Article Info

AMCC – Further Down the Spiral

image AMCC (AMCC) held a conference call last Friday to review preliminary FQ108 results. The company indicated in April that FQ108 would be $60M down from $70M in the previous quarter (see “AMCC Kicks the Distribution Habit“). The final tally now puts it closer to $50M, a quarter over quarter decline of nearly 30%. This is worthy of detailed examination.


The explanations for the decline were painfully detailed. The company covered quarterly revenue by product line and provided explanations for each of the shortfalls. The revenue information was welcome as it allowed a clear picture of quarterly product trends to be assembled. I don’t believe this level of detail was possible without the explanation on Fridays call as it allowed some information gaps from previous quarters to be sewn shut.

  Q207 Q307 Q407 Q108E Q208E
Processor 35.3 35.4 33.3 19
Transport 16.4 12.4 13.3 11
Storage 12.2 13.7 12.5 13.2
Non Focus 10.3 9.7 6.5 4.5
10GE PHY 2.1 5.5 4.6 3
Total 76.4 76.6 70.2 50 55-60M
Q207 Quake Revenue is for Sep 2006 only. Red number is my estimate.

Here are my takeaways:

  • The core transport business is still in secular decline. AMCC has good products here and has invested while competitors have not. It must be near a bottom.
  • The processor business had a massive shortfall. $8M of the $10M miss from April’s down guidance was this area as they expected to do $27M and ended up doing $19M. The company indicates almost all of this will come back, if you take them on their word. This is really the key question.
  • The Quake business isn’t generating the revenue it was supposed to. CEO Kambiz H. indicated “Large Price Reductions” took place in Datacom – I assume this is Quake. I have also heard Aeluros is doing well in this market with their CMOS PHY. Perhaps this technology has made it’s way into other product lines but the acquisition cannot be called a financial success. (see article from last year)
  • The storage market is an interesting long term asset. Kambiz referred to a competitor’s (assume Vitesse (VTSS.PK) ) lack of commitment to Raid-on
    -Chip (ROC) and that customers are seeking a credible supplier.

Explanations for the decline from Q207 were provided. Each point in the excuse/product matrix had even more detail than listed below. This was a particularly painful part of the call but I’ve tried my best to detail the data.

Excuse Processor Transport Non-Focus Storage Total
End Product Phaseout 2 2 6 0 10
End Product Transition 2 0 0 0 2
Inventory Corrections/Mergers 8 3 0 0 11
Demand Softness 4 0 0 0 4
Total 16 5 6 0 27

The objective of this exercise (I assume) is to identify what is coming back and what is not. After the ‘over-share’ of information the company indicated that the existing design wins should generate a normalized $67-68M in revenue, which is what you get if you add back in all of the shortfall except for “End Customer Product Phaseout”.

There was extensive discussion about future wins. This is all forecasting and I know from experience that such discussions are usually rife with confirmation bias. In order to understand the strengths of a company, you need to examine the losses alongside the wins.

Finally, the company committed to cut quarterly Opex 10% from $40M to $36M.


AMCC’s business now sells for less than Vitesse, which makes no sense to me given the risk still present in the situation there. Regardless of the detailed explanations, I firmly believe the company egregiously stuffed the channel and is in the process of correcting that mistake.

I know that AMCC has very good transport products and should be leveraged to any improvement in the Metro WDM space.

The storage business would be attractive to outside parties as this area is in a period of heightened M&A. Consolidation of suppliers provides a tailwind to all remaining participants. Either situation should be positive for AMCC.

The processor business is extremely competitive and perhaps AMCC can find a niche (Wimax CPE and Printers don’t sound like the same niche). Companies like Cavium (CAVM), Broadcom (BRCM) and Marvell (MRVL) have very successful category killers. AMCC needs a category killer and I don’t know what it is.

AMCC is the poster-boy for communication semiconductor industry consolidation, a core Nyquist thesis. It’s three product lines of roughly equal size do not share common customers at they system level and there are few efficiencies gained from keeping them together. Vitesse, Mindspeed (MSPD), and PMC-Sierra (PMCS) share the same structure. The combination of one or more of these entities is logical. Given AMCC’s large cash position and healthy balance sheet, they could fill the role of consolidator. Or as a caller from SAC Capital indicated, someone else might do it for them.

Author is long AMCC, holds an immaterial VTSS short position and has no position in any of the other companies mentioned.


Comments are disallowed for this post.

  1. My guess is that the 10GbE PHY droop is due to the move from Xenpak generation (with Quake PHY) to X2 (with Aeluros). I don’t believe the their current silicon in this space is competitive and the declines will continue.

    I think that this speaks to a larger question brought up toward the end of the aricle: the lack of strategy in small cap semi players. AMCC was not a 10GbE player. The did not understand the market and it cost them $60M.

