Seeking Alpha

Mark Castellani's  Instablog

Mark Castellani
Send Message
Recent INSEAD MBA graduate Please refer to my Linkedin profile: Follow me on Twitter @MarkCastellani and contact me at or
My blog:
Linkedin Profile
  • Fusion-IO: Clearing Some Of The Misunderstandings Between Wall-Street & Management 0 comments
    Dec 12, 2013 9:36 AM | about stocks: FIO, EMC, RVBD, STX, SNDK, CSCO, WDC, HPQ, IBM, NTAP

    My intention is to clear up the confusion between Wall-Street investors

    My clarification is in the form of quotes from recent management speeches; then I would comment on how investors might have misunderstood management and then followed by my clarification.

    I will post multiple posts given the word limit. Stay tuned. I hope it helps

    Quotes from UBS Global Technology Conference (11/20/13) - UBS Analyst: "Oh, speaking of architecture, when you look at some of the new architectures that are coming out -- scale out, etc. We hear things like, well, if you need to scale-out architecture with a Hadoop, per se, type situation -- not in Isilon, necessarily -- the need for Flash at the server decreases."

    Lance Smith - Fusion-io - President and COO response to the UBS Analyst "So the first thing is you said something that was kind of conflicting. Hadoop, Cassandra -- these are very much scaled-out types of architectures. And when we're scaling out, we're putting a lot of servers out there."

    My clarification: The UBS Analyst concluded that the adoption of Hadoop would be an impediment to Flash growth. His premise is the following: "Since Hadoop would lead to less Flash per server, Fusion would sell less Flash". While his premise is correct, his conclusion is far from; the benefit of Hadoop is that a customer does not have to put multi-terabytes worth of Flash in each Hadoop-based server. The customer would use less number of terabytes of Flash in SOME of these servers BUT they would put the Flash in all of her/his servers, including non-mission critical servers (i.e. it wasn't cost efficient to add Flash to non-mission critical servers in the past; now why not mentality). So rather than sell more flash per server, Fision would sell a lot more Flash per customer since only a fraction of percent of Hadoop servers use Flash (i.e. Hadoop is a driver of Flash growth rather than an impediment to it)

    UBS Analyst: "And I think when ioScale first came out, you guys argued that it's actually not going to be dilutive to margins, despite being a lower-end product, which originally was true. Now we're talking about margins going into the low 50s"

    Lance Smith - Fusion-io - President and COO response to the UBS Analyst "the way we designed ioScale is to be cost-optimized, not to be low gross margin. So we had to do some trade-offs. So ioScale is not the same product as ioDrive. Although it uses the same components, what we've done is we put twice the amount of Flash on a single controller. we came to market with ioScale, especially through the OEMs"

    My clarification: UBS Analyst concluded that introducing lower priced product (i.e. ioScale) would dilute earnings. However, what the analyst missed is that iOScale uses a single controller - it is designed to be cost efficient for customers who need a moderately priced solution for non-mission critical servers (i.e. it has lower cost of goods sold (COGS) in-order to maintain overall gross margins). In addition, iOScale is brought to market through the indirect channel (OEM), meaning lower operating expense (i.e. higher operating margins and higher net income) since Fusion would not have to pay expensive direct sales force (i.e. scalable product, leading to higher absolute operating income [$EBIT per product X more products without incremental fixed costs)

    UBS Analyst: "So it sounds like you're becoming more dependent upon OEMs for distribution. Is there any conflict there in terms of the more Flash you sell, the fewer servers you need? Or might some of them"

    Lance Smith - Fusion-io - President and COO response to the UBS Analyst "So I would say early on in the day, those were kind of the push-back that we got with the OEMs. Like, oh, wait -- if I use Flash, I will only sell one server instead of 10 servers, because you could do that kind of consolidation. They see higher ASP with the Flash -- that they are more fully-loaded and more full-featured servers. So over the years, what they have now seen is that it's a benefit to have Flash there. Because we are not selling the Flash; they're selling the Flash. And at the end of the day, they're getting a larger ASP sale. So they either sell a lot of cheap servers, or kind of one medium-fast server. And so we've seen that sort of benefit in that trade-off from that perspective. When you have the likes of HP, IBM, Dell, Fujitsu, and even Cisco -- these guys have good footprints, either in the networking side or in the storage side, or on the compute side. When we come up with full solutions for them -- and we have a couple of them; we have ION Data Accelerator. ION allows someone to build a full Flash array -- an all-Flash array, or a caching front end for a back-end stand, so it makes it very transparent. You don't have to worry about having a software-aware storage solution. And it also allows to be built into the channel. ioControl will address the small/medium enterprise, but it changes the economics of Flash.

    We combined Flash with hard disk drives so that a small/medium enterprise now no longer has to do a deployment to multiple storage solutions for every single workload that they've got. They typically will have a storage solution for, let's say, their email; for something like Microsoft Outlook.

    They'll have a storage solution for their CRM, or their front-end business for their customers. They'll have a back-end storage for their ERP and have another storage for archiving.

    This storage sprawling is what's being solved by ioControl in the acquisition we did with NexGen, because now somebody -- they can do one deployment, and they can say, look; I've got an application with performance requirements, but you don't need all-Flash array. You just need enough Flash to make it run fast."

    My clarification: The UBS Analyst assumes a channel conflict (channel conflict, in the simplest form, is that you ask a 3rd part to sell your products that would lead to a less demand for that particular 3rd party's products). However, as Fusion COO clarified, is that the OEM is actually seeing a benefit of pushing Flash to end-users in the form of much higher Average Selling Price (NYSE:ASP) per server. You might ask "Would the incremental Average Selling Price make-up for lost OEM server volume" My answer is (close), and make-up or not; OEM has no choice than to push for Flash adoption, given the influx of data volume, end users are demanding much higher performance servers (i.e. with Flash servers)

    What Would You Infer from the Sequence of Events Below?

    Shane Robison - Fusion-io, Inc. - Chairman, CEO (November 07 2013): "We've got work going on with Cisco, which is really exciting"

    Fusion-IO New Release (December 3rd 2013): "013 - Fusion-io (NYSE: FIO) today announced that Ted Hull, previously vice president of finance at Cisco since 2007, has joined Fusion-io as executive vice president and chief financial officer"

    Ted Hull Linkedin "VP of Finance - CISCO - Functional CFO for Global Sales, Service, Channels, and Marketing - strategic pricing, deal management, large solution deals

Back To Mark Castellani's Instablog HomePage »

Instablogs are blogs which are instantly set up and networked within the Seeking Alpha community. Instablog posts are not selected, edited or screened by Seeking Alpha editors, in contrast to contributors' articles.

Comments (0)
Track new comments
Be the first to comment
Full index of posts »
Latest Followers


More »
Instablogs are Seeking Alpha's free blogging platform customized for finance, with instant set up and exposure to millions of readers interested in the financial markets. Publish your own instablog in minutes.