Short-Term TSLA Price Movements - 2016

[AudubonB]

Develop on their own...or purchase from Tesla? What is the likelihood and what are the ramifications of that?

 

[TMSE]

Develop on their own...or purchase from Tesla? What is the likelihood and what are the ramifications of that?

Hmm... You are on to something. Neural net from Tesla sold like Windows OS? Could it be like Wintel (Microsoft + Intel) cartel of yesteryears? Dawn of nVidia + Tesla cartel? The likelihood of this happening is very low, as this neural net is not a distinguishable entity, but a closely coupled one with AI hardware, sensors, the placement/number of sensors, and more importantly constant learning and refining of this net based on the copious amount of data generated while driving. It is the integrated system that Tesla is after, just like Apple Mac.

 

[MitchJi]

Mitch,
I don't think that tesla is paying anywhere close to "few thousands of dollars". nVidia CEO might be quoting MSRP. Flagship customers usually get a much bigger discount, my guess is that Tesla might be paying high hundreds, comparable to a top-end GPU card.

I don't believe that their CEO lied on their earnings call when he said,in response to a question "a few thousand dollars ".

Your speculation on Tesla developing their own hardware is plausible, given the stellar semiconductor design team they have. It is likely they will develop their own Tesla drive using generic GPUs, either from nVidia or AMD. In fact, Tesla might already be using its own hardware with GPU from nVidia. Neither Tesla nor, nVidia has confirmed Drive PX usage in Tesla cars.

I'm pretty sure that it was confirmed, at the same time that they gave the price. One option is developing their own card with generic GPU's. I think that it's equally likely that they will develop their own custom hardware to do the radar associated DSP, possibly in combination with generic GPUs. Not that the distinction is important to the short term SP.

 

[generalenthu]

Hmm... You are on to something. Neural net from tesla sold like Windows OS? Could it be like Wintel (Microsoft + Intel) cartel of yesteryears? Dawn of nVidia + Tesla cartel? The likelihood of this happening is very low, as this neural net is not a distinguishable entity, but a closely coupled one with AI hardware, sensors, the placement/number of sensors, and more importantly constant learning and refining of this net based on the copious amount of data generated while driving. It is the integrated system that Tesla is after, just like Apple Mac.

I remember seeing a video somewhere, where Elon shot this down, basically saying that the entire system is too closely integrated with the rest of the car and it would be meaningless / too hard to make this a drop in solution for other vehicles.

 

[RobStark]

Ahahaha! Tesla doesn't need to take 90% of the global automotive market share for GM to be in trouble. Let's remind ourselves just how little GM sales had to drop before they went bankrupt. I'll give you hint: not anywhere near 90%, way less than 50%, even less than 25%, and under 10%.

LOL.

Tesla can take 50% of the compact sedan market and it would not materially effect GM profit .

Falling pickup and large SUV sales put GM into bankruptcy.

GM does need to meet CAFE goals in the USA and carbon emissions in Europe.

But there is more than one way to skin a cat.

 

[oneday]

I remember seeing a video somewhere, where Elon shot this down, basically saying that the entire system is too closely integrated with the rest of the car and it would be meaningless / too hard to make this a drop in solution for other vehicles.

I remember hearing that on the q3 conference call

Edit: maybe it was the ap2 press call?

 

[Krugerrand]

Tesla can take 50% of the compact sedan market and it would not materially effect GM profit .

Point me to the maths.

 

[DaveT]

Develop on their own...or purchase from Tesla? What is the likelihood and what are the ramifications of that?

In all likelihood, Tesla won't sell their neural net sofware because it's constantly evolving and it requires close integration with the car, hardware and sensors.

 

[Johan]

Point me to the maths.

https://www.google.no/amp/s/www.tru...-huge-first-quarter-profit/amp/?client=safari

Almost all their profit is from trucks and SUVs.

 

[MitchJi]

Develop on their own...or purchase from Tesla? What is the likelihood and what are the ramifications of that?

One ramification is that according to my interpretation of the Nvidia video (thanks for the link @DaveT), the two most important factors for the software quality are the quality of the AI software and the volume of training data. Nobody has the volume of data that Tesla has.

I remember seeing a video somewhere, where Elon shot this down, basically saying that the entire system is too closely integrated with the rest of the car and it would be meaningless / too hard to make this a drop in solution for other vehicles.

He was answering a question about a "kit". The list of hurdles he specified wouldn't rule out selling it to an OEM, assuming that they would implement OTA. Which Nvidia also said is necessary (another small moat).

In all likelihood, Tesla won't sell their neural net sofware because it's constantly evolving and it requires close integration with the car, hardware and sensors.

I think that they unlikely because, with the possible exception of a small startup, I don't believe that any other OEM would buy the technology from Tesla.

