2017 Investor Roundtable:General Discussion

[scaesare]

Usually automated storage and retrieval is a sign of misplaced priorities in manufacturing, as storage and retrieval add capital and operating costs (and more importantly cycle time to the inventory, and lags in cause and effect as downstream processes are often the place where flaws are found- they tie up money and effort and delay the discovery of flaws creating a junk exposure/risk) with no actual value added to the product.

So in trying to sort this, if an IC guy designed this thinking that buffers are good, there are differences between data and physical product, but Tesla is pretty smart, and they know that inventory turns matter, so this would be bad sorting on my part.

If color changeovers are so expensive that this facility is the only way to get more than one color on a truck, that might make sense.

I am from the school of low rate parallel machines that don't need a buffer in case they break down. Shigeo Shingo school I guess.

If it is to store finished product to avoid crossing the threshold, because the natural area of distribution is large, it is a natural capacitor, I would tend to rent space in mall garages, on a temporary basis. Or even rail cars or trucks.

If there is some sort of soak time required in the manufacturing process - reforming or something like that, it could make sense. Maybe they are moving a battery process step to occur in vehicle?

I see a pile of money, delay and misdirected effort in these types of buildings. Sometimes to serve a fragile high rate process with slow changeover time. If the inventory money and effort and lag costs were directed toward buying more manufacturing equipment or improving changeover speeds, ... usually it is better not to have these sorts of buildings.

Pretty sure I am missing something.

[Edit: AWK mind bending stream processing with no memory management and Shigeo Shingo single minute exchange of die( and other stuff) are the two things that set my expectation of high speed and low cost to a point where others think "impossible unrealistic." Appreciating these things creates conflict with those who don't...]

I wonder if the fact that Tesla is rather highly vertically integrated, whereas the plant when it was originally designed for NUMMI may have been laid out for more of a just-in-time manufacturing process plays in to this.

It seems to me that your "parts storage" for a JIT facility is the pipeline from your suppliers (their warehouses, ships, rail lines, trucks, etc...) to you.

For a vertically integrated environment, you are your own supplier for many of your parts, and thus you need some buffer space to accommodate those parts.

 

[elasalle]

Tesla’s manufacturing plant in Fremont is expanding, as an automated storage and retrieval facility is seen rising from the north end of the factory on Aug. 8, 2017. (Photo by Joseph Geha/The Argus)

Already, a new automated storage facility is rising quickly and two other additions are in the works, according to city officials.

The Fremont City Council approved Tesla’s 4.6 million-square-foot expansion master plan in December, which outlines 12 primary growth zones around the factory.

When complete, a new automated storage and retrieval facility rising from the north end of the Fremont Tesla factory will look like this rendering provided by the company. A similar facility will be built on the south end of the factory site as the company expands its footprint in Fremont. (Photo courtesy city of Fremont)

If fully built out according to plan, the plant would almost double its footprint in the city, helping the company reach its ambitious production goals of manufacturing electric vehicles for the mass market.

Fremont Tesla facility expansion underway

Wild Guess - but any chance this facility is multi-purpose and also solves the parking issue?

 

[mongo]

Mongo, you likely know this.

In changeover, setting the closing height is the long time constant: too open and you don't form the part and too closed and you break something that takes a long time to replace. On smaller presses you build up every die you have, no matter how deep the draw, to have the same closing height.

Is this standard procedure with the big stuff, as "that adds a lot of steel to the shallow dies" if the deepest draw is significant?

Warning: I am not a stamping engineer.

I think what you are describing is the result of a standard mechanical press having a fixed displacement based on the eccentric cam that drives it. If the die interferes, the entire mechanism jams and you suddenly have a whole lot of inertial trying to break things. Akin to hydrolocking an ICE.

The Schuler presses are hydraulic which provides much higher adjustability and control over the stamping process compared to a standard fixed stroke mechanical press. The maximum force is also limited to the hydraulic fluid pressure, even in a premature bottoming case. From their literature https://www.schulergroup.com/major/...s_broschuere_hydraulische_pressenlinien_e.pdf
It seems like switch over is fairly fast. The lead press also has a hydraulic bed which monitors the forces as the die reaches full stroke.

 

[22522]

Warning: I am not a stamping engineer.