    Without a fundamental corporate strategy, these companies are, in effect, hi tech whores who chase every buck that’s wafted in front of their noses leading to idiotic acquisitions, projects killed at tapeout (or worse), and hiring of ill-suited executives.

    I believe that there is a lot of value in these companies and under more seasoned management this value could be realized. It would require drastic action however.
    couple of my thoughts…

    – figure our what your company is about

    – sell off divisions that don’t fit in that strategy

    – stop investing precious R & D $$ in dead product lines. this one is especially difficult for most as these sectors are likely large contributors to the current top line. in this business, however, it seems to me that a large proportion of your investments are geared out 2 years. i guess its a lot like doing trapeze. you’ve got to let go of the swing in your hand if you want to make it to the next rung.


    Posted by m@ | July 12, 2007, 1:55 AM
  2. AMCC acquired a lot of units . VTSS developed most of its in house I guess latter should have a better and more thorough understanding of the direction and technology, though it comes with a large disproportional R&D cost.

    Posted by yf | July 12, 2007, 1:53 PM
  3. M@ – I think the battle for 10GE is far from over. I’m on the record as saying AMCC overpaid for Quake but I would not count them out at this point. This market will shift multiple times before it matures.

    yf – You could have a lengthy debate about which company wasted more money on acquisitions and not reach a conclusion. A better discussion is which acquisition was the best one.

    Posted by Andrew Schmitt | July 12, 2007, 2:20 PM
  4. Hi Andrew,

    Am I correct in reading the disclaimer at the bottom of your article that you are now SHORT VTSS (I didn’t know you could short the pink sheets) and LONG AMCC? When did this happen and why the change in strategy?


    Posted by phobos | July 12, 2007, 8:02 PM
  5. My short position in Vitesse is a personal hedge I never covered from years ago. I don’t consider it material and isn’t a reflection of my current opinion of Vitesse. Seeking Alpha & Yahoo require I disclose all positions.

    My position in AMCC is not near the magnitude of my previous Vitesse position. It shouldn’t be viewed as a one-for-the-other.

    Posted by Andrew Schmitt | July 12, 2007, 10:43 PM
  6. i’d be interested to see your math as to how VTSS is more expensive than AMCC. here’s my math:

    shares 218 284
    share price 1.2 2.8
    market cap 261.6 795.2
    debt 150 0
    cash -20 -264 (will burn 20 this qtr)
    Net EV 366.6 531.2
    Revenue 250 250 (a guess at steady state for both)
    EV/Sales 1.4664 2.1248
    ratio 0.737
    comparable price for VTSS at AMCC multiple 1.63
    –(what happens if AMCC is actually cheap at its current multiple?)

    as it stands today, VTSS also has a dramatically more efficient model with OpEx at 60% of AMCC for a comparable amount of sales (i didn’t take the $10-20M of restructuring that AMCC has coming out of its cash, though i could have). you may argue that VTSS has cut to the bone and we’ll see it a few years but no one really knows….

    you could also argue that VTSS deserves a premium directly due to it’s “various issues.” AMCC can survive a couple more years doing stupid stuff w/ its cash, VTSS won’t. the opportunity for a transaction that’s drives value is much greater w/ VTSS.

    it’s also tough to argue that AMCC isn’t fairly priced from a market efficiency perspective: it’s liquid, in a number of indexes, owned by large institutions, publishes financials, covered by a number of analysts, etc. they went through extraordinary detail on their last 2 calls – people know / have discounted the issues there.

    VTSS on the other hand: can’t be owned by institutions that can only own listed securities, is not an any indexes, has no coverage, no liquidity, no real financials w/o doing a lot digging (what the hell is net consumption anyway), etc.

    Anyhow, that’s what i’m thinking. Here’s a question for you:

    do you know who won the SFP biz @ CSCO?

    Posted by roger | July 14, 2007, 1:00 PM
  7. It isn’t clear how you get $366M EV for Vitesse.

    218M Shares x $1.20 = $261M (check)
    $261M $150M debt – $20M cash = $391.6 (you indicate $366M).

    Anyway, correcting for this doesn’t close the gap. This should.

    1. VTSS shares outstanding are higher due to PIK effects from debt.
    2. The 250/250 revenue estimate is skewed. Let’s not have a forecasting debate. Use 12 month trailing. You also do not know what next Q will look like for Vitesse. I would argue that AMCC channel stuffing is offset by Vitesse aggressively selling inventory to generate cash (as they should)
    3. I include debt prepayment penalties.
    4. I don’t use forward looking metrics.

    Elaborate on what you mean by SFP biz. SFP 1/2/4G or SFP ? SFP is irrelevant for the next 12 months… if you believe in a truly efficient market, which I do not.