 

[30seconds]

Take comfort as you choose; no one here expected 2/3 rds of the 2018 notes to be redeemed 1.5 years early, causing an outflow of about $1 billion (some of it absorbed by the hedge writers).. Dragging payments to key suppliers works great until they get spooked. That practice needs an alternate supply channel for mission critical parts.

At this point it wouldn't be hard for them to get $1B bond offering without an early call. The cash flows are solid enough

 

[30seconds]

Old news from July this year.. Trump doesn't want to deal with Koch Bros. Might bodes well for Musk Industries.

TRUMP TURNS DOWN MEETING WITH KOCH BROTHERS:
TRUMP TURNS DOWN MEETING WITH KOCH BROTHERS » REGATED

Today, Donald J. Trump once again proved that he is unbuyable. He stated on his Twitter that “I turned down a meeting with Charles and David Koch. Much better for them to meet with the puppets of politics, they will do much better!”

Donald J. Trump on Twitter

The Koch brothers (Charles and David) own Koch industries, the second largest private company in the United States. They are well known for the amount of political work that they usually do. However, they seem to be staying out of this year’s election as much as possible.

Yep, no need with all those rubels burning a hole in his pocket

 

[DaveT]

Dear beloved DaveT,

Admittedly I was blind and failed to see and you were absolutely correct at predicting that Mr. Trump was going to win the historic 2016 presidency. A bet goes 2 ways, I wish Trump had lost so We would get to watch you eating your socks instead. I thought that was your end of a fair bet, wasn't it?

If what I quoted above wasn't clear about my seriousness of this one way bet, don't worry I intend to hold up my end. It's just that the difficult tasks above might take some time to complete - You know writing sock eating manual for the POTUS and all, there is no guarantee that you'd see my video Before seeing a completed mexican wall.

In addition, as my investment in this silly electric car company is going down the drain due to my inability to predict the outcome of this election, I hope My failed bet did not cause you or your dear friends to follow and lose money as I did. On second thought, I am sure it didn't as you are the wiser and more cautious one. I am very naive as you can see, and like you always say, part of the naive herd on this forum. Probably the worst of them all.

As a current priority, I have to scramble for more funds to double down on this silly money losing car company should its stock declines further to $160 as they say.

Again I sincerely apologize for the delay in producing sock eating video and all which will be done after Mr. Trump completed the great Mexican Wall. Feel free to help him build faster if you can't wait.

I hope you'd understand the challenges I am facing.

Sincerely yours,
Xpert the sock eating loser.

I guess I misunderstood it when you said "you have my word"... I thought that was an actual promise. But I guess I was wrong.

I will eat my socks on TV if Trump wins...

@Xpert Is that literal or figurative? If Trump wins, will you post a video here of you eating your socks?

Yes Dave, i will eat my unwashed stinky socks and post the video here. you have my words.

From your most recent post it seems clear you have no intent of following through (whether it's eating socks, shaving head, or something else), so I won't hold you to it.

I think the best resolution if the following: add me to your ignore list, and I'll add you to my ignore list. And we'll move on.

 

[techmaven]

I don't believe that their CEO lied on their earnings call when he said,in response to a question "a few thousand dollars ".

So, the actual quote is:

... you could start from one Parker SoC and that allows you to have surround camera. It allows you to use AI for highway cruising. And if you would like to have even more cameras, so that your functionality could be used more frequently in more conditions, you could always add more processors. And so we go from one to four processors. And if it's a fully autonomous driverless car, a driverless taxi, for example, you might need more than even four of our processors. You might need eight processors. You might need 12 processors.
[...]
And so I think that depending on the application that you have, we'll have a different configuration and it's scalable. And it ranges from a few hundred dollars to a few thousand dollars. And so I think it just depends on what configuration people are trying to deploy.

So the few thousand dollars version likely is 8-12 processors. Tesla clearly isn't using that version. Putting together Tesla's comments and Nvidia's comments, here are a few things that are clear to me:

1) Tesla makes their own software and it is portable between Nvidia, AMD, and Intel hardware
2) That means Tesla isn't using much if any of the ADAS/automated driving software stack provided by other people
3) NVIDIA sells a Drive PX 2 platform and Tesla is using some part of it
4) What Drive PX 2 platform entails varies widely... anywhere from a Parker SOC (Tegra core + Pascal GPU), to a reference board with two of those SOCs and a variety of hookups, to 12 SOC beasts on multiple boards. Plus software stack, including NVIDIA reference neural nets
5) NVIDIA wants to sell the Drive PX 2 platform which could mean a few chips to a whole shebang, which costs a few hundred dollars to a few thousand dollars, not including back end processing servers, software licensing costs, additional consulting hours, etc.
6) Tesla is ahead of everyone

Not clear if Tesla is using merely a couple of Parker SOCs on even their own board, using the base Drive PX 2 hardware board with two Parker SOCs, or something more. At the bottom, the board cost is likely ~$500. Clearly NVIDIA would love to sell a whole slew of stuff to people and want to use Tesla's success as a means to do so. Which is perfectly fine. But when other people buy NVIDIA's Drive PX2 platform in its soup to nuts form, the only thing it shares with Tesla's AP2 is the fact both use the same Parker SOC at the bottom and a similar software approach.

