Description

(Apologises for the formatting below - we will try to also load to a web location in the next day or so)

Mobileye

Mobileye had been discussed elsewhere as a short idea. I will give references to those articles but I have been doing my own work on the short thesis. I will avoid repeating the issues that have been raised elsewhere but rather raise other points of relevance.

I apologise in advance that this document is rather dense and makes reference to a lot of web links. May I recommend that the documents I link to are downloaded or printed in case any of the source material gets pulled from web servers.

Summary:

The summary of our short case (more details below) would be:

1. Mobileye might be a ‘tier 3’ supplier rather than tier 2

2. The chip is co-developed with a third party (ST Micro) and we suspect the Auto OEM and Tier 1 relationships are significantly dependent on ST Micro

3. There is shipping silicon from other suppliers which appears a generation more advanced in computing power

4. Mobileye has had difficulty asserting its patents

5. TRW has an IP portfolio that might lead to litigation if other Tier 1 suppliers bring Mobileye based products into the US

6. The academic heritage of Mobileye may be less unique than generally appreciated

7. The scale of the options dilution is underappreciated

8. The margins may not be as high as first appears

9. Future sales pipeline growth may be less than hoped for

10. The company is organised under Dutch law and therefore a hostile takeover is difficult

I recommend reading the following articles to give a background on the short case:

1. Seeking Alpha – https://seekingalpha.com/pro/checkout/3445726?notice=pro [note – need pro membership]

2. Citron Research - http://www.citronresearch.com/wp-content/uploads/2015/09/MBLY-final-Sept-9.pdf

I will also recommend reading a number of notes by Morgan Stanley (MS) who have been the leading bulls on Mobileye:

1. Mobileye: Seeing Green – initiating as Our New Top Pick by Ravi Shankar et al – 26 August 2014

2. Mobileye: Engaging Phase 2; Raising PT to $80 – 17 August 2015

We would also refer to the following note from Goldman Sachs:

1. ‘Mobileye: Disruptive tech, strong positioning, autonomous roadmap; Neutral’ – 26 August 2014

The differentiated reasons we believe Mobileye is a short are the following:

1. The automotive supplier structure

The major automotive makers (OEMS) rely on a number of major suppliers (Tier 1s) who are true partners in the design and manufacture of major parts of a car – eg a steering system, a braking system etc. The Tier-1s in turn rely on tier-2 suppliers who make individual components or subassemblies.

Mobileye is apparently a tier-2 supplier. Morgan Stanley consider being a tier-2 supplier in the auto industry to be advantageous to Mobileye. To summarise their views (26 August 2014):

1. OEM pressure at arms length
2. OEM Tier-1 share restrictions
3. Low risk business as a tier 2 supplier

For the sake of brevity we will not reproduce the whole Morgan Stanley argument (see page 9 of the MS note). However we disagree for the following reasons:

1. Tier 2 suppliers do face pricing pressure

2. Several tier 1 suppliers are developing their own solutions (we have had conversations with a few of the tier 1 suppliers) [Page 4 of the Citron tabulates some tier 1 suppliers – we have spoken to some of the members of that list and have formed the view that they are developing complete integrated systems rather than the point solution that Mobileye provides. Indeed the tier 1s consider Mobileye little more than a component supplier]

3. OEMs work with selected tier-1 suppliers who have deep vertical expertise. A tier-1 might outsource a particular component to an external supplier – but there always is the risk that the tier-2 supplier will be swapped or dual sourced.

2. Mobileye may not even be a tier-2 supplier – the ST Micro relationship

Morgan Stanley essentially describes ST Micro as a ‘fab’ (ie outsourced semiconductor fabricator) for Mobileye (page 10 of MS 26 August 2014). On page 9 of the Morgan Stanley note of 26 August 2014 it is stated:

‘If the issue is due to a manufacturing problem, MBLY's contract manufacturer STMicro bears the risk.’

    And at page 37 of the same report it states:

MBLY does not manufacture the chip itself, which is contract manufactured for it by ST Micro out of a facility in France.

Thus the Morgan Stanley analyst implies ST Micro is merely a ‘contract manufacturer’.

However the following press releases seem to indicate significant ST Micro involvement beyond merely being a third party fab (our emphasis in the extracts):

1. http://www.mobileye.com/wp-content/uploads/2011/09/STMicroelectronic_Mobileye.pdf

Press release on Mobileye website dated 17 Nov 2005:

‘today announced that the two companies are committed to developing, producing, and commercializing chips for the visual-aid driving-assistance segment of the automotive market’

And later in the same:

‘STMicroelectronics intends to manufacture Mobileye’s EyeQ-1 System-on-Chip and both companies will jointly develop the second-generation EyeQ-2.’

