There has been a rumor flying around the twitterverse and other social media outlets that Cisco is about to announce that it is buying Nutanix. This rumor started circulating when a report from a Portland-based equity research firm was released, opining that as a last post, John Chambers was going to buy Nutanix. Now, as good a story as this is, it is a story, as the analyst quotes no sources. He is, however, quite bullish about the possibility and argues confidently that an announcement will be made as soon as the .NEXT Conference in June.
On the face of it, it may seem that there would be a lot of synergy in a purchase, as Cisco would gain what is perceived as the market leader in the hyperconverged market space. This is a market that is booming, with companies like Nutanix, SimpliVity (an often-mentioned target for Cisco’s checkbook, though that has come to nothing), and Atlantis Computing. Cisco would also gain a valid storage offering that is a thorn in the side of its former VCE frenemies, EMC and its federation partner VMware. However, there are several very good reasons why Nutanix is not a good fit for Cisco and, more importantly, why Cisco is most definitely not a fit for Nutanix.
Cisco does not have a good record of integrating storage-based companies into its product portfolio—Whiptail, anyone? For those who cannot remember, Whiptail was one of the first providers of an all-flash array and was previously a privately held company. Whiptail designed and commercialized the use of NAND flash memory as a replacement for hard disk drives, driving the all-flash array storage business. At the time of its purchase in October 2013, it was one of the leaders in that space. When Cisco announced its intent to acquire the company, it had lauded ambitions to integrate it into its UCS platform. However, the fact is that Cisco has completely mismanaged that acquisition. This is a telling issue for Nutanix: if it does not want to disappear, then it shouldn’t let Cisco buy it.
Next, Cisco likes to use its own hardware. Currently, Nutanix uses a four-node SLED device based on a Supermicro FatTwin. Dell’s OEM version also uses a four-node SLED device. Cisco’s UCS server hardware is based on blade or rack technology, and Cisco will not like paying Supermicro or Dell dollars for hardware. Further, it will take it some time to engineer a SLED device, thereby allowing competitors space to move in. (This could have the same result as the Whiptail purchase: a ruined company.)
One of the more obvious blockers is that Nutanix would likely have a very high purchase price due to its current market valuation of over $2B. Chambers would most likely have to pay a premium to persuade a change of ownership. That said, with well over $50B in liquid assets sitting in the bank, this would not be too much of an issue.
But perhaps the most convincing reason to think this is not going to happen is that Nutanix CEO Dheeraj Pandey has slapped those rumors right on the head, confirming plans for an upcoming IPO and, further, stating, “If we ever decide to be part of another company, Dell will have a front-row seat in the bidding for it. That is my word to Michael Dell and a part of our agreement.”
Hopefully this will put these rumors to rest. But then again, perhaps it will start a new one: “Dell is buying Nutanix.” Well, when has the truth ever stopped a good story?
I would be very surprised if Cisco was seriously interested in buying Nutanix, but if it is I wouldn’t think for a second that the 4 node SLED form factor is in any way important to the decision. Nutanix also has 2 node appliances (6000 and 9000 series) and single node appliances (7000 and 8000 series). The 4 node appliance is no more than a pragmatic packaging solution for fitting a product that requires 3 nodes to deliver a working cluster, into the smallest amount of RackSpace.
And for what it’s worth, I would not be surprised if Nutanix is reviewing its IPO plans following Atlantis Computing’s HyperScale launch. I can’t help but think that Atlantis has created a few waves with the pricing of HyperScale.