More on Parker SOC:
NVIDIA Details Next-Gen Tegra Parker SOC at Hot Chips

I haven't gone through and done a thorough AP2 COGS guess yet... but off the cuff, I'm guessing that the new hardware is about $750 to $1,000 more expensive than the old one. Note that MobileEye's ASP was something like $43 per OEM with the EyeQ3 chip. So the increase in price comes the new AP2 board itself plus 7 more cameras, of which 5 have to be wired up from far away (but 1 was already wired up), and the cost of the new board with 2 Parker SOCs. So a few hundred + wiring harnesses + cameras + moldings + labor. Of course, there's changes in the vehicle itself to support the new harnesses and the like, but that cost is minimal.

Assuming the additional hardware cost is $1,000, on 1 quarter's worth of cars @ 25,000, that's $25 million in additional COGS. Assuming a take rate of 60%, that's $75 million in additional revenue. Assuming $2,000 is recognized immediately, that's $30 million, with $45 million deferred in installments as software is rolled out in to the next year or so. Also, likely another 15% uptake rate in the future, so that's $22.5 million that will come over the next 2 years. Remember, that's per 25,000 vehicles.

Edit: altered for the base case = 1 Parker SOC, not 2.

 

[CALGARYARSENAL]

I guess I misunderstood it when you said "you have my word"... I thought that was an actual promise. But I guess I was wrong.





From your most recent post it seems clear you have no intent of following through (whether it's eating socks, shaving head, or something else), so I won't hold you to it.

I think the best resolution if the following: add me to your ignore list, and I'll add you to my ignore list. And we'll move on.

As small of a piece necessary....but sock must be eaten!

 

[techmaven]

Buried in those earnings are a $19.5 million reduction in pre-existing warranty reserves, $30 million in GHG/CAFE credit sales (mouse nuts) and the recognition of about $14 million reduction in COGS from prior quarters' gains from currency hedges. Repeatable?

Wow.. they really dropped on the warranty costs incurred. That's terrific. It means that the provision for warranty can drop.. that's repeatable. In 2015, Tesla was dealing with drive unit and battery contactor repairs and put in higher warranty liability and costs. In 2016, Model X didn't help things.

As for the accrued reduction itself, that’s a cash flow thing, not an earnings thing. The lower spend on warranty expense is an earnings thing and possibly repeatable.

The $14 million in currency hedges is repeatable if the U.S. dollar strengthens. If the dollar doesn’t strengthen then the currency hedge doesn’t increase, but then, the effect that the currency hedge is put in place to counter-act also doesn’t happen. I haven’t calculated which is better, but likely a weaker dollar would help overseas sales, so I bet Tesla would take the weaker dollar and not have the gain from the currency hedge. If the dollar weakens, then the cost of goods sold drops anyways.

I can’t find the break out of the regulatory credits for GHG/CAFE. I see the overall regulatory credits. Now, Tesla’s sales of CARB ZEV credits and other regulatory credits is repeatable, just not necessarily at each and every quarter. They still gain those credits each quarter. The worth of each credit is highly variable, but if the CARB rules remain in place while major automakers pivot away from CAFE rules and sell more gas guzzlers, then the value of each credit goes up. I think it is misleading that Tesla doesn't have a line item in the assets section for these credits as it comes as a surprise. However, given the highly variable nature of the price of these credits, likely Tesla can't put a value on them. Maybe just an accumulated amount? In any case, their balance sheet is actually better since these credits do exist and is worth something more than $0.

 

[vgrinshpun]

I hope Tesla dispatches it's VP of BD and Government Affairs, Diarmuid O'Connell, to talk to Trump's representatives that are in charge of his Infrastructure Plan. Here is what actually included in this plan:

• Pursue an “America’s Infrastructure First” policy that supports investments in transportation, clean water, a modern and reliable electricity grid, telecommunications, security infrastructure, and other pressing domestic infrastructure needs.

• Incorporate new technologies and innovations into our national transportation system such as state-of-the-art pipelines, advancements in maritime commerce, and the next generation of vehicles.

Battery Energy Storage is absolutely a large part of modernizing and making electricity grid more reliable, so incentivizing large grid connected BES projects makes a lot of sense and TE, as the only manufacturer that can provide large scale BES projects with batteries manufactured in US, could play a central role in this.