2. http://www.mobileye.com/wp-content/uploads/2011/09/CNNST.Money_.May08.pdf

7 May 2008:

‘STMicroelectronics (NYSE: STM), a leader in the design and manufacture of automotive ICs, and Mobileye N.V., the leading provider of Advanced Driver Assistance Systems (ADAS) to the automotive industry, today announced that the two companies have successfully sampled the second generation of their system-on-chip EyeQ2 for the vision-based driver assistance segment of the automotive market’

And later in the same:

‘By combining ST's automotive design and manufacturing expertise with Mobileye's strength in video-based driver assist systems we provide an optimal solution for innovative, automotive-market proven, and cost-competitive solutions for the automotive market.’

3. http://www.mobileye.com/wp-content/uploads/2011/09/ElectronicDesign.pdf

Article from Electronic Design on Mobileye’s website:

“As part of this trend, STMicroelectronics’ and Mobileye N.V.’s joint introduction of the second-generation EyeQ2 system- on-a-chip (SoC) 120-MHz processor marked an important milestone.”

In the same article the caption to photo 1 states:

‘This second-generation EyeQ2 SoC vision processor, co-developed and manufactured by ST Microelectronics and Mobileye N.V. ….’

4. ST Micro press releases state ‘co-development’ – for instance see http://www.st.com/st-web-ui/static/active/en/fragment/press/product_backgrounder/product_description/b0003.pdf

Dated Nov 2014

‘For example, ST collaborates with Mobileye on the development and production of chips for visual-aid driving-assistance such as lane-departure warning, forward collision warning, vision/radar fusion and pedestrian detection for active safety behind the wheel.’

5. http://www.prnewswire.com/news-releases/stmicroelectronics-and-mobileye-to-develop-third-generation-system-on-chip-family-for-vision-based-driver-assistance-systems-131044278.html

Press release dated 4 Oct 2011

States “[St Micro and Mobileye] today announced that the two companies are co-developing the next generation of their System-on-Chip processor for the vision-based driver-assistance segment of the automotive market”

6. http://www.semicontaiwan.org/en/sites/semicontaiwan.org/files/docs/st_technologies_for_a_green_safe_and_connected_car_1.pdf

Dated 28 August 2013

This ST Micro presentation, on slide 26 states (re EyeQ1, EyeQ2 and EyeQ3):

‘Co-development with WW leader Mobileye’

Incidentally it is worth going through this whole presentation to appreciation the complexity and number of items of silicon a car needs to become autonomous.

7. http://pdf.datasheetarchive.com/indexerfiles/Datasheet-039/DSA0088310.pdf

Oct 2006

This datasheet describes the Mobileye EyeQ1 and EyeQ2. It does mention Mobileye but the logo on the datasheet is from ST Micro and the contact information is for an ST Micro office. This also raises the issue whether it is Mobileye or ST Micro that has been selling the chips. Furthermore is Mobileye a tier 2 supplier to OEMs or a supplier to ST Micro, which in turn suppliers Tier 1 and OEMs – ie is Mobileye actually a tier 3 supplier?

8. Even Mobileye acknowledge that ST Micro is involved in the development of the chip:

Q4 2014 Earnings Call (26 Feb 2015):

Amnon Shashua: ‘The chip [EyeQ3] is developed in collaboration with our partner, STMicroelectronics, as were the previous chip generations, EyeQ1 to EyeQ3 [sic]’

(Source: Bloomberg transcript of conference call)

The above documents raise the following issues for us:

• How much of the know-how and IP belongs to ST Micro vs Mobileye?

• Who does the selling into automotive OEMs or their tier 1 suppliers – ST Micro or Mobileye?

• Who owns the customer relationship and the specification?

We acknowledge that ST Micro’s contribution may be the more basic building blocks of a chip design but we believe that STM has also contributed to the design process, sales process, the approval and manufacturing.

Readers may wish to check the number of employees, revenues or capex invested by ST Micro (STM FP <equity> on Bloomberg).

We would argue that a shared IP / sales position ought to receive a lower valuation than a truly unique IP / sales position.

Also of interest is the following extract from ARM’s results statement of 21 Oct 2015:

‘STMicro announced that it had licensed the new 32-bit ARMv8-R processor technology for chips targeting real-time safety-related smart driving applications and in industrial applications.’

