Reading the Earnings: The Data Center

Tech earnings continue to give us some interesting data points, and possible contradictions, in the overall tech space and in the networking space.  We also had some M&A, so let’s get to it.

Intel beat Street estimates in both revenue and profit, largely on the strength of business purchases of PCs and servers.  It’s Atom chips, largely targeting the netbook space, lost ground because netbooks are being eclipsed by tablets, a form factor where Intel has essentially no presence at the moment.  Server chip growth was the brightest spot, which shows that IT investment and data centers are getting some corporate attention.

Intel’s challenge is obviously the tablet and smartphone space, where the company has been investing to create a position.  The difficulty Intel faces is that it’s behind the curve here, having not predicted the smartphone craze and not predicted that smartphones were a leading-edge indicator of a broader network-appliance model of consumer electronics.  Yes, Intel can get positioned for the mainstream of the appliance market, but it will miss the opportunity to get in on early higher margins and also the opportunity to make itself a default brand, as it has with the PCs.

Related to Intel’s success was a decent quarter from F5, the company that is most identified with “data center networking” in our surveys, more even than Cisco.  F5 shows the value of specialization, but it may also be showing its downside because the pace of growth for the company appears to be slowing.  One reason may be that the data center network implications of virtualization, server consolidation, and the cloud appear to come in a kind of “wave-then-eddy” form.  First, the enterprise re-architects to create a flattish structure that can do load balancing and fail-over.  Then, as server count grows, it populates more of the devices with ports to accommodate the new devices.  The biggest revenue kick comes in that initial phase, and for the markets F5 has focused on (financial, government, etc.) the initial phase is now passing.  The company’s more limited product scope and sales presence makes opening new verticals more of a challenge.  They may be indicators of the issue Juniper will face with its QFabric product set, as that rolls out late this year; the old verticals may not be the best ones by that time.

Dell’s decision to buy Force10 to create its own networking product line to augment/replace OEM deals with Aruba, Brocade, and Juniper raises even more interesting data center issues.  It’s pretty clear that HP and Cisco are driving this particular decision; both companies offer networking and servers and thus a stronger and broader portfolio solution for the data center.  Buyers, more than ever, want to move into a major data center restructuring with a single player providing the critical tools.  For Dell, a purchase of Juniper would be unlikely; too much of the latter’s value is in the carrier space where Dell has less presence and focus.  Brocade’s storage products may be seen as contaminating that particular potential acquisition, and Aruba was likely not seen as a broad enough strategy.  Thus, Force10.

Obviously this raises questions about what IBM will do.  IBM, like Dell prior to this deal, is competing with HP and Cisco with an OEMed networking portfolio.  Will IBM now do a buy?  If so, will they pick Brocade as a cleaner (and cheaper) data center play or opt for Juniper because the latter has a strong carrier position that would mesh well with IBM’s cloud strategy?  We don’t think IBM is under the same M&A pressure that Dell is because IBM has such strong account control and acknowledged integration skills.  It might want better margins on its networking, but it also likely doesn’t want to upset the momentum of the moment with an M&A that would alienate at least one OEM partner and potentially strand some existing and developing accounts.  I’m on the fence here, for now.




Leave a Reply