I think what you are describing is the result of a standard mechanical press having a fixed displacement based on the eccentric cam that drives it. If the die interferes, the entire mechanism jams and you suddenly have a whole lot of inertial trying to break things. Akin to hydrolocking an ICE.

The Schuler presses are hydraulic which provides much higher adjustability and control over the stamping process compared to a standard fixed stroke mechanical press. The maximum force is also limited to the hydraulic fluid pressure, even in a premature bottoming case. From their literature https://www.schulergroup.com/major/...s_broschuere_hydraulische_pressenlinien_e.pdf
It seems like switch over is fairly fast. The lead press also has a hydraulic bed which monitors the forces as the die reaches full stroke.

Thank you. As long as changeover takes less than 10 minutes, you should not be getting behavior distortions to compensate.

 

[McRat]

GM Prices Senior Unsecured Notes to Fund Opel Pension Obligations - GM Inside News

Tesla $1.5B to fund the future.

GM $3.0B to close the books on its European past.

Not sure why folk insist on comparing General Motors to Tesla Motors. They could not be more different. They do not operate on the same business model. It does not matter they are both in the auto industry when it comes to financials. They are businesses. You might as well compare TSLA to XOM (Exxon). They are that different.

One is union, one is not (yet). This is important due to pension funding.
One has stock valued on dividends and profits, one on speculation alone.
One has $36 billion in land, buildings, and equipment, 70% of their market cap. Tesla holds $8 billion or 15% of market cap.
One made >$9 billion profit last year, and paid record profit sharing to union workers. The other? No profit.

So exactly how do those two business models line up when it comes to bonds? One issues bonds based on financial ability to service the debt today, the other would issue bonds that require the business to change significantly before they could be paid.

This is not a hit piece on TSLA, it is simply facts when it comes to unsecured debt.

One wants an unsecured loan and has a house and works hourly at a factory for the last 10 years.
One wants an unsecured loan but rents and is a promising intern heading towards being a plastic surgeon in Miami.

 

[ValueAnalyst]

Not sure why folk insist on comparing General Motors to Tesla Motors. They could not be more different. They do not operate on the same business model. It does not matter they are both in the auto industry when it comes to financials. They are businesses. You might as well compare TSLA to XOM (Exxon). They are that different.

One is union, one is not (yet). This is important due to pension funding.
One has stock valued on dividends and profits, one on speculation alone.
One has $36 billion in land, buildings, and equipment, 70% of their market cap. Tesla holds $0.6 billion or 1% of market cap.
One made >$9 billion profit last year, and paid record profit sharing to union workers. The other? No profit.

So exactly how do those two business models line up when it comes to bonds? One issues bonds based on financial ability to service the debt today, the other would issue bonds that require the business to change significantly before they could be paid.

This is not a hit piece on TSLA, it is simply facts when it comes to unsecured debt.

One wants an unsecured loan and has a house and works hourly at a factory for the last 10 years.
One wants an unsecured loan but rents and is a promising intern heading towards being a plastic surgeon in Miami.

I think the bond markets (past and present) probably see these two companies closer than equity markets (present and future) do. The fact is both companies have manufactured cars in the past, even though their futures are very different.

Also, Tesla's PP&E was $8.4 billion as of June 30, not $0.6 billion. What am I missing in what you said?

 

[brian45011]

ValueAnalyst said: ↑
Does anyone know why gigafactory is not being used as collateral? Why not back it with the asset and get a significantly lower rate? I don't see much of a default risk (clearly the bond market doesn't either) but even if Tesla defaults on this debt, it's not like gigafactory would be of much use anyway.

Leave their options open to leverage it in the future?

The ABL Credit Agreement through amendment 5 contained an exclusion of the GF in the definition of "Collateral" that could be attached to secure repayment of draws under the ABL. When the maximum commitment was increased June 16, 2017 from $1.2 billion to $1.825 billion that exclusion was deleted. Assuming the ABL creditors have made the proper filings under Article 9 of the UCC, the unsecured note would be behind the ABL creditors in priority to GF assets.

Tesla's investments in the GF have been reported as:


" In 2016, we used cash of $455.3 million towards Gigafactory 1 construction and expect to spend a total of approximately $770.0 million during 2017."

