There are a lot of stories coming out on whether the DoJ’s decision to try to block the HPE/Juniper deal ends up helping Cisco (HERE and HERE, for example), and Juniper’s CEO has also made that comment. The sense of the view is that if blocking the merger helps the current dominant player, how can the merger be anti-competitive? In order to know whether the claim is true, we have to look at how the success of the merger would hurt Cisco, and whether that hurt is a result of the merger itself or simply a reflection of market conditions. I’ve already come out against the DoJ on this, so I won’t go into that in detail here.
What “helps” Cisco, in the real world, is what raises its profits. How else can you measure benefit to a company? To do that, a company has to cut costs, raise revenues, or both. Cutting costs is a zero-sum game, so ultimately Cisco and every other vendor in the network (and any other) space is dependent on raising revenue to achieve a gain. That can be done by having buyers spend more (increase ARPU), or having more buyers (increase TAM or market share). It’s hard to invent new buyers in a mature industry, so the only way to get more buyers is to steal market share.
Cisco and Juniper have been locked in a market-share battle for almost three decades. That could mean that if the HPE/Juniper merger going through would create a more effective Cisco competitor, then having the merger blocked would indeed “help” Cisco. But my use of “could” rather than “would” is deliberate. Does the merger create a more effective competitor? Even if it does, is blocking it really a help? Let’s look at both points.
The problem in networking today, and the problem with tech in general, is that the TAM pie isn’t growing enough. Why does a company buy network gear? Two reasons. One, old gear is either broken or hard to maintain, so it needs to be refreshed. Two, new network benefits justify new network spending. This sets up the “race” that Cisco would run with Juniper today or HPE/Juniper if the deal goes through.
Suppose a company has a switch or router that’s getting long in the tooth. How do they replace it? It’s part of a broad deployment, in which over 90% of enterprises say there’s currently a dominant player. In fact, over 60% say that there’s only one player in a given part of the network, the part where our antiquated device lives. The vast majority of enterprises would replace it in kind, with the current version of the same device, from the same vendor.
Suppose a lot of gear was to be shifted out, though? Maybe it was deployed at the same time and so it’s all obsolescent. Maybe the vendor did something that turned off you or your management. Now, there’s a real chance there’d be a competitive bid, even without any new requirements.
Suppose a company does have a new source of network demand, a new mission that contributes a new business case? In many cases, this would require only some upgrades to current gear, perhaps some new line cards. In that case, the current vendor(s) supplying the devices to upgrade would win. In other cases, the new mission might require new gear, often deployed in a new enclave. That new mission would then justify taking a new look at vendors, a look where the incumbent vendor might be favored (if they’d made a good impression), or might not be.
Cisco, as the incumbent, has an advantage in total brown-field deployments, because those deployments favor the incumbent. If Cisco hasn’t messed up, they also have an advantage in “green-field” deployments for reason of their incumbency, but not as strong an advantage. They could be knocked off.
Cisco’s reorg clearly shows that Cisco itself thinks the current market situation poses threats to its profit growth, which it sort of does even statistically; the more share you have the harder it is to gain share, anti-trust concerns aside. But it’s also true that incumbents typically want to defend more than to attack, and so they are conservative in their product management. Cisco’s “fast follower” strategy is an example of this, and it goes back decades.
OK, suppose you are Cisco and you want to avoid competitive risks without a total remake of the market. What do you do? First, you try to defuse any differentiators that might arise from new tech developments. Think AI. Second, you try to exercise account control, strategic influence, to get companies to plan their strategies for new network missions in your own terms. That lays out where Cisco has to look at HPE/Juniper to decide if the merger is a threat.
Cisco’s reorg around AI is, IMHO, clearly aimed at keeping control of the one development that could create a new little enclave of network deployment. In-house AI means a cluster of GPU servers, perhaps requiring a data center network of their own. Not only that, rival Juniper’s AIops positioning has been effective against Cisco in areas where any sort of major network upgrade was proposed, and even won a few deals that displaced Cisco gear. HPE, of course, is an AI server vendor, and so has a definite involvement in any new self-hosted AI deployment.
In fact, HPE might well be a threat to Cisco’s account control even without Juniper. They already have the Aruba line, and their strategic influence among enterprise buyers is better than Cisco’s. For decades, my involvement with enterprises has shown that data center networking drives enterprise networking, and data center technology drives data center networking. Cisco has servers (UCS) but they don’t have a significant position in the enterprise data center market. HPE does.
So, from all of this, it would seem that blocking the merger helps Cisco. If that’s the case, given that Cisco is the incumbent, it could also hurt competition if the DoJ succeeds, but I’ve already given my views on that. However, it’s important to note that the only solution for Cisco in the long run is to add to the size of the network pie, and get the largest share of the addition. Yes, HPE/Juniper might be in a better place to do that, but HPE alone, and IBM alone, and maybe even Dell and Oracle and Broadcom alone, could do that too. Networks deliver stuff, and it’s the applications that create the stuff that builds business cases.
The big imponderable here is whether any of these Cisco competitors, including the now-disputed combo of HPE and Juniper, would be able to build any convincing new business cases. Or, of course, whether Cisco could. I personally doubt that Cisco would take a lead in developing new IT business cases, or that it could succeed if it did. But I also wonder whether the other players could. Only IBM has shown real appetite for business-case evangelism so far, but Oracle seems to be making some moves. Could Juniper drive HPE to do the same? Open question, I’m afraid, but I think the net here is that despite having just reported a decent quarter, Cisco does face risks down the road, and I think blocking the merger would reduce those risks.
