2017 Investor Roundtable:General Discussion

[ggr]

and the advantage is getting bigger as Elon launches more satellites for networking. GM and I'm sure others can negotiate deals with cell carriers to add data connectivity to their cars, but will have to do this per region/country, Tesla's network will work everywhere

If only that were true... I have some experience in this area through my previous employer (Qualcomm/Globalstar) and there are significant regulatory hurdles to get this kind of network operational in a surprising number of countries. Basically they want the traffic to go through their own ground station where they can legally control it or illegally intercept it. The various newcomers (OneWeb, SpaceX's satellites, Intelsat's new constellation, Iridium Next) expect to be available and profitable in enough places to put pressure on the others to also open up, but that's a long term plan. It won't "work everywhere" on day one.

(You know that Iridium got its name from the number of satellites in the original plan... for some unknown reason they wouldn't change the name to Dysprosium when it downsized though. )

 

[Reciprocity]

We’re Tracking Every Project Elon Musk Has Dreamed Up.

Elon Musk goal tracker on Bloomberg.

37 goals tracked
11 completed
2 late.

Interesting and quite fair.

What I think is important is to look at the 11 completed. I dont care that they where late. Look at the goals completed with SpaceX and Tesla. Its easy to get mad when AP2 does not yet exceed AP1, until you realize how infinitely small it is in the big picture of what has been accomplished in such a short period.

 

[Waiting4M3]

Something is wrong with this chart - just look at the yearly sales in your link. Sales went down from about 140k during the 2008 crisis, and started to recover by 2013, reaching the 140k plateau in 2014 and 2015, and then went down, as I mentioned, by 25% in 2016

I plotted using the monthly sale #, because I wanted to see if there is a drop off after March 2016:

For example, Jan 2017 I show a 16.7% increase YoY, from 5119 in Jan 2016 to 5976 in 2017.

 

[ggr]

I plotted using the monthly sale #, because I wanted to see if there is a drop off after March 2016:
View attachment 233142

For example, Jan 2017 I show a 16.7% increase YoY, from 5119 in Jan 2016 to 5976 in 2017.

I'd call that a pretty clear drop-off overall, though. In the first 2 columns there were only 5/24 months below 10,000 cars. In the last 13 months (since Model 3 reservations opened) only 3 months managed to top 10,000.

The irony of the announcement, to me, is that BMW had a 3-series electric car, the Active-E, and did the usual thing, lease-only, recall, crush.

 

[mutle]

The irony of the announcement, to me, is that BMW had a 3-series electric car, the Active-E, and did the usual thing, lease-only, recall, crush.

ActiveE was based on the 1 series.

 

[Waiting4M3]

I'd call that a pretty clear drop-off overall, though. In the first 2 columns there were only 5/24 months below 10,000 cars. In the last 13 months (since Model 3 reservations opened) only 3 months managed to top 10,000.

The irony of the announcement, to me, is that BMW had a 3-series electric car, the Active-E, and did the usual thing, lease-only, recall, crush.

In my earlier plot based on this data, the drop off started happening in late 2015. I suspect even before M3, maybe the MS/MX were already pulling 3-series buyers up.

 

[ggr]

ActiveE was based on the 1 series.

Oh, OK... looked like a 3 to me. I guess they all look alike ;-). A neighbor had one for a while.

 

[Reciprocity]

In my earlier plot based on this data, the drop off started happening in late 2015. I suspect even before M3, maybe the MS/MX were already pulling 3-series buyers up.

Anyway to pull this same data for A4 and C-Class?

 

[brian45011]

Unless you are aware that she is not, I suspect she is. Which, if true, does not bode well for Boston Energy consortium *plans*...

Why so? Ms Palaszczuk has said "“What I need to do now is talk to companies. I know there are a number that are interested in Queensland, and we need to work out who will be the best fit,” No Cookies | The Courier Mail

She did not go to Nevada merely to visit Tesla's facility:
"Premier Annastacia Palaszczuk and Governor of Nevada Brian Sandoval signed [in August 2016] the Statement of Intent which acknowledges an ongoing commitment to work collaboratively in sectors of significance to both the Queensland and Nevada economies."
Queensland and Nevada sign landmark agreement - TIQ

The consortium appears to have ambitious plans with ready access to investment capital :

"Most of the same partners have also unveiled plans for a similar sized battery storage “gigafactory” in New York state, and announced on Monday that this would be located at the Huron campus, the home of IBM (pictured above)...BEI will assist with project structuring, capital raising and global expansion, and Moss says these plants will be the “first of a series of proposed lithium-ion battery manufacturing hubs globally.” reneweconomy.com.au/boston-energy-consortium-advances-plans-for-queensland-battery-gigafactory-63187/

Does Tesla have an Australian "partner" or connection other than Mike Cannon-Brooke, the founder of software firm Atlassian, for the South Australia and Queensland opportunities?

 

[McRat]

which country does your GM's 4G LTE work in? how does it work globally?

All my cars and motorcycles are in the USA.

The UK Camaros and Corvettes come with 4G LTE in all trim levels. You'd have to check other countries.

Note: One reason GM left many countries? 60% more for Chevolets there than the US. Some countries are 100% over MSRP.

 

[erthquake]

Is anyone able to login into their My Tesla account? I keep getting "Username must be valid" messages with my login.

Edit: Nevermind. I had to reset my password and I can login again. Weird.

 

[McRat]

The speeds should be in the gigabit range and the latency a mere 25ms. The reliability should be greater as well, since they will be at a lower orbit than traditional geosynchronous orbit satellites that provide that dirt slow service we're used to out here in the boonies.

With latency as low as 25ms, SpaceX to launch broadband satellites in 2019

OK, that's impressive. Hughes sold me 256k avg (advertised 1mb) with >250ms latency for $199/m and had problems in weather IIRC.

 

[Drumheller]

OK, that's impressive. Hughes sold me 256k avg (advertised 1mb) with >250ms latency for $199/m and had problems in weather IIRC.

That.... that's just robbery. Are you living in Antarctica?

 

[McRat]

That.... that's just robbery. Are you living in Antarctica?

Charge what the market will bear. We could not even get DSL. 56k dial up.

 

[scaesare]

OK, that's impressive. Hughes sold me 256k avg (advertised 1mb) with >250ms latency for $199/m and had problems in weather IIRC.

I used to work in the sat-based IP delivery business... and I have one site I'm still affiliated with where Hughes is the best we can get. Your experience is not atypical... which is to say "it sucks", but it's better than nothing.

Latency is indeed the big issue. While there are a number of optimizations and tricks (primarily to avoid small round trip IP exchanges, and/or to batch traffic, DNS queries handled with local proxies, etc...), there's no getting around the latency associated with a geo-sync satellite. It's on the order of ~250ms 1-way... so I suspect your advertised speed is some sort of average based on the local experience.

The LEO birds should do significantly better in this regard. It still remains to see how well they deal with weather, however...

 

[MitchJi]

I've explained before that Tesla Network is basically not happening at all, with the possible exception of a small number of carefully selected cities. The tech won't be ready outside specified geofenced areas for a long time...

That's the last time you get an explanation of this; it's boring arguing with technoutopians who think self-driving cars will take over the world, because they're completely brainless on the topic, unwilling to listen to any evidence whatsoever.

I believe that your extreme skepticism is unwarranted. You are the equivalent of a layman arguing against climate change. Every expert disagrees with your opinion.

Of course you could be correct. OTOH every time Elon states that they are on schedule for the end of year demonstrator the odds improve. And Elon does have a record of accomplishing the impossible (landing rockets on floating barges for example).

