) reacts to price hikes positively.

Can you please correct the 'broken promise' false claims you made?
 
I think what's been driving all the maneuvering this year as far as pricing and available configurations is a question of balancing 1) keeping the number of variations of the Model 3 within Tesla's capacity to efficiently produce the car (to avoid the downtime issue in switching production from one variant to another which Neroden has been zeroing in on) with 2) having a selection of available variants sufficient to appeal to the consumer demand needed to match up with the targeted Fremont production of about 7k Model 3s per week.

That is, it isn't as simple as a demand issue, OR a production capability issue, but, more an issue of finding a product mix that can be realistically managed in terms of Tesla's production capabilities that has demand for at least 7k Model 3s per week. Think of that Elon line from a month or two ago, (paraphrasing) "It isn't if we have enough demand... the demand is there, we just need to make the car inexpensive enough for the people to buy the car."

I suspect that what we saw tonight reflects Tesla deciding that right-hand drive variants are a more efficient way of adding to their supply of what the market wants than offering the SR. Tesla may well feel it needs to go to adding right-hand drive availability to be confident about demand hitting 7k per week this year. To prevent adding right-hand drive availability from becoming utter production hell, drop the SR and LR from regular production. I suspect the market for RHD may be as much as 100k per year, (fwiw, I think the LR will come back once all markets have been cleared of their initial Dual Motor and P demand).

I think it is helpful to realize that while the long-term positioning for Tesla continues to look extremely compelling, they've seen a need to, and have made choices designed to tap a widening part of consumer demand to continue to go forward with raising production. We've seen several clear signs this has been done in a climate of softer present demand than previously anticipated. Tesla lowered prices twice this year, the second time, quite sharply. The Model 3 plan for Fremont has gone from 10k/week to 7k/week. At the end of February Tesla introduced the SR/SR+ about a quarter or two ahead of what they'd targeted just a month or so earlier (Q1 earnings call IIRC, they said summer or mid-year for these variants). Tonight, leasing was added for the US to tap more demand. Most evident to me that Tesla has felt pressure of finding the Model 3 product mix that tapped enough of the demand for the car was the media call for that 2/28 announcement. Elon consistently had said re Model 3 demand, ~500k-700k. When asked this question on the call, he said, with some agitation in his voice ~I don't know. Could only guess. 500,000? There's really no way to know~ (going off memory re exact phrasing, but, it was that big a shift). To be fair, I think since 2/28, things have been looking considerably more encouraging to those seeing the order rates. They raised prices back somewhat.

So, while the bearish ~"see, see, like we always said, demand is cratering" is nonsense, and, long-term I see Tesla selling millions of cars per year, I do think today a major aspect of Tesla's decisions re Model 3 offerings is finding a realistic expanded Model 3 product offering with enough demand to soak up Fremont production. This is neither a 'failure' or 'unthinkable, taboo, etc.' to consider.
 
Last edited: