Video is Driving 10GbE

This is by far the most over-used and misdirected belief and causes me physical pain when I hear it. Yes, video is playing a major role in the expansion of WDM transport networks. Yes, most of these new links are 10G. But much more growth is coming from the datacenter as bladeservers and computing infrastructure transition to 10GbE. Companies like Broadcom are driving down the cost of putting 10GbE in servers, and subsequently will drive demand for 10GbE switching equipment. Other trends such as re-using the 10GbE standard for datacenter storage interconnect amplify this trend.

At best, 100k new 10GbE connections will be deployed in the core network in 2007. Compare that with the 400k or so modules Cisco alone will sell in the same timeframe. In 2008 the gap grows much larger.

But everybody understands YouTube, so video remains the lip-synching poster boy.

10G is One Market

Few bother to drill down into the innards of module types and specifications to understand product mix, and are quick to dismiss the alphabet soup of module types and laser reach specifications. This is a huge mistake as the majority of unit volume gain in the next 5 years will come in a narrow segment of product and reach types.

The 10G optical market is really several different markets with little or no relationship to each other in terms of cost, market demand, competitive pressure, and growth. Little things, like the tendency for enterprises to deploy OM-3 fiber have an enormous impact on product mix.

Incumbent 10G Suppliers Will Benefit

Some existing suppliers will do well, but the mistaken view that 10G is one market drives investors to believe all companies supplying this market will benefit.  To borrow a much abused term, the distribution of growth in the 10G market is ‘lumpy’.

Opnext (OPXT) makes great hay about the importance of the 10G transition to their business and justifiably so given they are top supplier to Cisco for all types of 10G modules. While it is impressive that Opnext (as well as Intel (INTC)) supply the bulk of the volume today, it isn’t clear that they have adapted their supply chain for the demands of tomorrow.

My belief is virtually all unit volume growth will be in the low-cost, short-reach segments of the market, a market characterized by brutal price competition and 20-30% gross margins. Vertically integrated companies will be positioned best when volume ramps. Opnext is a somewhat vertically integrated manufacturer of long wavelength modules but heavily outsources both components and manufacturing of short reach modules.

It would not surprise me to see Opnext acquire a VCSEL manufacturer itself to remain competitive as the short reach market moves to higher volumes. I also believe this was the primary justification for JDS Uniphase’s (JDSU) acquisition of Picolight. (see “Why JDSU Bought Picolight“).

Such an acquisition still leaves both companies with a cost structure unsuitable to high volume commodity production.

I believe Finisar (FNSR) is the best example of a company well adapted to such a commodity environment as they are by far the most vertically integrated manufacturer of these types of components.

It will be interesting to see how all companies adjust in the volume shift in the 10G market. Perhaps Opnext decides to focus exclusively on long wavelength products, like they already have at 2.5G.

SFP+ is a key growth driver

SFP+ is a new low-cost 10GbE form factor and is a hot buzzword. It will indeed have a major impact on the market and is the key battleground in the competition between optics and copper interconnect. Industry insider consensus is SFP+ will not be ready for prime time in 2008 and I agree.

This is one trend where I admit I could be wrong as Cisco (CSCO) could change the whole equation quite abruptly. Cisco itself indicates only a small fraction of the modules it buys in 2008 will be SFP+, but doubts linger in my mind.

As readers know, I believe Cisco derives enormous profits from reselling optical modules. The transition to SFP+ is critical to Cisco for two reasons.

1. Lower cost modules mean lower cost 10GbE ports. Driving down 10GbE port costs is the most important vector for driving 10GbE adoption. Cisco must leverage their hegemony in 1GbE into 10GbE and not allow others to gain share. If they can drive costs better than anyone else they will capture the market share.
2. Cisco will make significantly more money reselling SFP+ modules than the existing X2 modules. Pulling in this transition just one quarter drops $25m of pure profit right to Cisco’s bottom line.

I am nervous betting against Cisco’s desire to make money and retain hegemony.