 

[ValueAnalyst]

Uber's margin in San Francisco:

City Contribution Margins
San Francisco Today

Gross Bookings 100%
Driver Commission (78.1)%
Safe Rides Fee 5.0%
Driver Incentives (1.4%)
Returns (0.4%)
Net Revenue 25.1%

Payment Fees (2.4)%
Mobile (Net) (0.2)%
Network (0.3)%
Other COGS (1.9)%
Sales & Marketing (2.9)%
Insurance (3.5)%
Other Opex (3.8)%

Contribution Margin 10.1%

Given that 80% goes to driver, it makes sense that Tesla Network would command 90% gross margin wIth steadily declining average fare as more offerings come online. Also, note that Tesla doesn't need to do any marketing and will be providing its own insurance (6.5% total for Uber). Finally, I would expect a rider's experience to be better in a Tesla than in an ICE/UBER combo where no competitor controls both the hardware and the software as Tesla does.

Nevertheless, I reduced my assumption to 80%, and I apply a level of r&d and sg&a after.

 

[McRat]

I believe that your extreme skepticism is unwarranted. You are the equivalent of a layman arguing against climate change. Every expert disagrees with your opinion.

Of course you might be correct. OTOH every time Elon states that they are on schedule for the end of year demonstrator the odds improve. And Elon does have a record of accomplishing the impossible (landing rockets on floating barges for example).

You are saying the Stock Market is Science? OK, where I can buy a block of room temperature fusion or anti-gravity?

 

[racer26]

I used to work in the sat-based IP delivery business... and I have one site I'm still affiliated with where Hughes is the best we can get. Your experience is not atypical... which is to say "it sucks", but it's better than nothing.

Latency is indeed the big issue. While there are a number of optimizations and tricks (primarily to avoid small round trip IP exchanges, and/or to batch traffic, DNS queries handled with local proxies, etc...), there's no getting around the latency associated with a geo-sync satellite. It's on the order of ~250ms 1-way... so I suspect your advertised speed is some sort of average based on the local experience.

The LEO birds should do significantly better in this regard. It still remains to see how well they deal with weather, however...

Yep, can't get around the limitation of the speed of light. Geosync orbit is 35,786km - 120ms at the speed of light, about 250ms round trip.

LEO birds will be ~36x closer. Should mean ping times more down in the 7ms range.

I expect they'll do fine with weather - the problem with a geosync bird is you're firing a directional signal at a single thing a few tens of metres across, 35,786km away. Its a fixed point on the sky, and so any weather obstructing line of sight will tend to scatter the signal. When aiming at a target 36x closer, the scatter will matter less. The LEO birds will also be a multi-point target, much more like GPS than a traditional Geosync comsat. GPS service is sometimes degraded in bad weather (because its more difficult to find a good signal from >4 birds at the same time) but rarely does weather impact it to the point of unusability.

 

[scaesare]

Yep, can't get around the limitation of the speed of light. Geosync orbit is 35,786km - 120ms at the speed of light, about 250ms round trip.

LEO birds will be ~36x closer. Should mean ping times more down in the 7ms range.

I expect they'll do fine with weather - the problem with a geosync bird is you're firing a directional signal at a single thing a few tens of metres across, 35,786km away. Its a fixed point on the sky, and so any weather obstructing line of sight will tend to scatter the signal. When aiming at a target 36x closer, the scatter will matter less. The LEO birds will also be a multi-point target, much more like GPS than a traditional Geosync comsat. GPS service is sometimes degraded in bad weather (because its more difficult to find a good signal from >4 birds at the same time) but rarely does weather impact it to the point of unusability.

Yup yup.

I'll point out that "round trip" for satellite folks means up to the bird and back down to the ground station. For computer folks that means to your destination and back.

So from an IP standpoint, a single packet round trip = two satellite round trips for a total of ~500ms packet latency. In reality with relay processing and ground station overhead ping times of 700+ ms weren't uncommon... although that's been some years ago...