2017 Investor Roundtable:General Discussion

[brian45011]

Straubel just gifted 22,000 shares to charity. I think that's a really positive sign that the SP is going to appreciate greatly in the near future.

The charitable deduction is valued at the average of the high and low on the date of donation. If he thought the stock was going to rise, why not wait? He has until 12/31 for it to affect the current year's taxes.

Ofc, the charity can cash in immediately, but usually that's not the case when they receive a gift in stocks. They keep it, and cash in only occasionally when they need some capital.

In the US, almost invariably charities immediately sell donated speculative equities because of their fiduciary duty. Fiduciary Duties of Foundation Directors Related to Mission Investing | Mission Investors Exchange

JB's approach of exercise and then donate some is similar to what Elon did last May. They recoup some of the funds the corporation withheld before issuing certificates when their individual tax returns are filed in the following April.

Still, I don't think it explains why he would exercise now ~12,000 options that do not expire until 2021 and later since he owns 326,301 shares and would have designated for the donation shares with a much lower basis and a holding period of over a year. Director Robyn Denholm also just exercised and sold 20,000 options that do not expire until 2021. Neither her nor JB's transactions were stated to be pursuant to a 10b5 plan

 

[techmaven]

sorry if this has been discussed already.
does anyone understand why energy storage revenue (excluding solarcity) dropped so sharply quarter over quarter?
my model has 47.2m energy storage revenue in q4 2016 and on 5.2m in q1 2017. how could that drop 90%? it's as if they had no completed installations/revenue recognition in the first quarter of 17? from the 10q below.

Energy generation and storage revenue includes sales of solar energy systems and energy storage products, leasing revenue from solar energy systems under operating leases and power purchase agreements and sales of solar energy system incentives. Energy generation and storage revenue increased by $191.2 million, or 841%, in the three months ended March 31, 2017 as compared to the three months ended March 31, 2016. This was primarily due to the inclusion of revenue from SolarCity, which we acquired on November 21, 2016, of $208.7 million, partially offset by a decrease in energy storage revenue of $17.4 million.​

Well...the Kauai project did not have any revenue recognition yet. It was PowerPack 2 with Panasonic/Japan cells that shipped in Q4. I think the PowerPack 2's that were made for Southern California Edison did make it into Q4. Those are the two largest single projects recently. Further, the Gigafactory came online in January and apparently PowerWall 2's were delayed by a rush of PowerPack 2 production early in the quarter. I suspect almost zero, if not actually zero PowerWall 2's were in Q1. And PowerPack 2's shipping in early February through March may not have had any revenue recognition yet.

 

[luvb2b]

so kauai gets recognized in q2 17? are other large projects that are expected to be recognized q2 17?

i'm really having a hard time wrapping my head around this 5.2m revenue for energy storage in q1 17. that's half of the q1 2016 figure!

i would have expected something closer to the 50m that was recognized in q4 2016.

Well...the Kauai project did not have any revenue recognition yet. It was PowerPack 2 with Panasonic/Japan cells that shipped in Q4. I think the PowerPack 2's that were made for Southern California Edison did make it into Q4. Those are the two largest single projects recently. Further, the Gigafactory came online in January and apparently PowerWall 2's were delayed by a rush of PowerPack 2 production early in the quarter. I suspect almost zero, if not actually zero PowerWall 2's were in Q1. And PowerPack 2's shipping in early February through March may not have had any revenue recognition yet.

 

[techmaven]

I don't know if anyone else noted this, but the exchange rates have been better for Tesla lately. The USD is weaker against the Euro by 4.2% since end of Q1, and almost 6% YTD. The battery cell costs from Japan is basically flat, as the gains the USD made against the yen have dropped back to break even. It is down 5% YTD. The USD/CNY chart looks pretty even from Q1, and 1% better from end of Q4.

A weaker dollar is good for Tesla overall, even though the cells are, today, priced in yen.

 

[techmaven]

so kauai gets recognized in q2 17? are other large projects that are expected to be recognized q2 17?

i'm really having a hard time wrapping my head around this 5.2m revenue for energy storage in q1 17. that's half of the q1 2016 figure!

i would have expected something closer to the 50m that was recognized in q4 2016.