Copper (10GBase-T) isn’t a factor this time

Jury is still out on this one. One thing for sure- it shouldn’t be a factor until 2011 or so and faces significant technology headwinds. 1GbE optical component growth stalled in late 2004 as it could not match the per port costs of Gigabit Ethernet over copper. We shall see if this time is different.

Author is long Finisar.

… and I vow to torture a small animal in my office every time I hear or read the words “Video” and “Explosion” used in the same sentence. You, your Marcom group,  and your conscience have been forewarned. (Just Kidding.)

Click through to the full post to view and comment.

One thing I didn’t mention – I am just as guilty of irrational exuberance in 2000.

Here’s a top 10 list of my most notable observations from the Executive Forum at the Optical Fiber Conference in Anaheim last Monday.

1. When the Carrier panelists (Verizon, AT&T, BT) spoke of SONET/SDH they used the past tense. “When we used to build a SONET/SDH ring”. They appear to have subconsciously moved on.
2. According to Ryan Limaye from Goldman Sachs, Sanford Bernstein issued a report analyzing what would happen to the Telcos if they became a dumb pipe. The result? Much smaller but more financially efficient companies with a better return on equity. All of the carriers with the exception of British Telecom were not open to the idea of being a dumb pipe.
3. Observation: Equipment guys are asking for 40Gb/s components, and component vendors are giving them the finger. The fact that component vendors no longer allow equipment vendors to outsource R&D with a low ROI is a very positive development for this business. People are finally saying no.
4. 40Gb/s demand pressure is from carriers who deployed the Cisco CRS-1 router and need something to hook it up to. Stephen Carlton from Fujitsu kicks in that 40G muxponders make up the other 50% of demand. Many speculate 40G will be bypassed altogether in favor of 100Gb/s Ethernet using new modulation schemes.
5. Jerry Rawls from Finisar (FNSR) - “It is impossible to deliver the level of R&D expected by customers given current margin levels”. “Custom Tweaked Requirements are hard coded into the Telecom psyche and may be impossible to remove”.
6. Fariba Danesh from Avago (private) - ”Finisar pricing is making things difficult for us” (I didn’t get a precise quote… but this definitely was the jist).
7. Jo Major from Avanex (AVNX) – “Huawei still tends to buy at the component level.” Something I’ve covered in depth (see Dr. Strangelove, Or: How I learned to stop worrying and love Huawei)
8. Mike Nishiguchi from ExceLight – “Optical Modules are footprint compatible but the components that go into them are not. Why? We should think about this.” On outsourcing (heavily paraphrased): “The Japanese are not good at giving people titles and job descriptions. This is because they are not good at partitioning responsibility. And this is why I believe Japanese companies are not good at outsourcing their manufacturing”.
9. Gary Wiseman from Intel (INTC) Promises $150 SFP+ module in 2008. Claims 10G-BaseT will be delayed due to 2.5us latency, 6W of power, and new cabling. He clearly wants to keep copper interconnect out of 10Gb/s.
10. Stephen Carlton from Fujitsu (FJTSY) – “Don’t design your network for Video because much bigger drivers exist in the near future. Japan is trailblazing these applications with a cashless society, hyper-mobility, and other application innovation. Their greatest lead isn’t in FTTH, it is in application innovation.” Example: Take a photo of your sushi and your mobile will tell you when it was caught and what boat it came from. Stephen’s comments were unique and a refreshing break from the “Video Broadband Explosion” cited again and again by many talking heads.

Many of these quotes are paraphrased though I feel I captured the intent of the speaker. If I did not, please speak up. Some of these topics have follow up articles that I may release publicly.

Full disclosure: I am long Finisar.

Quantitative Data

• Consumption was 60-62M. Historically I take the midpoint number for all calculations, yielding the following product line growth.
• Opex cut radically from 33M to 28M. If I had realized the magnitude of the cut during the call I would have asked what projects were rationalized to make this happen. Opex was $42.7M 12 months ago, a reduction of 34%
• Capex was $2.6M, $1.4M for lease and $1.2M for masks (low)
• Cash at 16M, down 9.6M. Cash burn of 4M due to expanding inventory, 4.8M professional fees. Payables grew 1.4M too, so actual burn was around 11M.
• I have questions into the company about interest payments.

