When you ask questions you sometimes get solid answers. Sometimes, you get answers that are less than solid. Are network equipment vendors jumping more into the enterprise market? Yes, unquestionably they are. Is that the right move? Yes, but…. That makes the question we need to look at simple; how solid is that “yes, but”, which depends on how strong the possible exception is. Let’s explore it and see.
Say you’re on the Titanic, and they’re playing “Nearer My God to Thee” or “Autumn”, depending on what version of the story you like. You jump into the cold sea and you see a wooden deck chair nearby. Do you grab for it? A likely response would be unequivocal; “Yes!” The cold sea is fatal fairly quickly. A more reasoned response is “Yes, if nothing better is perhaps just a bit further away.” And a thoughtful one is “In any case, you need to find something that actually gets you entirely out of the water, like a lifeboat.” In the telecom capex slump, is the enterprise market a step toward salvation, or just a temporary reprieve from the cold?
Let’s get one point out of the way here, which is that there’s nothing wrong with getting drier, even if you’re still a bit wet. It’s smart for vendors to look to enterprises to pick up the sales slack created by the pressure on telecom capex, even if what enterprises can do is going to be used up fairly quickly. It’s smarter to be asking how long the enterprise respite can last, because steps will have to be taken to provide a drier and warmer place before it fails too.
Capex depends on ROI. Every organization who buys network gear has to present the CFO organization with a business case to justify the spending. What the CFO wants is proof that the ROI on that business case will at least match if not exceed the ROI target, which is usually the buyer’s internal rate of return on invested capital, the average of the stuff that’s been done up to that point. What’s been happening in the telecom space is that the profits generated by services have failed to keep up with Wall Street’s required level of growth, required to justify a nice steady appreciation in stock price. That’s because revenues have failed to grow fast enough, which means that profits can’t be sustained without cutting costs, cutting capex, and hurting vendor sales.
This same ROI dependency exists for enterprises, but it’s a little more complicated there because most enterprises don’t buy network equipment to support sales of services. They’re buying it to connect IT elements with workers, customers, and partners, in order to build an information framework for their operations. They usually think of this as a means of supporting “productivity”. That means that a new investment in networking has to be linked to a productivity improvement whose value is greater than the investment by the ROI target level or more. For modernization/renewal, it usually means that the cost of the modernization has to be lower than the original cost, or that any increase is justified by an incremental benefit.
For the great majority of the 70 years or so that companies have used IT, we’ve typically seen a new productivity paradigm that justified additional IT investment come along roughly every 15 years. That has resulted in a new wave of investment in IT, and so in a sense the enterprises has refreshed its value roughly at fifteen-year intervals. However, we have had no such refresh cycle in this century, which means that the paradigms that have justified additional enterprise investment in any form of IT, including networking, are getting stale.
This is the source of what (to go back to my opening) I’ll call the “lifeboat problem”. Enterprises are a bit behind operators in reaching an ROI impasse, but not all that far behind. It’s very difficult to model just where things are in that regard, even with a lot of comments from enterprises to use as primary data. My current attempt yields a rough four-year timeline, after which enterprises will cease to offer network vendors any relief from capex pressures at the telecom level. After that, while telecom capex pressure will likely continue unless something happens (like subsidization), enterprise capex pressure will probably continue to grow and even exceed that on the telco side. Unless there’s a lifeboat.
Networking along cannot ignite a new productivity paradigm. Information delivery has value to the extent that what’s being delivered has value. If we want a new productivity paradigm we need a new relationship between information and the stakeholders in a company’s business. That new relationship is a mixture of the information and the delivery requirements of the relationship, which of course depend on who the delivery targets are. However, a change in the information relationship can also be used to manage costs, and right now we have an example of both of these uses, examples in the real world of M&A.
Broadcom has bought VMware. Broadcom is a chip company, and chips in general (and AI chips in particular) are of high value in the market. VMware is a middleware company, a provider of virtualization tools that aim both at virtualizing compute resources and network resources. If all Broadcom wanted was better revenues, why not buy AI companies? Could it be that Broadcom is seeing a network future were capex pressure continues, and where network devices like routers and switches are totally commoditized? Could they be seeing a future where the top layer of networking and the bottom layer of IT combine into a new virtual-services-middleware game? Could then be seeing the death of networking as an independent set of vendors and products? They could be.
HPE is buying Juniper. HPE is primarily a server/software company, a compute piece of the great IT puzzle, but they already have (with Aruba) a network technology element too. Computing as a part of the capex story is under pressure, and so is networking. Could HPE be buying Juniper because they believe that the combined company, eliminating a lot of jobs made redundant by the combining process, could be more competitive in a commoditizing market? They could be, but both Broadcom and HPE might also be seeing the lifeboat.
Could a new productivity paradigm require a level of fusion between compute, information, and networking that we simply cannot create today when all the pieces are supplied by separate vendors with separate financial priorities? Could Broadcom be looking to create a fusion-middleware story, the foundation for a new foundation layer for IT? They could be. And could HPE be seeing that the next wave of productivity, necessarily rooted in the compute side, will then empower growth again in the network domain, and be buying into an asset it intends to revalue with a new productivity initiative, driving a new IT cycle of growth. They could be.
And here, as you probably guessed, is where we need to return to to the Titanic. We have a couple of swimmers, we have a nearby deck chair, and a more distant lifeboat. So far, neither of our swimmers has made a convincing choice of destination. It would be smart, probably, to strike out for the deck chair and ride cost-driven opportunism for a bit; it’s a fairly certain source of profits and ROI, but only for a while. The danger is that floating on that deck chair will still sap strength and waste time, and just maybe so much so that the lifeboat floats out of reach. I hope both Broadcom and HPE can struggle through the figures of speech here, and reach the right conclusion.
