You are currently viewing SemiWiki as a guest which gives you limited access to the site. To view blog comments and experience other SemiWiki features you must be a registered member. Registration is fast, simple, and absolutely free so please, join our community today!




Page 2 of 2 FirstFirst 12
Results 11 to 18 of 18

Thread: Strong Q3-2017 results ASML

  1. #11
    Admin Daniel Nenni's Avatar
    Join Date
    Aug 2010
    Location
    Silicon Valley
    Posts
    4,005
    Thumbs Up
    Received: 710
    Given: 2,021
    Quote Originally Posted by IanD View Post
    5 EUV layers is correct for N7+, mask count saving is 10 layers -- this might be the reason for the two numbers you heard...
    5 layers for TSMC 7nm and 10 layers for Samsung 7nm.

    0 Not allowed!
    Now available in print or Kindle: "Mobile Unleashed: The Origin and Evolution of ARM Processors In Our Devices"

  2. #12
    Blogger
    Join Date
    Sep 2012
    Location
    Greater New York City Area
    Posts
    109
    Thumbs Up
    Received: 42
    Given: 4

    No new EUV orders suggests test/digestion of new tool Memory @ 77% orders = huge "exp

    ASML reported a good quarter, coming in at revenues of Euro2.5B versus guide of Euro2.2B and EPS of Euro1.29 versus guide of Euro1.07. 22 ArF units were shipped versus 16 in the prior quarter a sharp uptick which we think is reflective of the early wave of memory business.

    No new EUV orders were received, versus 8 last quarter, which we think implies a "digestion" period where customers are trying to figure out if the new tool is all that ASML has promised. It does appear that the new tool has fixed many of the reliability and uptime issues but we are sure customers want to take time to take it for a test ride before committing further.

    The bigger news, that will likely drive the stock, is the drop in forward guidance to Euro2.1B which translates to an EPS of just over Euro1.0. Even if we assume a 15-20% standard "beat", it will still be short of Q3.

    The company expects to ship 6 EUV systems in Q4 which is what it has shipped year to date.

    At 77% of business, memory orders are in nosebleed territory even well above memory centric players like LRCX. As memory does not need EUV any time in the near future we can translate to most all of these orders being for ArF systems which at least keep the margins high.

    We didn't see or hear of much progress on EUV other than normal block and tackling. Pellicles are here to stay which obviously means that throughput takes a 10% hit.

    Lumpy, Bumpy business

    No EUV orders and huge ArF orders make for lumpy business characteristics. Obviously the industry is all about memory right now which also means no need for EUV (in the near term). The guide down to Euro2.1B in revenues seems unusually large given the long lead times in litho systems usually allow manufacturers better "scheduling" of revenue. We are somewhat surprised that ASML couldn't do a better job of managing this lumpiness. Shipping 6 EUV tools in Q4 is also a higher goal but I guess they will be less busy with ArF tool shipments.

    Huge "exposure" to near term memory business
    At 77% of orders, memory is driving the company. As memory does not need EUV right now, it is clear that most was for ArF which sports higher gross margins but does not move the EUV ball forward. Given that NAND is less litho intensive its clear why AMAT and LRCX are out performing on a relative basis.

    Most importantly, we are quite concerned about the very high level of memory business because if memory falls off a cliff these orders could be canceled or pushed out creating a hole in revenues before EUV kicks in.

    The bottom line is that ASML is very highly dependent upon memory remaining strong to bridge revenues until EUV ramps up which is really at 5NM and a few years out in 2020. This is a significant risk in our view especially given the volatile nature of the memory market and fickle nature of capex spend related to it. We think memory related orders are much less reliable than EUV orders or logic/foundry orders.

    EUV (in)digestion period?
    The lack of EUV orders versus 6 last quarter suggests that customers have enough tools in the hopper in the near term and are likely figuring out how well the new tool versions work.

    While we have heard positive comments out of TSMC relative to their commitment to EUV we are still a way off from real insertion at 5NM versus initial, few layers, at 7NM.

    Intel continues to lag behind Samsung who is in the lead in EUV followed by TSMC (recently on board )with Intel still taking a cautious route.

