The deal HPE has made to acquire Juniper Networks isn’t yet final, but the competition between the yet-to-be-combined entity and its natural rival, Cisco, is already starting. I’ve had 211 enterprises tell me that they’ve heard from both companies already, positioning for supremacy in a battle that will surely unfold. There’s also a battle over the loyalty of channel partners, as this piece shows. The article also includes some comments from the CEOs of the rivals, and they open some interesting opportunities for analysis.
The article opens with HPE CEO Neri comments on HPE becoming “a networking company at the core. Something that probably Cisco has forgotten now for a little bit, and I think that’s the big opportunity, which ultimately is to give customers a more modern, AI-driven experience at the networking level….” It’s hard to tell for sure from this whether Neri thinks networking is the core of AI, the core of any transformation, or a place where AI needs to be applied as a priority. Since the last of these has been Juniper’s theme, I suspect it’s the meaning, and the article says Neri went on to call networking the core foundation, “before you even think about cloud or AI.”
It sure looks like HPE is staking the future of the Juniper deal on networking, which is strange given that HPE’s differentiation lies in the compute/software side. Cisco’s CEO Robbins is content to counterpunch in a fight he sees is already held on Cisco’s terms, saying “we have more technology that brings more value to our customers in the infrastructure layer than anybody else.” The piece cites a channel player in support of this; “Cisco is still the dominant player in the networking space, Long View Systems’ Irvine said. Security, he added, is one of HPE’s biggest gaps.”
Only 178 of the 211 companies I’ve cited commented explicitly on the HPE network-first strategy. The most common remark (offered by 151) was that given HPE’s acquisition of Juniper, it’s inevitable HPE would focus remarks on networks. But 75 enterprises also pointed out that the data center is where HPE’s IT gear is installed, and the data center network is the place where the majority of enterprise network capex is spent. The Juniper deal gives HPE a shot at displacing Cisco there, and Cisco thinks that HPE risks its own IT sales if they fail. That’s perhaps the biggest insight point of the whole HPE/Cisco duel.
Enterprises overwhelmingly accept that the data center network policy sets network policy overall, so a win or loss in the data center network has major implications. Juniper, as probably everyone knows, has made security and AI the cornerstones of its positioning, and the channel comment I cited above suggests that resellers believe HPE is weak there. That, again, suggests that HPE’s Neri is smart pushing networking, and the Juniper deal, because network security from Juniper could help fill the gap.
AI is another factor in the game here, as the article notes. I think that the concept of AI-specific network needs driving enterprises is consistent with the widely held views that AI will be a driver in enterprise business transformation. As I’ve said in prior blogs, that demands AI be applied to business analytics, which in turn means that it needs to be used on highly confidential core business data, the kind enterprises don’t want to share with others. That’s why the great majority of enterprises think self-hosted AI is critical, and of course self-hosted AI is hosted in the data center, and its use impacts the data center network more than any other part of the network.
Juniper has made an explicit distinction between AI as a tool in network operations, and AI as a driver of network traffic. Cisco, less so; their AI focus has been more on the former, perhaps because enterprises say that Juniper’s AIOps tools are their best differentiator. Perhaps there’s another, broader and more interesting reason.
Often it’s possible to look at competition like warfare, in which there’s both strategy and tactics. The latter are the most visible from the outside, and that would naturally color the channel view of the HPE/Cisco face-off. Strategy, though, is the core truth, and what tactics is only valid to support. What’s the strategy here?
First, Cisco has the strategic advantage of network incumbency in its favor. Right now, with almost 90% of all network spending being little more than routine replacement of aged gear, the incumbent has the advantage because the trickle of gear subject to modernization fits inside an established infrastructure most likely dominated by Cisco. The status quo works in Cisco’s favor.
HPE has the strategic advantage of IT incumbency. Business transformation has to be grounded in applications and data before delivery matters much, and so for decades it’s been the IT vendors who have had the most strategic influence in the data center, including the data center network. But this influence is only useful if it can project transformational application value that drives network change. Something that can not only justify but perhaps demand a change in a big chunk of data center network infrastructure is a threat to the status quo, and so to Cisco.
For HPE to make a go of the Juniper deal, they need a strategy of business transformation, one that promotes new benefits to drive new IT deployments, which in turn would drive new data center network requirements. However, raising them now could be a negative because the Juniper deal hasn’t closed, and thus it could raise interests HPE can’t fully capitalize on. So, might HPE be focusing on a pure network theme, to draw Cisco’s positioning into its natural, incumbent-defensive position? Perhaps.
Or perhaps they’re messing this up. HPE, in the IT space, has been as much a seeker of the status quo as Cisco and Juniper have been in networking, for the same reason. If projects make up less than ten percent of IT and network spending, then why try to push against the tide?
There are two possible directions that network infrastructure could take. The first, the status quo, is the “no-differentiation-and-commoditization” direction, and the second the “new-revenue-driver-found” direction. If the first of these directions is taken, then the incumbent wins because the new-gear-needed decision results from the aging out of a piece of that incumbent’s gear, which favors the incumbent. The only way for a non-incumbent to win in networking is for something to come along to change that new-gear decision, to create an opportunity to differentiate that’s meaningful to buyers. Cisco doesn’t need to do anything; it’s HPE that does. Do they, or does Juniper, understand that? That’s the important question, and of the 211 enterprises I’ve been citing here, only 7 said that HPE said anything at all that could signal this sort of realization. They need to do a lot more.