Noting that Mobileye appears to have switched to a MIPs processor is STMicro licensing the ARMv8-R for other ADAS related features? Is Mobileye likely to merely be one component in an array of products produced by STM?

[After writing the above we found this in an F-1 filing (http://d1lge852tjjqow.cloudfront.net/CIK-0001607310/8447d5e5-ecec-4131-b177-5f1d8bd65238.pdf?noexit=true ) :

All of our EyeQ integrated circuits are manufactured by STMicroelectronics N.V., which is a leading supplier and innovator of semiconductor devices dedicated to automotive applications. Many of our customers are also direct STMicroelectronics N.V.’s customers, which allows us to benefit from their existing relationship with STMicroelectronics N.V. and gives our customers familiarity with STM’s manufacturing processes, including Quality Assurance, Customer Care, Failure Analysis and Manufacturing Standards. We believe that our relationship with STMicroelectronics N.V. also creates significant cross-selling potential for our products. Further, as a leading provider of integrated circuits, STMicrolectronics N.V. has the capacity to increase production of our EyeQ chips as our sales increase.

We believe that this confirms our view that without ST Micro, Mobileye would have only a limited relationship with OEMs and would find the sale process harder.

3. We do not believe the following assertion ‘To replicate Mobileye would cost billions’

In the MS note of 26 August 2014 it states:

The Tier-1s probably realize that trying to catch up and compete with MBLY may take several years and billions of dollars of effort.

We disagree for the following reasons:

1. Mobileye itself has not spent billions. If one strips out the management share options we believe the actual development cost for Mobileye has been less than $250M ($105.5M on a gross basis between 2012 – 2014; unfortunately we have no gross R&D numbers prior to 2012).

2. We have previously invested in private startup semi-conductor companies that have built sophisticated processing chips (to use but one example - SIMD, multicore processor with RISC / VLIW, H.264 decoding etc). We estimate that the cost to develop starting from scratch would be less than $100M and quite probably less than $50M

3. Several tier-1 suppliers / OEMs are already working on vision systems. Our conversations indicate that several tier-1 suppliers have deep internal silicon expertise. The Suhail Capital article gives examples of automotive makers (Toyota, JLR, Mercedes) who are using solutions that are not sourced from Mobileye

4. iOnRoad – we will discuss this company in a separate section

5. Semiconductor suppliers such as Texas Instruments already provide hardware at least as powerful as Mobileye.

4. ADAS / semi-autonomous / autonomous vehicles

Part of the excitement around Mobileye concerns discussions relating to autonomous vehicles.

Current technology (advanced driver assistance system) such as offered by Mobileye really offers alerts (eg you are too close to the car in-front). Moving to fully autonomous systems will require integration with the steering, acceleration and braking systems. Of the total system vision will only be one small part; furthermore it will not be the only sensor system – as well as proximity sensors (eg laser, radar etc) we believe a fully autonomous car will need inputs from the steering (which way are the wheels pointing) and the braking system (what is the road surface like, what is the current safe braking distance etc).

We believe it is unlikely that OEMs / tier-1 suppliers will allow Mobileye to provide the full system integration.

5. How sophisticated is Mobileye?

1. Smartphones / cameras

As the first step in this assessment we suggest you take your smartphone out of your pocket and take a photo of a colleague. If your phone is less than two years old it probably has an 8 megapixel (MP) or higher camera and also does face detection. It is pretty hard to find cameras or phones these days without similar specifications.

2. The MobilEye image sensor:

The following page (http://www.mobileye.com/technology/development-evaluation-platforms/cameras/ ) appears to suggest that Mobileye uses a Micron MT9V022 / MT9V023 image sensor with a resolution of 752 x 480 pixels (this equates to 0.36 MP (ie a fraction of the resolution of a smartphone))

We cannot confirm which generation of EyeQ this camera works – but the following pages confirm that at least one Mobileye product uses this camera - http://www.bollanddrivingsolutions.ca/mobileye.html and http://mobileye.com/en-en/technology/processing-platforms/seeq2/ ]. For reference typical VGA resolution is 640 x 480 but the term can be stretched to include 752 x 480.

We appreciate that resolution is not everything – especially for video – but we are surprised that the Mobileye system does not use a better image sensor. The counterarguments might be that the sensor is adequate and a higher resolution requires more processing power. Nonetheless the point we would make is that resolution is often considered a sign of technical prowess. (We will refer back to the camera resolution below).

