Fusion-io, which provides solid-state storage for so-called hyperscale data centers, has delivered some strong growth courtesy of customers such as Apple and Facebook. But these deployments can be a bit lumpy quarter to quarter.
Fusion-io's outlook highlights the lumpiness and indicates that it needs to diversify its customer base. Fusion-io's December quarter was fueled by Apple and Facebook, two companies that accounted for 50 percent of sales. "Facebook, Apple and HP combined represented 69% of total revenue; and 34%, 16% and 19% respectively," said Fusion-io CFO Dennis Wolf.
Fusion-io reported fiscal second quarter earnings of $1.7 million, or 2 cents a share, on revenue of $120.6 million, up 43 percent from a year ago. Non-GAAP earnings in the quarter were 13 cents a share. That's the good news.
The problem: In the next two quarters, Apple and Facebook won't be buying as much from Fusion-io since they can squeeze more efficiency out of what they have. As a result, Fusion-io said that third quarter revenue will be about $80 million with a non-GAAP operating loss of $10 million to $15 million. Few observers saw that projection coming. For the fiscal 2013, Fusion-io said revenue will be about $420 million to $440 million.
Simply put, Apple and Facebook have stocked up their data centers with solid state flash storage and have paused spending. The good news is that in six months that spending is likely to resume.
The even better news for Fusion-io is that it's likely to diversify its customer base due to trends in big data and the reality that more data centers will have solid state storage. There may become a time when spinning disks in the data center are a relic.
Sure, the biggest worry with Fusion-io is that its well-heeled competition catches up and poaches sales. On a conference call with analysts, CEO David Flynn noted that the roughly six-month indigestion period for both Apple and Facebook purchases is part of the solid-state data center growth curve.
It's not so surprising from our point of view when we look at how Apple and Facebook moved along. We worked with Facebook from when we were less than 1 year old as a Company, grew alongside them and watching how they learned to leverage solid-state technologies in their infrastructure. These things simply do take time and have major implications. It is a huge transformation in the capacity of the data center and its efficiency levels. So, the fact that there's --takes a little longer because they're looking at how things are re-architected, it also -- they're having to optimize and decide, well how much does it take? We had one customer tell us recently that their estimations were wrong, that they were able to determine that they could use one-third fewer servers than what they had even previously calculated because now they were able to get additional throughput from it. Now, luckily they have a very, very large deployment, so that is not a material impact to the size of the thing, but it is indicative of the fact that they are still tweaking and tuning and extracting every last ounce of efficiencies that they can get.
And then Flynn mentioned Apple's dynamic with its data center.
Facebook is more open about this kind of stuff in that we work with them, but even there, I would be loath to make assessments about why they might be focusing on efficiency in the cost of their business, but our guys could probably infer that and of course, Apple is extremely private about it, so I won't really make any speculation on that point. What I will say is, though, if you look at those two customers and how they have performed in the seven quarters since we IPOed and look at the relative sizes to each other, they have tended to be oscillating to where one does 70% while the other one does 30%, and then the other one does 30% and the other one does 70% of the pool of strategic. So, the variability in their purchasing patterns is inherent. What you have here is the opposite of the law of large numbers. The law of smart numbers means that things don't always line up. And so I think that's maybe a better way to think about this. There is going to be things pushed in a quarter or pulled in a quarter or pushed out of the quarter. Sometimes they offset each other, and that has been the case up until now.
Analysts didn't like Fusion-io's outlook and openly fretted about Apple and Facebook demand. By the way, the customer concentration is disclosed in regulatory filings. Most analysts, however, think Fusion-io can diversify in a few quarters.
Benchmark analyst Gary Mobley stuck with a buy rating on Fusion-io and said:
First, a diversified set of customers should help drive a revenue recovery by 4Q13 (June) with Apple and Facebook no longer representing ~50%-60% of Fusion-io’s revenue. Second, we believe Fusion-io would make a great strategic fit with storage system and server OEMs such as HP, EMC, etc.
Mobley's bet is that even if Facebook and Apple don't ramp spending again, Fusion-io can benefit from new customers and partnerships with the likes of Cisco and NetApp.
Overall, Mobley's bet sounds fairly realistic---unless you believe that solid state storage is done in the data center.