"We had cumulatively incurred and capitalized costs of $1.98 billion and $825.3 million, respectively, for Gigafactory 1 as of June 30, 2017 and December 31, 2016."

It appears $1.155 billion was expended in the 1st half of 2017 against the total of $0.77 billion expected for the year.

The primary reason the GF is of minimal value as collateral for debt is that the Panasonic lease grants the lessee rights of Quiet Enjoyment.

Some have asserted the bond solicitation is not really a recent decision to provide a contingency cushion against force majeure events, but something that has been planned for awhile.

 

[ggr]

Does anyone on TMC have knowledge or likely indications on what options they give to M3 reservation holders whose number comes up and are contacted to configure, but who wish to wait until some option like dual motor AWD becomes available?
I would hope these reservation holders are put in a dedicated queue so that when M3 AWD is available they get their cars ahead of holders who have higher numbers than theirs, who are not in that line. For some, how this is handled could determine whether they get 100% of the federal tax credit or 50%. Once the M3 AWD wait queue is drawn down, then position on regular reservation line and geographic location would again become the only considerations.

We can only go on what has happened in the past, with Model S and X. Basically, the queue determines when you get the offer to configure your car. But you can just sit on that offer. Only when you configure does your car enter the production queue. At that time you'll still be ahead of anyone who hasn't yet been able to configure, but then production batching takes over so the cars still won't be produced in exactly the order they were configured.

 

[McRat]

I think the bond markets (past and present) probably see these two companies closer than equity markets (present and future) do. The fact is both companies have manufactured cars in the past, even though their futures are very different.

Also, Tesla's PP&E was $8.4 billion as of June 30, not $0.6 billion. What am I missing in what you said?

DOH! Sorry, my bad. Stupid Right to Left timeline on TSLA financials. You're right about PP&E, but it's still not near GM. Different ballpark, 14.5% is not near 70%.

 

[Starno]

This is the first post (April 2017) in which I mentioned "non-dilutive debt," which shows I had not expected Tesla to do what it just did until 2018.

Elon & Team proved me conservative once again. Simply unbelievable.

Elon and Team are the type of people when in college that while everyone were struggling with some maths problems, they were just doing something else because they finished in 10 minutes.
Not surprising imo.

 

[22522]

ValueAnalyst said: ↑
Does anyone know why gigafactory is not being used as collateral? Why not back it with the asset and get a significantly lower rate? I don't see much of a default risk (clearly the bond market doesn't either) but even if Tesla defaults on this debt, it's not like gigafactory would be of much use anyway.



The ABL Credit Agreement through amendment 5 contained an exclusion of the GF in the definition of "Collateral" that could be attached to secure repayment of draws under the ABL. When the maximum commitment was increased June 16, 2017 from $1.2 billion to $1.825 billion that exclusion was deleted. Assuming the ABL creditors have made the proper filings under Article 9 of the UCC, the unsecured note would be behind the ABL creditors in priority to GF assets.

Tesla's investments in the GF have been reported as:


" In 2016, we used cash of $455.3 million towards Gigafactory 1 construction and expect to spend a total of approximately $770.0 million during 2017."

"We had cumulatively incurred and capitalized costs of $1.98 billion and $825.3 million, respectively, for Gigafactory 1 as of June 30, 2017 and December 31, 2016."

It appears $1.155 billion was expended in the 1st half of 2017 against the total of $0.77 billion expected for the year.

The primary reason the GF is of minimal value as collateral for debt is that the Panasonic lease grants the lessee rights of Quiet Enjoyment.

Some have asserted the bond solicitation is not really a recent decision to provide a contingency cushion against force majeure events, but something that has been planned for awhile.

I think you can only use something for collateral once...

 

[Reciprocity]

Usually automated storage and retrieval is a sign of misplaced priorities in manufacturing, as storage and retrieval add capital and operating costs (and more importantly cycle time to the inventory, and lags in cause and effect as downstream processes are often the place where flaws are found- they tie up money and effort and delay the discovery of flaws creating a junk exposure/risk) with no actual value added to the product.

So in trying to sort this, if an IC guy designed this thinking that buffers are good, there are differences between data and physical product, but Tesla is pretty smart, and they know that inventory turns matter, so this would be bad sorting on my part.

If color changeovers are so expensive that this facility is the only way to get more than one color on a truck, that might make sense.