Various notes suggest that Mobileye has a 'vision' system based on deep learning and neural networks. Our views are:

• it may sound a ‘semantic’ issue but Mobileye’s system is a ‘feature detection and interpretation’ system – not a ‘vision’ system – ie it detects objects in the foreground (eg a moving object) (by comparing background between subsequent images) and then interprets whether the object is moving closer or not; ie it does not ‘see’ an image and note that to the left there is a Ford driven  erratically by a grumpy guy; whilst to the right there is an individual with a limp, an eyepatch and bad dress sense who is crossing the road

• we are personally aware of university research into neural network based image recognition systems since at least the late 1980s

• it is not clear to us how much ‘in-field’ ‘learning’ the Mobileye system does – in fact we fear that ‘in-field’ learning is a risk factor – you do not want two, otherwise identical cars, to interpret the same environment differently – this incidentally also raises issues for insurance if and when cars become autonomous

• In summary ‘deep learning’ neural networks appears like ‘magic’ to investors but is a well understood technology and there are plenty of other teams able to replicate this

6. The SoC design

SoC or System on a Chip is a standard semiconductor term. Mobileye’s SoC comprises of 2 processors. We think the current version of the EyeQ2 chip involves two MIPS processors (see http://mobileye.com/technology/processing-platforms/eyeq2/  - half way down it states ‘Two Floating Point MIPS34K hyper threads RISC CPU’ ). However previous versions might have involved ARM9 processors (see http://mobileye.com/wp-content/uploads/2011/09/SoCIPQualifcation.pdf )

It is not clear to us why Mobileye switched from an ARM architecture to a MIPs architecture. Changing a semiconductor chip architecture is a big deal and most companies do not do this unless there is a very very pressing reason. As well as cost it takes a lot of time – this also of course means that some of Mobileye’s invested R&D dollars would not need to be spent by a competitor.

7. Competitors

Though Sunhail Capital has previously highlighted competitors to Mobileye we did not appreciate the level of competition until we started looking into the details. We would like to highlight the following links / comments:

1. Texas Instruments

1. In the following release Texas Instruments states that they have shipped 15 million chips for ADAS:

‘Confirming a leadership position in the automotive market, today Texas Instruments (TI) (NASDAQ: TXN) announces that more than 15 million of TI's advanced driver assistance systems (ADAS) System-on-Chip (SoC) devices are on the road.’

(Source: http://newscenter.ti.com/2014-07-30-Texas-Instruments-surpasses-shipment-of-15-million-SoCs-enabling-the-automotive-industry-to-redefine-advanced-driver-assistance-systems - date 30 July 2014 – we accept that this refers to not only vision systems but the point is that Mobileye is not the only player of significance – as discussed below the latest figures from Mobileye are 5.2M units in its lifetime)

2. TI’s Front Camera solution is at the following page:

1. http://www.ti.com/lsds/ti/processors/dsp/automotive_processors/tdax_adas_socs/applications.page#fr

2. We would recommend readers watch the following video from TI re their TDA2x system (http://link.brightcove.com/services/player/bcpid3936518607001?bckey=AQ~~,AAADeK19QZE~,BDUED4gHjR-YvTA_lSqgnbyawHVyHfVZ&bclid=3897436132001&bctid=3881591689001 ) – in particular note that this system works with 1280 x 720 ie 921,600 pixels – the demonstration then goes on to 1080p60 (it will be recalled from earlier that Mobileye appears to work with cameras with a pixel count of 752 x 480 ie 360,960 pixels)

3. The TDA3x works upto1920 x 1080 p 60 (ie 2MP - see http://www.ti.com/product/tda3#features )

4. The key point with the above is that the TDA 2x and 3x appear to be a generation ahead of what we can discern as the maximum resolution for Mobileye.

2. PLK:

1. PLK appears to be a spin out from Hyundai (see http://plkast.en.ec21.com/company_info.html ) and is offering its ADAS system to third party OEMs. Its clients include Valeo and Continental (Tier 1 suppliers to the auto industry) – see http://plktech.malgum7.gethompy.com/en/about-plk/customers/ .

3. Details of Freescale’s offering is at http://www.freescale.com/applications/automotive/adas/front-view-camera:APLFVCAM

4. Renesas offering is at http://www.renesas.com/applications/automotive/adas/solution_kits/adas_starter_kit/index.jsp

5. ARM has started focusing on the ADAS market - http://www.arm.com/markets/embedded/advanced-driver-assistance-systems.php

8. Research:

1. The following page on the Mobileye website points to articles regarding research by the company; there appears to be no links beyond 2011:

http://www.mobileye.com/technology/mobileye-research/

2. In the Q2 2015 earnings conference call (6 August 2015):

The proximity to Hebrew University, which has one of the best Computer Vision Science departments in the world, is a very big advantage because we're heavily related to the university. And we usually take the best students while they're already in a progress of their studies and we never had a problem to grow.