Kauai was a Solarcity PPA. I suspect they can't sell off those future revenues while the project wasn't considered "done." As of the last shareholder's letter:

"Final testing on the KIUC project will be completed very shortly, triggering the start of revenue recognition."

I would expect this to be part of the next batch of revenue to be sold off. For PPA's, it appears, it can be 3 quarters between production and revenue recognition on the bigger projects. Obviously, some of these projects can be completed intra-quarter. I think the SCE Mira Loma project was rare in that it was essentially intra-quarter which inflated the Q4 numbers. As the stationary storage business ramps, I think we will see a lot of lumpiness as a result.

 

[esk8mw]

The charitable deduction is valued at the average of the high and low on the date of donation. If he thought the stock was going to rise, why not wait? He has until 12/31 for it to affect the current year's taxes.



In the US, almost invariably charities immediately sell donated speculative equities because of their fiduciary duty. Fiduciary Duties of Foundation Directors Related to Mission Investing | Mission Investors Exchange

JB's approach of exercise and then donate some is similar to what Elon did last May. They recoup some of the funds the corporation withheld before issuing certificates when their individual tax returns are filed in the following April.

Still, I don't think it explains why he would exercise now ~12,000 options that do not expire until 2021 and later since he owns 326,301 shares and would have designated for the donation shares with a much lower basis and a holding period of over a year. Director Robyn Denholm also just exercised and sold 20,000 options that do not expire until 2021. Neither her nor JB's transactions were stated to be pursuant to a 10b5 plan

Executives tend to exercise options/sell stock in single larger swaths as opposed to a drip drip drip over time. There's a few reasons for it:

Executives are encouraged/pressured to minimize the overall number of reportable transactions. Investors truly do not understand Form 4s at all and essentially every action is viewed negatively. Thus, better to do it all at once where possible. The expiring options were the trigger forcing him to do a transaction this time and the others were probably just added in.

Executives tend to not sell shares unless they have a defined need - real estate purchase or other major investment are the 2 most common reasons I see. Note that selling shares to cover taxes/exercise price should never be viewed negatively. 90%+ of transactions are handled this way - most do not have the liquidity to pay cash, especially for options.

Risk of filing errors. Form 4s are a huge burden. They have to be prepared and filed within 2 business days of each and every transaction. Many companies use outside counsel and other services to prepare these and it's expensive. The costs of failure are embarrassing "we screwed up" filings and a mea culpa in the annual proxy statement. Not good. Better to minimize the Form 4s. And again, each Form 4 invites fresh scrutiny from investors who have no idea what they are talking about or looking at.

I know right now you are thinking "but these minor concerns wouldn't influence behavior when there's potentially millions being left on the table by selling too early if they are truly bullish! They must be trying to get out at the top!" You are wrong. These institutional biases are strong and most definitely influence behavior. Add on to that the huge desire to avoid the appearance of trying to "time the market" by peers or investors and you get executives acting against their financial interests all the time. It's a cost of doing business in the top levels of a public company.

For those who don't already know, I do this stuff for a living.

 

[luvb2b]

i'm estimating global auto-pilot take rate around 48% using data from the 10q. comments on the methodology appreciated:

10q mentions 81.1m of ap2 revenue recognized in q1 17 with no mention of deferrals. assuming that's all the ap 2.0 revenue, and assuming 9k vehicles sold in q4 2016 w/ap2 hardware and 24.5k vehicles sold in q1 2017 w/ap2 hardware:

then if take rate = r:
33,500 vehicles x r x $5000 = $81,100,000
r = 48.42%

that gives approx. 1 time recognition in q1 2017 as:
9,000 vehicles x 0.4842 x 5000 = 21.8 m revenue of ap2 vehicles from q4 16 recognized in q1 17.

and 81.1m - 21.8m = 59.3m = ap2 revenue from vehicles sold in q1 2017.