As for the next generation of vehicles, I think that Tesla could make a convincing case that BEVs certainly qualify as next generation of vehicles, and the ones which can boost US energy independence to boot. So the logical thing to do would be to expand $7,500 federal tax credit from phasing out after attainment of 200,000 cars by a specific manufacturer, to phasing out upon attainment of specific overall target, regardless of the manufacturers contributing to this target. Let's set such a target, for example to be 5 or 10% of all Light duty vehicles sold in US annually. This will encourage all manufacturers to pursue BEVs, as the ones that sell most before hitting the target will reap the largest incentives for their customers.

 

[GoTslaGo]

2 company's this close to a merger vote typically move in unison, we all know that but that wasn't being discussed???

For anyone interested, I looked into some pending mergers and their charts. There is a remarkable concordance of their price movement patterns on the monthly charts with daily periods AND their intra-day charts.

I only looked at 5 mergers and posted 3. One out of the 5 mergers does have a discordant intra-day chart pattern. All of them have similar monthly/weekly patterns.

Not sure what it means or if it disproves @Papafox's hypothesis of shorts moving TSLA/SCTY.

However, since many of the other mergers do not have much short interest it does imply that there may be a more "market maker" action going on in the background, "guiding" the merging stocks together.

Here is a link to the TA, I don't want to clutter things up here with it much more.

TSLA Technical Analysis

 

[MitchJi]

So, the actual quote is:

... you could start from one Parker SoC and that allows you to have surround camera. It allows you to use AI for highway cruising. And if you would like to have even more cameras, so that your functionality could be used more frequently in more conditions, you could always add more processors. And so we go from one to four processors. And if it's a fully autonomous driverless car, a driverless taxi, for example, you might need more than even four of our processors. You might need eight processors. You might need 12 processors.

That is not the actual quote I was referring to:
Tesla’s Autopilot chip supplier NVIDIA on new self-driving system: ‘It’s basically 5 yrs ahead and coming in 2017’
NVIDIA reported its financial results for the last quarter yesterday and surprised Wall Street.
<snip>
On a conference call with CEO Jen-Hsun Huang following the results, analysts were particularly interested in the company’s push in AI and the automotive industry, especially since Tesla’s started delivering every single one of its vehicles with NVIDIA’s Drive PX2 supercomputer.

Huang offered some very interesting insights into how he sees Tesla’s self-driving program playing out.
<snip>
He says that by introducing the necessary hardware for full autonomy now, Tesla “sent a shock wave through the automotive industry”:
<snip>
When NVIDIA and Tesla confirmed the use of Drive PX2 in the new Autopilot and Self-Driving Capable hardware suite, people started speculating about the price of the supercomputer. Huang didn’t confirm it exactly during the call yesterday but mentioned a vague “few thousand dollars”.
Which makes sense, given that Tesla added $2k to the base price for AP. It also makes both strategic and financial sense for Tesla to develop their own chip, hopefully before the M3 launches If they do that it will probably have a bigger impact on the Nvidia SP than the Tesla SP, which was the main reason for my OP.

Putting together Tesla's comments and Nvidia's comments, here are a few things that are clear to me.

1) Tesla makes their own software and it is portable between Nvidia, AMD, and Intel hardware

Basically their software will run on any hardware that has the required compiler.

2) That means Tesla isn't using much if any of the ADAS/automated driving software stack provided by other people

6) Tesla is ahead of everyone

I agree with both of those points. Sounded like zero software provided by other people. Huang said five years ahead, and IMO that might be conservative.
“And I think what Tesla has done by launching and having on the road in the very near-future here, a full autonomous driving capability using AI, that has sent a shock wave through the automotive industry. It’s basically five years ahead. Anybody who’s talking about 2021 and that’s just a non-starter anymore. And I think that that’s probably the most significant bit in the automotive industry. I just don’t – anybody who is talking about autonomous capabilities in 2020 and 2021 is at the moment re-evaluating in a very significant way.”

Huang continued by saying that autonomous driving is not a “detection problem” but an “AI problem” and he insists that it’s going to be solved in 2017.

Of course, Huang is a little biased because if Tesla manages to solve that problem in 2017 like he predicts, the first self-driving cars will be powered by his machine, the Drive PX 2, and the rest of the auto industry is likely to turn to NVIDIA for their own in-car processing power.
Elon said very adamantly "don't bet against this" (at least?) twice as safe as a human with the current sensors. He said that in the future 10x to 100x will be possible with better sensors.

 

[stolemkd]

I know that many of you are worried about the 2016 elections and the future of the Tesla. I would say that Tesla will rise regardless of the elections. Tesla Motors will Produce 500,000 Vehicles per Year - Don’t Mind the 2016 Elections !!! - CARS DISTRICT