We accept that the Hebrew university may have one of the better computer vision departments globally but we do believe that investors would be wise to remember that there are an awful lot of other computer vision departments and researchers. The following lists may be of some interest:

• https://www.quora.com/What-are-the-best-schools-for-studying-Computer-Vision

• http://academic.research.microsoft.com/RankList?entitytype=2&topdomainid=2&subdomainid=11&last=5

9. Peculiar broken links on website:

Mobileye is a $10Billion market cap company. Yet the website appears to have peculiar broken links:

1. On the following page click on the ‘cameras’ link at the bottom- http://www.mobileye.com/technology/development-evaluation-platforms/

2. On the following page click on the ‘Mobile EyeQ3’ link in the last paragraph –- http://www.mobileye.com/technology/processing-platforms/  

10. iOnRoad:

‘iOnRoad’ is an app for the iPhone which provides driver assistance. It would be comparable to using TomTom or Google maps on your phone rather than an in-built satnav. It retails for $0.99. [See http://www.ionroad.com/ and https://itunes.apple.com/app/id569507817?mt=8 ]

There are a number of points about this app that we think are extremely important:

1. The TomTom analogue

Cars used to have in-built satnav systems as expensive options in high end cars. The launch of TomTom’s standalone devices showed that there was demand but also reset the price significantly lower.

The entry of Google (GoogleMaps) and Apple (AppleMaps) allowed navigation apps to be run on phones / tablets and lowered the price consumers were willing to pay even further.

Auto OEMs do continue to offer navigation as an in-built option but the ‘reset’ in price means that the price premium has collapsed – though volumes have grown.

2. How much unique IP does Mobileye have?

Part of the bull case for Mobileye is that it has unique image processing algorithms and that these algorithms must be run on specially designed silicon closely coupled with the image sensor to get the best advantage.

However iOnRoad as an app running on standard (mobile phone) hardware negates all these issues.

3. Litigation between Mobileye and Picitup (iOnRoad) in NY District court:

It would be easy to dismiss iOnRoad as a simple, basic consumer app. However Mobileye certainly did not and took iOnRoad (and its parent company Picitup) to court. The ruling, by the New York District Court on 5 March 2013 is available at http://www.leagle.com/decision/In%20Adv%20FDCO%20140116-000348/MOBILEYE,%20INC.%20v.%20PICITUP%20CORP.#

The court considered if iOnRoad infringed Mobileye’s patents; we will defer from delving into the depth of each patent but rather merely extract the key conclusion for each patent from the above link (our emphasis):

1. ‘867 patent:

‘The Court thus finds that Mobileye's only arguments on this issue are unfounded, and further concludes that iOnRoad's app does not infringe the asserted claims of the '867 patent because it does not employ a claimed scaling factor, for precisely the reasons iOnRoad argues.’

2. ‘101 patent:

‘Accordingly, the Court concludes that the iOnRoad app does not infringe the asserted claims of the '101 Patent. ‘

3. ‘621 patent:

‘The Court thus concludes that the iOn-Road app does not infringe the asserted claims of the '621 Patent.’

Note that Goldmans (26 August 2014) state:

The company has 18 US patents, six European patents, 29 US patent applications, 24 European and other non-US patent applications and eight provisional patent filings.

So the question for investors is how much is Mobileye’s IP worth?

Other issues in court - advertising

In addition the court considered if the iOnRoad product advertising could be confused by consumers with the high end Mobileye products fitted in luxury cars. The conclusion was:

‘Based on the Comparative Ads Survey, a reasonable jury could find that the comparison's between iOnRoad's app and similar systems in luxury vehicles are impliedly false and mislead a substantial percentage of consumers.’

We think this is actually an ‘own goal’ for Mobileye as it illustrates that for some consumers a $0.99 app on a mobile phone provides functionality that is apparently similar to that of a Mobileye system.

    Other issues in court - branding       

We have heard bulls suggest that Mobileye will become the ‘Intel inside’ for ADAS for cars; and that specifying Mobileye as the technology provider will help sell cars. The court did not think that the Mobileye brand was particularly strong:

‘the Court has already noted that the evidence is rather weak that Mobileye's mark has even acquired secondary meaning, much less that it is "widely recognized by the general consuming public of the United States." MOBILEYE simply is not a famous mark.’

A useful summary of the case is at http://tushnet.blogspot.co.uk/2013/03/are-mobileye-and-ionroad-confusingly.html

4. Litigation in the patent court:

iOnRoad has also gone to the US Patent court to try to overturn Mobileye’s patents – the decision of the judge was made on 27 August 2013 – see http://ipr-pgr.com/wp-content/uploads/2013/11/ionroad17.pdf

The key conclusion of the judge was:

‘For the forgoing reasons, we determine that the information presented in the Petition establishes that there is a reasonable likelihood that iOnRoad would prevail in showing unpatentability of the challenged claims of the ʼ621 patent.’

There was a denial for a rehearing on 2 October 2013 regarding the Suzuki prior art – see http://www.aiablog.com/wp-content/uploads/2014/01/iOnroad_v_MobileyeTechs_IPR2013-00277_Paper23_2013OCT02.pdf

[It seems that the Suzuki prior art was rejected by the court due to questions relating to the date; in the rehearing iOnRoad tried to present more evidence that it had priority but this was not accepted due to it not having originally been provided; not due to the merits of the prior art. We presume other litigants, or courts in other countries would be able to refer to the Suzuki (potential) prior art.]

The case was settled on 13 Dec 2013 – see http://www.uspto.gov/ip/boards/bpai/uspto__ptab_inter_partes_review_petitions-terminated_cases__as__of__20140402.pdf -(search for case IPR2013-00227) - however we have not been able to ascertain the terms of the settlement

5. Harman

One might ask why we are focusing on a relatively small company such as iOnRoad. Interestingly it was  acquired by Harman on 10 April 2013 – see http://www.harman.com/EN-US/Newscenter/Pages/HARMANacquiresiOnRoad.aspx#.ViYBw_mrRaR

Harman is a $7.4bn market cap, $6.1bn revenue company focussed on audio and on in-car systems.

We therefore believe that over time Harman will provide the resources for iOnRoad to continue and possibly already be engaged with auto OEMS to displace Mobileye.

6. It is worth considering that the iOnRoad app is processing a higher resolution video stream than the Mobileye device appears to; without specialised hardware

7. Incidentally there is at least one other ‘augmented driving’ app we have found on the itunes store - Augmented Driving By imaGinyze - https://itunes.apple.com/us/app/augmented-driving/id366841514?mt=8

11. TRW / Magna:

Both Magna and TRW are well established tier 1 supplier to auto OEMs. There has been various litigation between the parties regarding vision systems based on the Mobileye chips. See

https://docs.rpxcorp.com/lits/851/14401/miwdce-76604.pdf?Signature=%2F4PoMkcxKMRyBXG2QCW3NKDyYWo%3D&Expires=1445333583&AWSAccessKeyId=AKIAI2UWKALIEYBVOKDA (note that this link does not always work)

    and

http://cdnet.stpi.narl.org.tw/techroom/pclass/complaint/Complaint_pclass_14_A021b_TRW%20v.%20Manga(ITC).pdf

(also here with a slightly different format - http://cdnet.stpi.narl.org.tw/techroom/pclass/complaint/Complaint_pclass_14_A021a_Magna%20v.%20TRW(ITC).pdf )  

Also see:

http://www.itcblog.com/20130924/trw-automotive-files-new-337-complaint-regarding-certain-vision-based-driver-assistance-system-cameras/

Frankly we are trying to understand the different strands of this litigation but the key points for us are:

• Both TRW and Magna are Mobileye licensees

• TRW asserted some prior art from patents and knowhow it had obtained when it acquired Lucas

• Much or all the litigation appeared to have been settled in 2013 or early 2014, out of court, in time for Mobileye’s IPO

The above assumption that the litigation was now settled meant we were relaxed about any issues; however we then found the following dated 31 July 2015: http://www.usitc.gov/secretary/fed_reg_notices/337/337_907_public07312015sgl_0.pdf suggesting on-going ITC investigations.

The key points for us are:

• Morgan Stanley highlights the number of tier 1 suppliers who are Mobileye customers

• BUT it is not clear to us what the overall patent situation is – if other tier 1 suppliers or auto OEMs ship into the US are they likely to get embroiled into patent cases involving Magna and TRW?

• How much of Mobileye’s patent estate is anticipated or is obvious due to pre-existing IP at TRW, at Magna or at other tier 1 suppliers or automakers themselves?

• How enthusiastic is an automaker going to be to integrate into its core offering (as opposed to an add-on option) a product, which might suffer ITC sanctions or other litigation?