 

[esk8mw]

i'm estimating global auto-pilot take rate around 48% using data from the 10q. comments on the methodology appreciated:

10q mentions 81.1m of ap2 revenue recognized in q1 17 with no mention of deferrals. assuming that's all the ap 2.0 revenue, and assuming 9k vehicles sold in q4 2016 w/ap2 hardware and 24.5k vehicles sold in q1 2017 w/ap2 hardware:

then if take rate = r:
33,500 vehicles x r x $5000 = $81,100,000
r = 48.42%

that gives approx. 1 time recognition in q1 2017 as:
9,000 vehicles x 0.4842 x 5000 = 21.8 m revenue of ap2 vehicles from q4 16 recognized in q1 17.

and 81.1m - 21.8m = 59.3m = ap2 revenue from vehicles sold in q1 2017.

I don't have data for you, but I would be absolutely shocked if the AP (not FSD) take rate is less than 75%.

 

[techmaven]

i'm estimating global auto-pilot take rate around 48% using data from the 10q. comments on the methodology appreciated:

10q mentions 81.1m of ap2 revenue recognized in q1 17 with no mention of deferrals. assuming that's all the ap 2.0 revenue, and assuming 9k vehicles sold in q4 2016 w/ap2 hardware and 24.5k vehicles sold in q1 2017 w/ap2 hardware:

then if take rate = r:
33,500 vehicles x r x $5000 = $81,100,000
r = 48.42%

that gives approx. 1 time recognition in q1 2017 as:
9,000 vehicles x 0.4842 x 5000 = 21.8 m revenue of ap2 vehicles from q4 16 recognized in q1 17.

and 81.1m - 21.8m = 59.3m = ap2 revenue from vehicles sold in q1 2017.

I believe you are calculating 100% recognition of EAP option revenue and I suspect that isn't the case. It would make more sense that they've only recognized about 50% of the EAP option revenue, or possibly even less. They have to hold back some for future upgrades. I would expect that EAP has one more big chunk of revenue to be recognized and the rest would be recognized according to the given schedule for other software upgrades.

 

[luvb2b]

i thought so too, but there's no mention in the 10q of any revenue left unrecognized from ap2. if anyone has any disclosure to indicate some ap2 revenue was left on the table i'm most appreciative of sharing that source.

50% couldn't be correct as a fraction of revenue recognized anyway as it would imply a 99% global take rate (clearly too high).

I believe you are calculating 100% recognition of EAP option revenue and I suspect that isn't the case. It would make more sense that they've only recognized about 50% of the EAP option revenue.

 

[Oil4AsphaultOnly]

Yeah, that's been predebunked. First of all, subway tunnels are about the same diameter as the tunnels he proposes. Second, there'll be traffic congestion at the entrances and exits. Third, it takes 20 car tunnels to have the people-moving capacity of 1 subway tunnel. I wish him best of luck with the boring technology, but the idea of layers of car tunnels is actively stupid.

Not sure how the first two issues will be addressed, but I think we can look to the hard disk interface changes of the early 2000's from IDE to SATA for an analogy to addressing the 3rd issue.

It might be possible to send more data (aka passengers) down a cable (tunnel) if you increase the signaling rate (speed) and latency (wait time between "cars"). Yes a subway train might be able to carry 20 cars (80 people?) worth of passengers at a time. But with 5 minute gaps between every train @ 30mph (NY subway average speeds), that's 80 people distributed over every 2.5 miles of tunnel. Assuming each station is 2.5 miles apart would yield 80 people moved every 5 minutes every 2.5 miles.

At 120mph (simplifies calculations), with a 10 second gap between sleds (that's an 880 feet gap, ~1/6th of a mile), you can deliver 30 cars worth (a 15 second gap, ~1/4 mile, would yield 20 cars worth) of people over 10 miles. Your data rate is higher in this scenario, with better addressability (dropping people off where they want to be) too!

 

[luvb2b]

here's a typical disclosure in the 10q.

Gross margin for total automotive increased from 22.0% to 24.4% in the three months ended March 31, 2017 as compared to the three months ended March 31, 2016. This was primarily due to lower material and manufacturing costs as we further improved our vehicle production processes as well as favorable product mix shift and higher option uptake. Additionally, we recognized Autopilot 2.0 revenue during the three months ended March 31, 2017.​

I believe you are calculating 100% recognition of EAP option revenue and I suspect that isn't the case. It would make more sense that they've only recognized about 50% of the EAP option revenue, or possibly even less. They have to hold back some for future upgrades. I would expect that EAP has one more big chunk of revenue to be recognized and the rest would be recognized according to the given schedule for other software upgrades.