    Its not like everyone is clamoring to use EUV ASAP and the remaining issues still need work. After some critical positive breakthroughs over the past year which created some EUV momentum we may be slowing a bit in overall progress as some remaining issues may be more pesky/difficult.
    The focus on memory does take investor focus away from EUV issues which is likely a positive. But we would anxiously await a return for EUV orders in Q4. If not, we would get much more concerned as customers could be pushing out.

    ASML direct challenge to KLAC's dominance
    With little fanfare, ASML announced that it had shipped its first eXplore 6000 E beam, reticle "inspection" system to a foundry/logic customer (we assume TSMC). It does take 24 hours to review a mask so it probably does do "inspection". While its not clear if this is a full inspection tool or just a "review" tool it almost doesn't matter as it undoubtably will be tightly tied to ASML's EUV tool and be aimed at enabling customers to get EUV off the ground. ASML management did throw KLAC's "optical" under the bus during the conference call.

    ASML also announced that it shipped the first "pattern Fidelity Management" tool the ePfm5. The first joint tool developed since the HMI acquisition. While its clear that this is a "review" tool, it adds to the metrology/inspection arsenal that ASML is developing for EUV ramp.

    Right now , KLAC is somewhat "left out" of the game of doing "at wavelength" or below wavelength E beam EUV inspection/metrology and has cobbled together an "optical" alternative because they have nothing else to respond with after cancelling their program.

    We view this as a very clear risk to KLA's otherwise dominance in the inspection/metrology business. KLA is also getting squeezed on the other side with E beam offerings from AMAT. KLA needs abetter response.

    Nosebleed valuation?
    At 25-30 times forward EPS, we think ASML's stock is likely overdone especially when taking into consideration the near term risk of memory and little progress on EUV with no orders.

    This valuation is double the equivalent US tool company who are right now just trying to get to a 15 multiple.

    ASML has always, historically traded at a premium due to European investors who are stuck with relatively few successful tech companies to invest in and drive up the valuation. But if I were an investor who could buy US and European stocks I see no reason to own ASML with the added near term risk at double the valuation of its higher growth US counterparts.

    If we were Intel, we would think about unloading some more ASML stock at these levels as it recently did. INTC invested $4.1B in the $50 range and has now more than tripled their money. A more than $8B profit is more than enough to buy all the EUV scanners Intel needs until Moore's law runs out.

    Maybe I can buy some shares of Uber and use the profits to take rides for the rest of my life.....what a concept...

    Collateral Impact
    The memory spend confirms what we have heard from AMAT and recently LRCX, but the 77% raises eyebrows. The large wave of ArF systems for memory likely confirms the view that 2018 will be great for fabrication tools as litho tools are the leading indicator of fab outfitting.

    If fabs are ordering/taking litho tools now, then they will be buying lots of dep and etch in a quarter or two, obviously supporting a strong 2018 for dep and etch makers.

    We are a bit more concerned about KLAC as metrology/inspection tools have similar timing to litho tools as fabs need to order KLA tools before dep/etch tools to nail down their process first before filling out a fab. We might be a bit more concerned about a similar digestion period as customers figure out KLA's new tool generation and new 7NM process flow before ordering more tools.

    Overall ASML's report is both positive and cautionary at the same time as it confirms the memory strength but also confirms the risk as the EUV ramp remains slow and memory remains fickle.

    2 Not allowed!
     

  3. #13
    Admin Daniel Nenni's Avatar
    Join Date
    Aug 2010
    Location
    Silicon Valley
    Posts
    4,005
    Thumbs Up
    Received: 710
    Given: 2,021

    Smile

    And this just in:

    Keeping Leadership in 7-nano Era: Samsung Electronics Seeks to Buy Up Next-gen Semiconductor Mfg Equipment | BusinessKorea

    Samsung is now in talks to buy 10 EUV lithography tools and it is almost the equivalent to the total number of EUV lithography tools that ASML can be produced a year. ASML believes that it can produce about 12 EUV lithography tools this year. It is the only company that manufactures EUV lithography tools in the world.Since Samsung Electronics has been exclusively buying next-generation semiconductor production equipment before its competitors, the company is seeking to overwhelm its competitors again with the same strategy, according to market watchers. In this regard, Samsung Electronics and ASML said, “We cannot officially confirm the fact.”