• In view of the above we question the rapidity of take-off with the US for ADAS. (Outside of the US patents have different rules and we defer from an opinion)

[Any help from readers in understanding the IP position would be appreciated]

12. Help – DCF:

We have avoided directly criticising too much of the Morgan Stanley research but, as value investors we felt like screaming when we read the following regarding their DCF ( page 5 - 6 MS note of 17 August 2015):

Firstly (our emphasis):

We use a 9.6% cost of equity and 9.5% WACC, which is right in the middle of our auto coverage universe.

We are not quite clear what precisely is his auto coverage universe but we would have thought that MBLY might perhaps have one of the higher costs of equity.

Secondly – we found the following, which we present with no comment:

‘We do not dilute the share count for stock comp as we assume MBLY will use its strong FCF generation to mop up dilution (starting soon, though maybe not immediately). None of these assumptions make a material difference to our valuation.’

…. ok, we give up, we will in fact comment, or rather quote the following from Mobileye’s 2014 accounts (see 20-F):

The Company has adopted a stock option plan (the “2003 Plan”), whereby up to 11% (out of the Company’s issued and outstanding aggregate number of shares of all classes) options may be granted to employees and service providers for purchase of the Company’s Ordinary shares. In May 2014, the Company increased the pool of options to be available under the 2003 Plan to up to 18% of the issued and outstanding shares of the Company.

But the most interesting thing about the share options is elsewhere in the 20-F (due to formatting issues with tables we will summarise rather than quote) – in the year to December 2014 Professor Amnon Shashua received $30,725,796 and Ziv Aviram received $29,474,309 in ‘pension, retirement, options and other similar benefits’. The footnote to the table however states:

1. Substantially all of these amounts reflect expenses recorded with respect to 16 million options granted to our Founders at a weighted average fair value of $4.82 per share on the grant date computed in accordance with ASC 718. These amounts do not reflect the actual economic value realized by the executive director.

So, if we have interpreted this correctly the senior management received share options priced at $4.82 per share. On the date of the Morgan Stanley report (17 August 2015) the current price of Mobileye was $57.87 (front page of MS report).  Companies buyback shares using the market price not the accounting price for the issuance of share options – so if, to use the Morgan Stanley analyst’s assumptions, the company were to buy back shares issued to avoid dilution we estimate it would cost the company a ‘mere’ $ 848.8M (16 M x $(57.87 – 4.82)).

[Incidentally we recognise that 16M x $4.82 is $77M which is greater than the $60M ($30,725,896 + $29,474,309) that is recognised as payment to Amnon Shashua and Ziv Aviram in the 20-F under ‘pensions, retirement, options and other similar benefits’. We are unsure why there is a mismatch – is it tax?]

It may be fair to say that the above share options were issued as a result of the management having succeeded in IPO-ing the company and may not be fully reflective of the rate of future share based and share option based compensation. Currently the company has two equity compensation plans for which 11.8M shares remain available as of Dec 31, 2014 (see annual report for 2014 under ‘Share-based Compensation’).

However our point is buying back shares to offset future equity compensation could have a significant impact on cashflow available to shareholders. Furthermore the company has been a regular issuer of share options – even when it was private.

13. High margins, R&D and pipeline:

Generally investors prefer high margin companies. In the case of Mobileye it receives money from its OEM customers in two ways:

• Selling product (ie chips) at a ASP of $43 – 45

• A non-recurring engineering (NRE) reimbursement when it works on a specific project for an OEM

However Mobileye does not recognise the NRE through its revenue line but rather as a negative charge for its R&D. Consequently unless one is careful the margins for Mobileye optically appear higher than they are. Secondly the growth in the NRE fee ought to be an indicator of future product revenue (ie an OEM is likely to ask for specific work on a product a year or two ahead of launching a car with that product integrated).

NREs were described in the following way in the prospectus (http://d1lge852tjjqow.cloudfront.net/CIK-0001607310/2cc54cf3-8d5f-463d-bf4e-8e12c9a2bb96.pdf?noexit=true ):

Our research and developments expenses are partially offset by non-refundable Non-Recurring Engineering (“NRE”) reimbursement that we receive from OEMs attributable to specific development programs with the OEMs. Any such reimbursement is not contingent upon success of the program. We retain all the rights to our work on these programs.

Based on information in the 20-F for 2014 for the gross and net R&D spend we estimate the following NREs (ie difference between gross and net revenue):

    The key questions for us are:

• Why is the gross R&D spend not reported quarterly?

• Why is the NRE not reported quarterly?

• Why is the NRE flat over three years? And what implications does that have for future OEM product adoption?

