I think that the HPE acquisition of Juniper is the most significant company-side event in the network equipment market, and I think it’s critical for both HPE and Juniper. But execution is critical, and last week, HPE offered the first projections of a year’s worth of the combination. The Street didn’t like it. I asked Google Gemini and NotebookLM to analyze the report; view it here if you’re interested, and contrast it with my views here.
The reason why the HPE/Juniper deal is so critical boils down to the combination of the role of AI in the future of networking, and the fact that Cisco dominates the network space today. Juniper could not hope to counter Cisco effectively alone, and HPE was at risk to Cisco’s dominance if they relied only on their Aruba gear. That could have hurt their server business in an AI-centric future. HPE has good strategic influence; better than Cisco, so it makes sense to use that to push AI-centric networking, particularly if that makes the HPE server position in AI stronger and more secure.
The core of the HPE story to the Street was a focus on what the CEO called “the IT industry’s most promising opportunities in networking, cloud, and AI”, and shifting to “higher-growth, higher-margin businesses”. That implies that HPE is planning to focus on the potentially best market opportunities rather than trying to chase every server and network deal.
This sort of shift isn’t likely to be easy or fast, though. You can see that by the fact that HPE announced both a 10% dividend increase and the allocation of $3 billion to share-buyback programs. This has the effect of making the stock more attractive quickly, something you wouldn’t need if those higher-growth and margin opportunities could be realized right away. HPE’s presentation offered a number of steps they were, or planned, taking.
“Creating a new networking industry leader” is a goal, perhaps, more than a step. However, HPE is right in that their acquisition “created an opportunity to disrupt the status quo”. What I think is missing is an identification off just how that was created, and could be exploited. IMHO, the big deal lies in HPE’s strategic influence, which is almost double Juniper’s, and in the fact that you don’t need to network AI until you have AI to network. We’ll come back to this point.
“Capturing profitable growth in the AI infrastructure market”, also more a goal, so my comments on the first point above are even more relevant here. The good news is that their next subhead, “Accelerating high-margin GreenLake and software services growth” is at least more prescriptive. GreenLake is about AI hosting, which is what creates AI network needs.
Another positive step is “Capitalizing on unstructured data market gr4owth with our own IP [Intellectual Property, not Internet Protocol] through Alletra MP. Enterprises’ own data is a critical component of AI as far as enterprises are concerned, and so any real AI hosting initiative has to address data access. A server/software vendor is in a better position to do this than a network vendor.
“Driving customer transition to next-generation server platforms” is a final realization of the reality of HPE’s fulfillment of the Juniper potential. That’s true for several reasons. First, obviously, you want AI to drive your server business if you’re a server provider. Second, server sales is your opportunity to exercise strategic influence early in any build-out, while network needs necessarily follow hosting decisions. So this is good, but I think the two reasons should have been raised early and driven hard at every subhead thereafter.
“Optimizing cost structure to sustain long-term profitability” is, I think, almost an admission of weakness. Taken on its surface, this implies that without cost management the realization of the AI networking opportunity (and the hosting opportunity) wouldn’t necessarily be enough to build a great HPE business. I also don’t think this is what HPE intended to say. The value of the cost side is that it could make the sum of the parts of HPE and Juniper less costly than the two were separately, which means that there is an almost-immediate benefit to the bottom line. That’s good if the sales cycle for AI self-hosting is likely to be long, which it is. This is also why they push their raising of the dividend and share buyback pool; investors don’t have to wait.
So, in all, this doesn’t sound like a bad story, and I don’t think it is one. I do think it’s badly told, and I also think that one of the early decisions HPE made, which was to retain Juniper’s semi-independence as a business, has a risk. Juniper is Juniper, a follower in terms of creating AI hosting, which makes it a weak AI network driver. HPE is a server/software vendor with more strategic influence by far. Does the retention of a Juniper autonomy weaken HPE’s ability to exploit their own strategic AI assets? I think it could.
I’m not saying that Juniper doesn’t have a good AI story. Mist AI is the most-recognized AI network tool on the market. But people don’t want or need an AI network without the AI in place, and Juniper alone cannot make that happen. HPE/Juniper can build demand for AI-native networks, then fulfill it, but to do that HPE has to be a decisive strategic leader in the story.
HPE’s stock fell sharply after this conference presentation, and didn’t significantly recover the following day. At the very least, this shows that the presentation to the Street missed the mark, since no company makes a Wall Street presentation to kick their stock in the butt. The worst? That the Street recognized what I noted here, which would mean that HPE isn’t doing everything it needs to do to realize value from the Juniper deal. That could even mean that the deal will end up a net negative, because for sure it will be distracting.
I think there are issue here. While I think the HPE/Juniper deal is in fact the most significant company event in the industry in a decade or more, I also think that HPE management hasn’t grabbed the brass ring, or perhaps even reached for it, and that could be a major risk for the combined company and a major opportunity for Cisco.