    Posted by Andrew Schmitt | July 16, 2007, 12:04 PM
  8. so, let me get this straight: you want to compare a peak yr where AMCC clearly stuffed the channel to a yr where VTSS went through enormous upheaval? i am not trying to be critical, but that seems to be a little nonsensical–after all, at the margin, how much biz did AMCC get because people thought VTSS was going under? It’s not about forward metrics, it’s about ballparking where both companies are today. additionally, looking at trailing ignores 2 key facts: 1) VTSS has one of the best growth semi divisions (storage) in all of semi ==and it did not really scale until 2007 and 2) AMCC has lost 2 key customers forever (NOK due to merger, only one survives RACK due to the fact they are dying)

    VTSS selling extra inventory generates cash, does not generate extra revenue. In my math, I had been giving them credit for sale of substantial excess inventory and the potential fab sale, before taking that out at the last minute, as it was unnecessary to the argument (hence the $25M difference – sorry).

    the PIK is paid in extra debt, not shares, and it’s not a big number.

    Including prepayment penalties is interesting but is a rounding error in this exercise… do you then include the 30M shares that AMCC has w/in striking distance of today’s price? It’s not like that I’m sure you’ll readily acknowledge that VTSS has far, far less of an option overhang at 1.18…

    as to market efficiency, the argument isn’t that it’s efficient, it’s that AMCC has alot more characteristics that lend themselves towards efficiency than does VTSS.PK

    was asking re: SFP 10GbE @ CSCO….don’t know the answer, thought you might

    Posted by roger | July 16, 2007, 8:24 PM
  9. The number one priority at Vitesse from dawn till dusk is to generate cash, particularly by monetizing existing inventory. If they were managing the revenue intelligently (I firmly believe they are) this will drive revenue forward. Think of it as the Wal Mart Automatic markdown. People may criticize such a practice but the number one priority is to generate cash. This activity will drive revenue growth, but not unnecessarily organic growth. I have no way to estimate this effect (you can’t just look at inventory depletion).

    Like I said I am not going to get into a forecasting debate. The fab sale is a fair add in. The options issue you mention is true as well. And I agree that AMCC should command a more efficient price. We could argue skillfully back-and-forth and maybe you are right and Vitesse is cheaper than AMCC.

    But, at the end of the day, I think there is a lot more that could explode at Vitesse than at AMCC. I will re-state what I said – my opinion about AMCC is much less vehement than my opinion was about Vitesse nearly a year ago.

    On SFP, 10G SFP is actually SFP-plus, and as I said before any volume there in the next year is not going to be material. The market will disagree with me and attach significant importance to whomever ‘wins’ – so much for an efficient market. Vitesse has won all of the EDC sockets for SFP-plus but it is a first gen design than will be shortly replaced. Vitesse can win the replacement but there are at least 5 other players angling for the same business. The odds at this point prevent anyone from legitimately naming the winner. On the module side, it’s tough to say who ‘won’ because the volumes are based on a single system that will be shipping low volume. I have much more in depth opinions here I will not disclose.

    In the end, Cisco will buy most of the volume and screw the supply chain. The low cost supplier with the most fixed cost leverage will win.

    Let me ask you a question – what do you think the Vitesse storage division is worth?

    Posted by Andrew Schmitt | July 16, 2007, 8:56 PM
  10. Andrew, what do u think of the VTSS July CC. i believe VTSS is alive & well. Current Qtr backlog is 15% higher than 3rd qtr. Q3 was strong qtr

    VTSS Financial Metrics

    3Q07 $60.5 mil Consumption
    CASH $21.2 mil, $4.8 mil Increase

    A/P $13 mil, Down $3.6 mil.
    A/R $12.7 mil, Increase $2.6 mil
    Inventory $31.8 mil (Down $7.8 mil – $4.2 mil write off, $3.6 sell off)
    Legal Fee 3.7 mil. Increase $200,000

    OpEx 22.3 Mil.

    Gross Margin 55-60%
    R&D 25-28%
    SG & A 11-14%
    % of Consumption
    NPD 54%
    EPD 22%
    SPD 24%

    % by Region
    N. America 50%
    Asia 38%
    Europe 12%

    Posted by alphnasx | July 31, 2007, 12:45 AM
  11. It was a very good result. I asked the question clarifying the 15% backlog Q/Q number.

    Posted by Andrew Schmitt | July 31, 2007, 8:16 AM
  12. i was thinking 55 – 57 mil for last qtr. CRG did mention some market were weak & some market were strong

    Posted by alphnasx | July 31, 2007, 9:22 AM
  13. Andrew; A rumor for a while, what do you think today’s deal of VTSS selling a portion of storage asset?

    Posted by FC | August 24, 2007, 1:30 AM