Qualitative & Colorful Analysis (because investing in Vitesse will drive you insane without it):

• Increased ‘consumption’ (Vitesse speak for revenue until an auditor is hired and they can remove their financial training wheels) was the best news of the call. While the growth rate was not as good as last quarter, considering the struggles of others in the industry, this was a good result.
• As before, no margin guidance was given. The company indicated distribution inventory with channel partners like Nu Horizons (NUHC) was drawn down substantially. As readers of this blog know, nothing kills margins faster than greedy use of distribution. A reduction of channel inventory will impact cash in the short term, but long term will enhance margins by at least 3-4 pts. I will wager their gross margins are in the 58-59% range. As stated in the past, Nu Horizons is facing a big headwind as Vitesse rationalizes it’s business.
• If my margin assumptions are correct, the company is break even excluding professional extortion expenses.
• Please, not another quarter of multimillion dollar professional expenses. Just think of how those expenditures are viewed by employees on the inside; Cash controls must be tight to the point where you can’t get paper for the laser printer, but $5M gets paid to guys who are 10 IQ points away from “Would you like Fries with That”.
• The best thing I heard was the imminent board changes. Mayor McCheese would be an improvement.
• Cash is tight but company indicated inventory growth is over, and professional extortion expenses are rapidly declining. If revenue stays flat, and the company can hit these two targets, they should be cash flow neutral. Executing this plan should be the most important goal for the company and I wish I heard that on the call.
• Brief but significant mention of additional financing. I don’t know what to make of this other than their customers are not comfortable with a razor thin cash margin. My guess is they are looking to open a line of credit. This is fine by me so long as no temptation exists to spend it.  If Nu Horizons can finance inventory at LIBOR +1-2% I see no reason why Vitesse cannot. I hope that this is what the company does, and I sure hope they don’t use it.

Summary

In an environment where comm semi companies briefly emerge from the surf before being pummeled along the post-bubble reef of telecom wreckage, Vitesse appears to be treading water against tough odds. I’m much more positive today about the company than I was last August at a lower market price.

Key Priorities for FQ207 (Unfortunately, two weeks from closing)

1. Reduce cash burn by eliminating professional waste fees and burning off inventory.
2. Aggressive pursuit of M&A of valuable units such as Ethernet in order to catalyze industry consolidation. Placing the assets of this group in a larger, better funded existing operator would be a win for everyone.
3. Hire an auditor. Get audited results. Plead for mercy from NASDAQ.

Original Post Published at 10:40AM.
Previous Relevant Postings:

• Q306 Conference Call Notes
• Vitesse Investment Thesis
• Q306 Conference Call Notes
• Trickle Down Economics and Channel Stuffing

Author is long Vitesse and short Nu Horizons

Flat Revenues

The flat revenue guidance for FQ407 of $104-$110M was particularly disappointing after a flat FQ307. This revenue plateau is entirely due to 11% quarter over quarter reductions in multimode LAN/SAN revenue. Other product areas made up for the shortfall resulting in flat top line revenues.

Sales of LAN/SAN 850nm VCSEL based products dropped from $62.6M in FQ207 to an expected $49.5M in FQ407, a reduction of $13M in revenue.

Finisar is the leading provider of multi mode optical modules in the world. This is a very big drop, something that the financial community isn’t questioning sufficiently.

The culprit, as explained to me by Finisar, was an inventory correction tied to 4G Fibre Channel SFP products. At one point these modules were on allocation and the company witnessed a classic case where customers double order in order to assure a large allocation of scare product.

The sheer magnitude of the shortfall ($13M) seems awful large to be tied to such a narrow product area, though I have received assurances from the company this is the case. Is this inventory correction prolonged by a move to 8G Fibre Channel? What else is at work here? Comments and thoughts are welcome.

Metro and Telecom Revenue Growth

The strength in Metro modules, primarily high-end tunable transceivers, is very notable. This is not where I expect Finisar to take share since  their competitive advantages are in multi mode LAN/SAN market. This new strength comes not just from customers shifting away from 300-pin MSA modules, as well as taking share from companies like Optium (OPTM), (AVNX) and Intel (INTC).

The company mentioned component supply chain constraints as the reason they could not deliver sufficient 10G modules to meet demand. A little known fact is that Finisar, unlike all other module makers, has taken the additional step of vertically integrating their analog IC requirements. Instead of purchasing components from Mindspeed (MSPD), AMCC (AMCC) or Vitesse (VTSS.pk) Finisar designs and fabs their own laser drivers, clock and data recovery, transimpedence amps, etc. This is not an easy thing to do and I suspect was the cause of their 10G hiccups.

This in-house IC expertise makes them the most vertically integrated module maker in the world. Customers like Cisco certainly notice and appreciate the future pricing benefits this will bring regardless of short term problems.

$1M in Cisco X2 module revenue

$1MM in X2 module revenue at Cisco (CSCO) is a drop in the bucket. Cisco purchases about $40M in 10GE modules on a quarterly basis. This is not negative for Finisar, it’s an an indicator that the ramp at Cisco has lots of runway.

Wall St. does not understand the architectural transformations underway at Cisco and how this will impact the module supply chain. While all analysts agree this is positive for Finisar our internal research indicates they are underestimating the revenue impact.

 Author is long Finisar

JDSU paid $115m in stock for Picolight, roughly 2.5x to 3x Revenue. While not a huge amount of money, it is a respectable sum when compared to public valuations of other small optical component companies like Stratos (STLW) Avanex (AVNX), or Bookham (BKHM)..

Picolight is interesting because they are the last pure-play provider of VCSEL based modules, using technology developed in house. The reality is they are a laser supplier that was forced to package their products up in order to extract maximum value. Particularly interesting are their long wavelength (1310nm) VCSELs. Finisar (FNSR) is a major VCSEL supplier, and aside from acquiring 1310nm VCSEL technology from Genoa and Infineon, has yet to commercialize it.

Low cost, low power 1310nm lasers are suddenly are a lot more valuable with the advent of the 10G-LRM standard for transmitting 10GbE. LRM that can extract much more value from 1310nm lasers by using crafty EDC (Electronic Dispersion Compensation… but let’s call it crafty for the benefit of most readers) electronics to take 1310nm wavelengths and transmit them further over existing fiber. It is a crucial technology for reducing the cost of optics in 10GbE deployments. Vitesse (VTSS.PK) is a key EDC chip supplier into SFP= based EDC designs.

Picolight, like any good VCSEL supplier, will make the argument that VCSELs are inherently cheaper to manufacture than the DFB or FP lasers used in the majority of 1310nm applications today. JDSU, in purchasing Picolight, is betting their VCSELs will enable a cheaper LRM solution. VCSELs also have low power characteristics that make their use ideal for small form factor SFP+ applications.

What does not make sense is that JDSU has previously dismissed the optical components business and chosen to focus on subsystems and test equipment. But JDSU would not have acquired Picolight unless they were focused on fielding a competitive 10G LRM module. This indicates that JDSU senses an opportunity in the market that didn’t exist a year ago. They are in essence, admitting a mistake through their actions.

The fact that a company that publicly dismissed the optical module market and has now reversed that stance is very notable. Today, JDSU is very much not engaged with Cisco, the largest module customer, for next generation 10GbE modules. Perhaps the Picolight purchase is an admission that this situation must be reversed, and a bet that long wavelength VCSELs are the means to make it happen. JDSU just needs to convince Cisco that the technology is ready for prime time.

Author is long Finisar and Vitesse and short Cisco as a hedge.

This acquisition is surprising. Cortina previously focused more on the core of the network and has either built or acquired products designed to enable the next generation of WAN equipment for carrier applications. Immenstar is tightly focused on fiber to the home chipsets, a much more access oriented application. I suspect that Cortina is looking to couple it’s carrier class ethernet solutions with Immenstar’s high density OLT solution. Still, it is not a pretty fit.