 

[techmaven]

here's a typical disclosure in the 10q.

Gross margin for total automotive increased from 22.0% to 24.4% in the three months ended March 31, 2017 as compared to the three months ended March 31, 2016. This was primarily due to lower material and manufacturing costs as we further improved our vehicle production processes as well as favorable product mix shift and higher option uptake. Additionally, we recognized Autopilot 2.0 revenue during the three months ended March 31, 2017.​

Strange that in the most recent 10-Q, there isn't this from the most recent 10-K:

Summary of Significant Accounting Policies



Revenue Recognition

We recognize revenue for products and services when: (i) a persuasive evidence of an arrangement exists; (ii) delivery has occurred and there are no uncertainties regarding customer acceptance; (iii) pricing or fees are fixed or determinable; and (iv) collection is reasonably assured.

Automotive Revenue

Automotive revenue includes revenues related to deliveries of new vehicles, sales of regulatory credits to other automotive manufacturers, and specific other elements that meet the definition of a deliverable under multiple-element accounting guidance including free access to our Supercharger network, free internet connectivity, and future free over-the-air software updates. These deliverables are valued on a stand-alone basis and we recognize their revenue over our performance period, which is generally the eight-year life of the vehicle, except for internet connectivity which is over the free four year period. If we sell a deliverable separately, we use that pricing to determine its fair value; otherwise, we use third party pricing of similar options, our best estimated selling price by considering costs used to develop and deliver the service, and other information which may be available.

As of December 31, 2016, and 2015 we had deferred $291.2 million and $138.2 million related to the purchase of vehicle maintenance and service plans, access to our Supercharger network, internet connectivity, autopilot and over-the-air software updates.

At the time of revenue recognition, we record a reserve against revenue for estimated future product returns. Such estimates are based on historical experience and are immaterial in all periods presented.

I still think that they have to hold to the practice of recognizing autopilot revenue over time...

 

[Fallenone]

i'm estimating global auto-pilot take rate around 48% using data from the 10q. comments on the methodology appreciated:

10q mentions 81.1m of ap2 revenue recognized in q1 17 with no mention of deferrals. assuming that's all the ap 2.0 revenue, and assuming 9k vehicles sold in q4 2016 w/ap2 hardware and 24.5k vehicles sold in q1 2017 w/ap2 hardware:

then if take rate = r:
33,500 vehicles x r x $5000 = $81,100,000
r = 48.42%

that gives approx. 1 time recognition in q1 2017 as:
9,000 vehicles x 0.4842 x 5000 = 21.8 m revenue of ap2 vehicles from q4 16 recognized in q1 17.

and 81.1m - 21.8m = 59.3m = ap2 revenue from vehicles sold in q1 2017.

They specifically said carried over Q4 2016 EAP revenue was 35m in the Q1 update letter.

 

[Buddyroe]

Time for our morning M3 eye candy. What a beautiful car.

Tesla Model 3: new high-res shots of the latest silver release candidate spotted

If that's THE car, I gotta buy out someone's spot!

 

[techmaven]

assuming 9k vehicles sold in q4 2016 w/ap2 hardware

Basically no vehicles produced in Q4 had AP1 hardware, so the only cars sold with AP1 in Q4 are the cars leftover from before Q4. Assume 9,000 or so, 5,500 vehicles in transit, another 2,500 demo vehicles were AP1. Out of 22,252 delivered in Q4, I think there were more like 13,000 AP2 vehicles.

 

[brian45011]

sorry if this has been discussed already.
does anyone understand why energy storage revenue (excluding solarcity) dropped so sharply quarter over quarter?
my model has 47.2m energy storage revenue in q4 2016 and on 5.2m in q1 2017. how could that drop 90%? it's as if they had no completed installations/revenue recognition in the first quarter of 17? from the 10q below.