    0 Not allowed!
    Now available in print or Kindle: "Mobile Unleashed: The Origin and Evolution of ARM Processors In Our Devices"

  4. #14
    Admin Daniel Nenni's Avatar
    Join Date
    Aug 2010
    Location
    Silicon Valley
    Posts
    4,005
    Thumbs Up
    Received: 710
    Given: 2,021
    The CEO of ASML (Peter Wennink) will be speaking at the TSMC 30th Anniversary celebration today in Taipei along with Nvidia CEO Jensen Huang, Qualcomm CEO Steve Mollenkopf, ADI CEO Vincent Roche, ARM CEO Simon Segars, Broadcom CEO Hock Tan, and Apple COO Jeff Williams. They will also do a panel moderated by Morris Chang. Should be interesting. I will let you know how it goes.

    0 Not allowed!
    Now available in print or Kindle: "Mobile Unleashed: The Origin and Evolution of ARM Processors In Our Devices"

  5. #15
    Member
    Join Date
    Aug 2016
    Posts
    4
    Thumbs Up
    Received: 0
    Given: 0
    So by 2019, in terms of EUV usage by layers, is this the best guess? SEC LSI - 10, TSMC - 4-5, SEC DRAM - 2, everyone else - 0.

    The silence from Intel is deafening!!

    0 Not allowed!
     

  6. #16
    Admin Daniel Nenni's Avatar
    Join Date
    Aug 2010
    Location
    Silicon Valley
    Posts
    4,005
    Thumbs Up
    Received: 710
    Given: 2,021
    Quote Originally Posted by Ghostofmoore View Post
    So by 2019, in terms of EUV usage by layers, is this the best guess? SEC LSI - 10, TSMC - 4-5, SEC DRAM - 2, everyone else - 0.

    The silence from Intel is deafening!!
    Remember this one?

    Intel exec says fabless model “collapsing”
    Intel exec says fabless model "collapsing" | EE Times

    SAN FRANCISCO – It’s the beginning of the end for the fabless model according to Mark Bohr, the man I think of as Mr. Process Technology at Intel. Bohr claims TSMC’s recent announcement it will serve just one flavor of 20 nm process technology is an admission of failure. The Taiwan fab giant apparently cannot make at its next major node the kind of 3-D transistors needed mitigate leakage current, Bohr said.

    Hahaha........... Could the IDM model finally be collapsing?

    0 Not allowed!
    Now available in print or Kindle: "Mobile Unleashed: The Origin and Evolution of ARM Processors In Our Devices"

  7. #17
    Member
    Join Date
    Aug 2016
    Posts
    4
    Thumbs Up
    Received: 0
    Given: 0
    I actually think Bohr will prove to be right on the future of the fabless model - you could argue thay Apple et al already are 'virtual IDMs', but Intel has the advantage of being a true IDM - manufacturing only has one major customer. They can still compete on shrink without playing around with EUV in my view. The question is can they compete on cost (much more doubtful).

    0 Not allowed!
     

  8. #18
    Top Influencer
    Join Date
    Mar 2012
    Posts
    432
    Thumbs Up
    Received: 138
    Given: 111
    Quote Originally Posted by Ghostofmoore View Post
    So by 2019, in terms of EUV usage by layers, is this the best guess? SEC LSI - 10, TSMC - 4-5, SEC DRAM - 2, everyone else - 0.

    The silence from Intel is deafening!!
    From reading their VLSI Technology paper on 2nd gen 10nm this year, I got the feeling Samsung LSI had not developed SADP yet, at least for metallization, which is why they had been counting more heavily on EUV. It seemed quite strange to me since Samsung planar NAND had been using SADP (the appropriate way for metallization) for quite a while. The other FinFET foundries already use SADP or even SAQP.

    0 Not allowed!
    Last edited by Fred Chen; 2 Weeks Ago at 05:23 PM.
     

Page 2 of 2 FirstFirst 12

Tags for this Thread

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •