I was recently asked in an e-mail about my opinion on PMC-Sierra. Here's the repsonse I gave:
To me, PMC-Sierra's like Cisco and EMC in that I'm fairly sure that it's going to perform very well over the next couple of years, but I just don't feel comfortable holding it. On the Layer 1 side, PMC-Sierra dominates the WAN Access chipset market. That is, the chipsets that go into T1 and T3 (DS-3) equipment, E1 and E3 equipment, OC-3 and OC-12 connections, and so on. While most of the industry's focus has been on high-speed OC-48 and OC-192 connections, the kind that Applied Micro and Vitesse make chips for, the lower-level market segments have also been seeing tremendous growth. After all, all the backbone traffic has to be coming from somewhere; and contrary to what one might think, this market's far from commoditized. Advances are still being made when it comes to offering more integrated, cheaper, low-power solutions, and PMC-Sierra leads here.
Meanwhile, at Layer 2, they're the undisputed leader in the ATM switch chipset market. While IP is the transport platform of choice at the network core, and although ethernet dominates in the LAN market, ATM is still widely used in the metro/WAN access arena, and the growth that this technology's seeing is still tremendous. I think that well over 30% of PMC-Sierra's sales, perhaps over 40%, come from the ATM market, and for the time being, there doesn't seem to be much of a slowdown in sight.
Then way am I hesitant to buy in? Well, for starters, I'm very cautious about the T1/T3 and E1/E3 markets going forward. These are very old, Bellhead-driven technologies that are going to get disrupted by newer, cheaper solutions like ADSL, VDSL, OFDM, LMDS, MMDS, and "wireless fiber" solutions from companies like Terabeam. All of these technologies will offer far lower $/mbps when compared to T1/T3 and E1/E3. Meanwhile, as more and more fiber gets deployed into metro markets, a lot of companies might opt for direct gigabit ethernet solutions.
This brings me to my next concern, which is the coming fall of ATM. Right now, everyone's looking at 10-gigabit ethernet to be a major threat to fibre channel. However, it seems to me that the technology 10-gigabit ethernet will be most detrimental to is ATM. When combined with the advanced QoS features that are now being incorporated into ethernet standards, when 10-gig ethernet rolls out, it'll make absolutely no sense for any carrier to use ATM anywhere. Granted, PMC-Sierra's attempting to be a major player in 10-gigabit ethernet as well, but so's Broadcom, Applied Micro, and numerous other companies. It's very unlikely that PMC-Sierra will be able to get anywhere as dominant a position in this market as they have in the ATM market. Furthermore, the deployment of 10-gigabit ethernet and gigabit ethernet over copper could eat into the company's OC-3/OC-12 SONET sales. While ethernet can be run over SONET, the products often won't be the same as the ones PMC-Sierra currently sells for present-day OC-3/OC-12 ATM/SONET and packet-over-SONET solutions, and a lot more competition will exist.
Of course, PMC-Sierra is rapidly expanding into new markets, and some of their new offerings include OC-48/OC-192 SONET and DWDM products, switch fabric chips, network processors, VoIP processors, DSLAM chipsets, and even 3G base station chipsets. However, all of these markets are going to prove quite competitive fot the company as it goes up against entrenched leaders, and for the time being, the bulk of PMC-Sierra's sales still come from traditional WAN access markets.
These markets are still growing very rapidly, and I expect them to continue to do so for quite a while. In fact, there's a good chance that one year from now, PMC-Sierra's stock will be far higher than where it is today. However, with so many good investments out there, I feel more comfortable passing up on this one for the time being.
Eric |