I am from the school of low rate parallel machines that don't need a buffer in case they break down. Shigeo Shingo school I guess.

If it is to store finished product to avoid crossing the threshold, because the natural area of distribution is large, it is a natural capacitor, I would tend to rent space in mall garages, on a temporary basis. Or even rail cars or trucks.

If there is some sort of soak time required in the manufacturing process - reforming or something like that, it could make sense. Maybe they are moving a battery process step to occur in vehicle?

I see a pile of money, delay and misdirected effort in these types of buildings. Sometimes to serve a fragile high rate process with slow changeover time. If the inventory money and effort and lag costs were directed toward buying more manufacturing equipment or improving changeover speeds, ... usually it is better not to have these sorts of buildings.

Pretty sure I am missing something.

[Edit: AWK mind bending stream processing with no memory management and Shigeo Shingo single minute exchange of die( and other stuff) are the two things that set my expectation of high speed and low cost to a point where others think "impossible unrealistic." Appreciating these things creates conflict with those who don't...]

Great analysis and probably correct, but I dont know much about this stuff. The first thing I thought of was that this system was going to be used in part to remove any storage requirements from the main factory, while maintaining the access through automation. My point being that they are never going to be able to make parts just in time and many parts today are stored in the main manufacturing facilities today. If they move the stored parts out to this new facility, they can free up space for more production while still having good access to parts as they need them. To me it might be more about optimizing space then optimizing inventory. This system gives you a best of both worlds as they can build vertically and still have good access to the parts.

 

[Reciprocity]

The used schuler press was $50mil. I would expect that a new one to cost at least triple that. I don't know how much that building costs, though.

First shuller press was $6M, used from Detroit. New ones are $50M:

Tesla Factory - Wikipedia

NUMMI auctioned off[26] the press lines, robots and other equipment to Toyota's other US factories[27][28] while Tesla purchased over $17 million of manufacturing equipment and spare parts in 2011, at significant discounts compared to new equipment.[29] Tesla bought $50 million worth of Schuler SMG hydraulic stamping press lines used from Detroit for $6 million, including shipping costs.[30]

 

[DurandalAI]

First shuller press was $6M, used from Detroit. New ones are $50M:

Tesla Factory - Wikipedia

NUMMI auctioned off[26] the press lines, robots and other equipment to Toyota's other US factories[27][28] while Tesla purchased over $17 million of manufacturing equipment and spare parts in 2011, at significant discounts compared to new equipment.[29] Tesla bought $50 million worth of Schuler SMG hydraulic stamping press lines used from Detroit for $6 million, including shipping costs.[30]

Dang. I had pulled that $50mil amount off of an Electrek article.. Apologies for the incorrect information, and thank you for the correction.

Edit: I attempted to edit my original post to fix this, but it appears the forum will only allow me to edit/delete my most recent post.

 

[Bobfitz1]

We can only go on what has happened in the past, with Model S and X. Basically, the queue determines when you get the offer to configure your car. But you can just sit on that offer. Only when you configure does your car enter the production queue. At that time you'll still be ahead of anyone who hasn't yet been able to configure, but then production batching takes over so the cars still won't be produced in exactly the order they were configured.

Perfect. Thanks @ggr.

 

[MP3Mike]

Thank you. As long as changeover takes less than 10 minutes, you should not be getting behavior distortions to compensate.

I think it takes way longer than 10 minutes to change out all of the dies. That is why Tesla says that they make weeks worth of a particular part before they switch the dies to start making the next part. So every die change they can eliminate increases their press efficiency.

 

[donauker]

I think it takes way longer than 10 minutes to change out all of the dies. That is why Tesla says that they make weeks worth of a particular part before they switch the dies to start making the next part. So every die change they can eliminate increases their press efficiency.

This exactly.

If you take a factory tour you will see that this warehouse addition is absolutely essential to reaching the production levels targeted for Fremont.

 

[brian45011]

Are details on the bond issue already out? I may have missed it. Just seeing stock price took a nosedive, related or coincidence?

Should be soon, but's here the other side's speculation:

By Claudia Assis and Ciara Linnane, MarketWatch

'We think it's a terrible bond,' without many of the customary bondholder protections, says analyst

Tesla Inc. (TSLA) has been shopping its first pure corporate bond offering this week, and a price at the levels whispered on Wall Street would be a big score for the Silicon Valley car maker.