Energy generation and storage revenue includes sales of solar energy systems and energy storage products, leasing revenue from solar energy systems under operating leases and power purchase agreements and sales of solar energy system incentives. Energy generation and storage revenue increased by $191.2 million, or 841%, in the three months ended March 31, 2017 as compared to the three months ended March 31, 2016. This was primarily due to the inclusion of revenue from SolarCity, which we acquired on November 21, 2016, of $208.7 million, partially offset by a decrease in energy storage revenue of $17.4 million.​

SCE's Mira Loma installation (80 MWh) and American Samoa (6 MWh) were included in 4Q16. KIUC (52 MWh) was not completed in time to be included in 1Q17. The other big installation, 6 MWh for CREEC, was included in 3Q16

IIRC, SCE and American Samoa were outright purchases: CMEEC and KIUC are 20 year PPAs. The CMEEC involved Brightfields Development LLC so it probably is already in a Tax Equity SPE and contributes to the NCI losses.

The KIUC situation is more of a black box. I don't fully understand the nuances among the various alternatives for Tax Equity structures. Tax Equity 101: Structures

I think there must be investor ownership in the project when it is placed in service for a Partnership Flip structure to afford the investor the ITC and accelerated depreciation. Investor ownership on the commissioning date may not be required for a Sale-Leaseback structure. I do not think an Inverted Lease Structure would be used for KIUC because TSLA/SCTY couldn't use any of the tax benefits considering their huge tax-loss carry-forwards.

SCTY's last 10k stated:

"Generally, significant loss allocations to the investors have arisen in situations where there was a significant difference between the fair value and the cost of the assets sold to the joint venture financing funds in a particular period accompanied by the absence of guarantees of minimum returns to the investors by us, since the capital contributions by the investors were based on the fair value of the assets while the calculation is based on the cost of the assets."

A bit ambiguous, but can be "interpreted" that a large NCI loss could be reported in 2Q17:
i.) if the KIUC project is transferred to a tax equity investor during 2Q17, and
ii.) the "FMV" paid by the tax equity investor is significantly higher than TSLA/SCTY's capital cost of the project.

TE's revenue depends to some extent on how quickly Advanced Micro-grid Solutions/Macquarie Capital draw PowerPacks against their commitment.

 

[luvb2b]

They specifically said carried over Q4 2016 EAP revenue was 35m in the Q1 update letter.

thanks for that. ok so the split for ap 2 revenue:

q4 16: 35m / 5000 = 7,000 vehicles
q1 17: 46.1m / 5000 = 9,220 vehicles

regardless of what number i use for q4 ap2 vehicles (9000 or 13000 as suggested by @techmaven) i get a > 50% take rate in q4 16.
and the take rate in q1 17 drops to 37% (9220 / 25011 deliveries). how does that happen?

i'm all ears trying to understand this.

 

[T3slaTulips]

thanks for that. ok so the split for ap 2 revenue:

q4 16: 35m / 5000 = 7,000 vehicles
q1 17: 46.1m / 5000 = 9,220 vehicles

regardless of what number i use for q4 ap2 vehicles (9000 or 13000 as suggested by @techmaven) i get a > 50% take rate in q4 16.
and the take rate in q1 17 drops to 37% (9220 / 25011 deliveries). how does that happen?

i'm all ears trying to understand this.

Did it say that they recognized 100% of the EAP revenue? They've not achieved feature parity with AP1 yet, nor 4 camera AP, so my guess is that the percentage of revenue recognized is still less than 100%. (But this is a very wild-assed guess based only on the way I would do it, and not about how it is done or the way it should be done).

 

[schonelucht]

i would have expected something closer to the 50m that was recognized in q4 2016.

The energy storage business was seriously overhyped. Tesla didn't have a convincing product and the market wasn't ready to buy. Simple as that. Now with version 2 of the Powerpack and Powerwall, Tesla has a chance to redeem itself. But they got off a bad start in 2017Q1 with a supply problem. Gigafactory output is/was low enough that the Hawaii order pushed back residential Powerwall deliveries because they couldn't fulfil both at the same time. I expect Q2 to be better but still not anything that would resemble a serious level of output.