Immenstar is a hybrid Chinese-American company with strong ties to UT Starcom, a company that at one point was poised to supply NTT (NTT) with large amounts of FTTH equipment. They stumbled (or were tripped by Passave, now owned by PMC-Sierra (PMCS) ) and Mitsubishi Electric took the majority of the lucrative and prestigious NTT contracts. As a result, Immenstar has not been tremendously successful in the marketplace. I suspect that this acquisition was Immenstar’s only option.

Cortina is one of the more interesting startups out there today, and appear to be pursuing a strategy of strategically integrating underperforming assets acquired from both public and private companies with a goal of consolidation. This is a core Nyquist strategy.

It’s unfortunate that companies like Centillium (CTLM) are not actively engaged in attempting to sell their own FTTH assets. As an owner of Centilluim equity I find this extremely frustrating. It’s clear that Cortina would have been an interested suitor.

Hat Tip: Venturebeat

Author is a Centillium stockholder

I’m not trying to start a debate about whether WiMAX will succeed. My cycling buddy would define success as 5M WiMAX subscriber adds in a single year (not total). What isn’t clear is how Intel benefits even if the large investment they are sinking into WiMAX pays off.

Intel has a poor history of developing any successful products in the communication space, let alone mixed signal communication. Historically, Broadcom (BRCM) and TI (TXN) (among others) are far more capable. Odds are these vendors would seize the infant technology from the cradle and produce a superior product.

Even if Intel manages to be a leader in WiMAX chipsets, the market opportunity is a pimple when compared to their existing $35BB business of mostly CPU’s and companion chipsets. Even if WiMAX chips generate $1BB, it is still noise. Intel lacks a whole suite of products to wrap around this hard-won hypothetical WiMAX market share to generate additional pull through revenue. They sold their mobile CPU business to Marvell (MRVL). PC CPU’s and chipsets don’t count because they would have won those anyway.

When I asked someone at Intel not involved with WiMAX this question, his response was that Intel sees ultra mobile computing as a key initiative, and WiMAX is the enabler. This makes sense if you ignore the fact existing wireless carriers don’t want WiMAX.

Virtually all carriers are divided between UMTS (HSDPA) and CDMA2000 (EV-DO) as evolutionary 3G technologies. They don’t care about WiMAX because they have a big installed base (2G) they must migrate. There is Clearwire (aborted S-1 excepts here), and Sprint (S ), which has committed to building out a WiMAX infrastructure. There is also Korea, which never met a communication infrastructure it didn’t love (ATM, ISDN, DSL, two flavors of FTTH, 3G – what haven’t those guys deployed?). But the opportunity is dwarfed by the size of the existing mobile voice and data standards.

If Intel was really committed to growing a mobile wireless business they would identify a product that could both fill their fabs AND meet the needs of existing carriers – not build something counter to their desires.

Here are my three possible theories:

1. Intel has concluded that WiMAX will be an incredibly disruptive technology with undeniable performance/price advantages. New carriers will rise as the incumbent carriers foolishly embrace their legacy technology. I don’t have the technical firepower to form a conclusion here, but I doubt that the Intel Borg-collective has figured out what many others have not. Perhaps too many WiMAX guys at Intel have read “The Innovators Dilemma” and jumped the gun.
2. The big push for WiMAX has nothing to do with making lots of money. Instead, it is an initiative that will allow Intel to build a new, separate corporate division focused on mobility and wireless, allowing them to compete in many other markets. As a culture, they prefer to lead rather than follow, so better to champion a new standard than chase competitors who are already established. Even if WiMAX doesn’t pan out, they can move their new wireless R&D Death Star into another market and start taking share.
3. Emerging Markets. There are huge untapped expanses of the world devoid of Internet connectivity and more importantly, Intel CPUs. WiMAX as a greenfield technology can bring low cost connectivity to these markets and enable the sale of more PCs. Existing markets already awash with connectivity just aren’t that important.

I think WiMAX is destined to be what I call a crack filler, an alternative technology that fills in the areas where incumbents are too expensive or not available. WiMAX will be to 3G what Satellite is to Cable – best case.