Investors, however, may have to curb their enthusiasm: the bonds are lacking many of the customary bondholder protections for this much risk.

"Anyone who looks at a lot of high-yield bonds would expect more robust protection against future debt," said Valerie Potenza, head of high-yield research at Xtract Research, a sister company of Debtwire. "We think it's a terrible bond, but people seem blinded by the Tesla story."

See also:The market hasn't cheered Tesla's earnings this much in nearly two years (http://www.marketwatch.com/story/ tesla-to-sell-15-billion-of-debt-to-help-fund-model-3-production-2017-08-07)

Tesla (TSLA) earlier this week said it plans to sell $1.5 billion in senior notes due 2025 (http:// www.marketwatch.com/story/tesla-to-sell-15-billion-of-debt-to-help-fund-model-3-production-2017-08-07) to help shore up its balance sheet ahead of the Model 3 production ramp this fall (http://www.marketwatch.com/story/elon-musk-gets-the- 35000-tesla-model-3-price-tag-he-wanted-but-extras-cost-more-2017-07-29).

A successful Model 3 production scale-up is the cornerstone of Tesla's expansion plans, which include new passenger and commercial vehicles, solar-power products and the ability to produce cars at a rate of 500,000 by the end of 2018.

The bond offering is being whispered at a coupon of 5.25%, according to market sources, reflecting Wall Street's continued strong appetite for riskier debt amid still-low interest rates on typically safer debt.

But Tesla faces "sizable near-term credit risks" associated with Model 3 production and sales, Moody's Investors Service said in a note Thursday.

Moody's assigned a B2 corporate family rating to the credit, placing it a full five notches into speculative, or " junk," territory to reflect those risks. For Moody's, that includes a "make-or-break" launch for Model 3, minimal proprietary technologies to keep Tesla's competitive advantage, formidable competitors and credit metrics that reflect its junk credit profile.

Moreover, the deal is likely the first of many, according to analysts at CreditSights, which means competing debt in the pipeline that may entice investors.

Read also:Another analyst joins the Tesla bulls, upgrading the stock to buy (http://www.marketwatch.com/story/another- analyst-joins-the-tesla-bulls-upgrading-the-stock-to-buy-2017-08-08)

High capital expenses and negative free cash flow "will be the reality" for Tesla for the foreseeable future. While Tesla holds an early-mover advantage in the electric-car space, barriers for electric vehicle penetration are still high, not to mention competition from rivals with bigger pockets and better distribution networks.

"We expect the deal will sell and perform well in a hot market, but we see 5.25% as inadequate compensation for the risks of the business and weak asset protection," CreditSights analysts wrote in a note.

Tesla's cash burn is another factor that might impact its ability to service its debt. The company had about $3 billion of cash at the end of June, but is expected to spend about $2 billion of that in the second half. Moody's is expecting "significant" capex and cash burn through 2018.

"Debt service could become an issue depending on how much debt they sell," said Xtract's Potenza.

Then there's the weaker structural protection Tesla is offering bondholders.

As a single B-rated credit, the Tesla notes might be expected to come with a full high-yield covenant package, a set of protections for the bondholder to guard the investment by preventing the company from using monies that could be assigned to interest payments for other purposes. The offering is lacking these protections.

"The covenants are weak," said Potenza. "There are no restrictions on the company issuing further unsecured debt, on selling assets and on paying dividends or making investments. There is nothing to stop them from issuing notes on the same terms in an unlimited amount going forward."

 

[3Victoria]

The comment about batching would seem to fit with the 40% uptime comment.

This exactly.

If you take a factory tour you will see that this warehouse addition is absolutely essential to reaching the production levels targeted for Fremont.

My impression is that at the start they had oodles of floor space so storing parts in it was easy and effective (essentially 2D storage). With adding Model 3 line(s) they need this space, and so are now floor space constrained. Building a 3D storage facility is both floor space conserving and likely more time efficient. Also, this fits more into the new alien dreadnought 3D philosophy.

 

[Curt Renz]

Auto-parts suppliers remain in denial: Musk's Electric-Car Vision Doubted by Major Auto-Parts Suppliers