• How many OEMs and how many OEM projects does the NRE reflect each quarter or annually?

14. Are there some missing units or is there a misprint?

The following press release ST Micro (http://www.st.com/st-web-ui/static/active/en/fragment/press/product_backgrounder/product_description/b0003.pdf dated November 2014) states:

Safety  

ST provides a broad range of solutions to increase vehicle-occupant safety, including devices for airbags, anti-lock brakes, traction control, electric power steering and suspension systems and is the leading supplier of chips for automotive airbags and anti-lock braking systems (ABS), which currently represent the largest portion of automotive safety electronics.  

The Company is also a leading player in advanced safety systems that help avoid or minimize the severity of traffic accidents. For example, ST collaborates with Mobileye on the development and production of chips for visual-aid driving-assistance such as lane-departure warning, forward -collision warning, vision/radar fusion and pedestrian detection for active safety behind the wheel. These solutions, adopted by several Tier1 suppliers and car manufacturers in Europe, the US, and Japan, are now deployed in nearly ten million vehicles4.     

It is not entirely clear whether the ten million vehicle number refers to Mobileye equipped vehicles or the entire ST safety portfolio. Furthermore the footnote (footnote 4) refers to a previous ST Micro press release dated 24 Oct 2012 which stated Mobileye and ST Micro have shipped 1 million chips (see http://www.st.com/web/en/press/cn/t3353?s_searchtype=keyword .

For reference in its conference call dated 26 Feb 2015 (Q4 2014 Earnings Call) Mobileye stated it had shipped ‘over 5.2M units’ in its lifetime. We are therefore intrigued by the ten million unit reference but uncertain if it is a typo or refers to the whole of ST Micro’s auto safety product range. This also does, however, again raise the issue whether perhaps ST Micro is selling the Mobileye chip as part of a bundle of products to Tier 1 suppliers?

[We plan on confirming that the ‘ten million’ is a typo or refers to all safety products with ST Micro IR shortly]

15. Is Mobileye widely adopted?

We think it is a reasonable to consider whether Mobileye is widely adopted or if adoption is concentrated in a few OEMs currently. Based on data in an F-1 filing (prospectus – see http://d1lge852tjjqow.cloudfront.net/CIK-0001607310/8447d5e5-ecec-4131-b177-5f1d8bd65238.pdf?noexit=true ) we estimate the following:

Growth at Mobileye has been driven by GM, Nissan and BMW. Most intriguing however is why Honda sales have not grown in 2015 over 2014?

16. Is a takeover premium justified?

In their note of August 26, 2014 Goldman Sachs state that they believe Mobileye deserves a valuation based on M&A possibilities:

‘In our valuation framework for our European Tech coverage (see Exhibit 40),we assign a 30% M&A weighting for companies that screen as potential acquisition targets given strong technology, strong customer bases, and potential savings from cost synergies for a potential acquirer. We believe Mobileye meets all of these criteria:’

However we believe that investors may not have fully appreciated the ability of Mobileye management to block a hostile acquisition as Mobileye is a Dutch organised holding company. From the 2014 Annual Report (our emphasis):

In evaluating a response to a takeover offer, whether hostile or friendly, our board of directors is required under Dutch law to take into account not only the interests of shareholders, but also the interests of all other stakeholders in the Company, including employees, creditors, customers and other contract parties. Under this legal standard, our board of directors would be authorized to reject a takeover offer that it views as less beneficial to the interests of our employees and other stakeholders than a competing offer, even if the rejected offer were for a higher price than the offer that is accepted. This rule would also apply in an “auction” situation where the Company is actively encouraging competing offers to acquire the Company or its business.

On additional issue to consider from an M&A point of view is that currently Mobileye pays tax at 10%. A non-local purchaser might need to adjust the tax rate depending on inter-country tax treaties in his or her model.

Conclusion:

Ideally we would like to provide a valuation for Mobileye. However the only conclusions we can come to are that the market share forecasts embedded in the current share price are too high, that the competitive threat is underestimated and the IP position is not clear. Furthermore the share based and option based compensation has significant implications on cashflows (ie sharebuybacks to offset the dilution and cash inflows due solely to options exercises). In essence Mobileye is a short – but it might fall further than other commentators have highlighted because of the issues highlighted in this note.

ROTR

Catalyst

1. Increasing free float

2. Recognition of massive options dilution of public market shareholders

3. Recognition of the competitive position of other players

4. Recognition that there is a lack of IP exclusivity

5. Recognition of the excessive valuation