Either Intel is pouring hundreds of millions of dollars into a foolish endeavor or something is unseen.

Expect to see other semiconductor companies investigating their accounting practices. Board members of these companies are probably scrambling in an ass-covering frenzy as I write this.

Vitesse illustrated the way sell-in accounting can be abused, and has moved to a sell-through model since earlier this year. I think this will be an industry-wide trend in an effort to avoid the appearance of impropriety.

Sell-in revenue is recognized when inventory is shipped out from a component supplier, whether it be to an end customer, contract manufacturer, or distributor. Sell-through revenue is recognized only when it reaches the end consumer of the component. The difference is that companies that use sell-in recognize shipments to distributors as revenue, companies that use sell-through do not, unless the sale is non-cancelable/non-returnable. 

This allows companies operating under sell-in accounting to use distribution as a ‘buffer customer’. This has a number of advantages, allowing manufacturers of components to control inventory costs. It also can enable companies to ship product at the end of the quarter to magically hit Wall St. targets.

Component distributors like Nu Horizons (NUHC), Avnet (AVT) and Arrow Electronics (ARW) are the prime beneficiaries from the sell-in practice. Distributors make money marking up goods they receive from suppliers and reselling them to end customers. With sell-in they must pay their supplier when goods are received. With sell-through they pay their supplier only when they sell the goods.

Distributors demand additional margin from companies using sell-in because of the negative liquidity required, and subsequently derive much higher profits, albeit with significantly higher risk. This process has become more profitable with cheap and easily accessible liquidity.

I wrote a long post (see “Trickle Down Economics and Channel Stuffing“) examining the potential risks to Nu Horizons as Vitesse moved from a sell-in model to a sell through model.

Nu Horizons is particularly exposed to this effect because Vitesse is their #2 supplier, and Vitesse will no longer be using sell-in accounting. It is yet indeterminate what impact Vitesse’s accounting change and subsequent operational changes will have on Nu Horizons. It should be noted that Nu Horizons has issued a press release indicating they did nothing wrong, though they have not discussed what impact the Vitesse accounting change will have on their operating income.

If other, larger suppliers of specialty components move away from sell-in, this will have a permanent, long term, and negative structural impact on the P&L’s of distributors as their highest margin business will go away.

In an age of sophisticated just-in-time shipping and real-time inventory management systems there is little need for sell-in accounting. Sell-in can be abused by a supplier to manage revenue and earnings. At its worst, it makes engaging in accounting fraud significantly easier. At best, this leads to lower gross margins in exchange for less revenue and earnings volatility. Personally, I find such a tradeoff unacceptable.

Full Disclosure: I am long Vitesse Semiconductor and short Nu Horizons.

The shock and awe portion of the release were the 9 accounting sins committed by the company. I will not list them here, read the release.

The conference call in July provided a preview that prepared investors for this fact. Of greater concern to me was the lack of an explanation of what the future plans were for Vitesse.

In particular, the lack of audited results from FY04/FY05 now and forever leaves a big void with respect to the future intent of the company. Investors need to know what the companies objective is with respect to relisting on the Nasdaq. When can this be anticipated, or as one astute reader points out, are we looking at a future similar to Riverstone Networks? (Lightreading link)

I provided a more extensive and private opinion directly to the company, but this summarizes what I feel needs to be done.

1. Make a clear and concise statement on what the plan is as a result of not having 2004/5 financials. What operating model will the company seek?
2. Provide a release timeline for FY06 results based on what is known today. Previously, the company expected to release these by end of year. Now, it is indeterminate.
3. Provide timeline for the independent operation of the company. When are the consultants (and expense) gone and when do the new auditors come on-line.

It may be that my expectations for an immediate plan following the release of the report are unrealistic. If so, a date should be provided to investors when a go forward plan addressing the above issues will be available.

Such a plan will allow investors to benchmark the ability of management to manage the crisis to a conclusion.

I have other, stronger opinions about the board, but the reality is that management has little control over the makeup of board. It is up to the investment community to catalyze change there.

Full Disclosure: I am long Vitesse