One of the most compelling symptoms of a market transformation, sometimes even a driver of one, is M&A. HPE’s announced-but-not-yet-consummated acquisition of Juniper is surely a major event in networking, and surely a major step for HPE. How can we read the symptoms of the deal to get an advanced indicator of its market impact? I blogged on the deal back in January (HERE and HERE) but there’s a recent development reported in CRN.
My key points on the proposed deal were first that it leveraged a basic truth, which was that data center strategy was driving network strategy, and that future transformational shifts in both networking and IT were likely to require movement in both technology sets, most easily accomplished by a vendor with a solid position in both areas. Now, we’re hearing that HPE is looking for channel partners with skills “especially in the security area and also around data center networking.”
Juniper has done well with its network-security business and it’s certainly active in the data center networking space. It’s also a major player in the AI-in-the-network area. HPE’s statement includes a commitment to the first two but not the third, but the HPE channel chief quoted above also said “The AI opportunity is out there, it’s big, it’s a large opportunity, but there is still time.” What can we say about the relegation of AI to the future, and to early HPE interest in the channel to market Juniper gear?
One obvious question, given the fact that Juniper already has a channel program, is why HPE is looking for Juniper-symbiotic channel partners. In the data center networking space in particular, this would seem to invite a channel conflict unless HPE wants to replace the Juniper partners. However, the HPE channel head says there’s a lot of enthusiasm for the combination product family among both partner sets. That, of course, could mean that both partner sets are at risk, and I’d suspect that HPE plans to revamp their program overall, likely emphasizing a different and broader skill set and pruning out partners who can’t offer that.
Channel, of course, is usually a product conduit that’s highly dependent on leads from the vendor, which means marketing programs. The alternative is a feet-on-the-street program that relies on the partner to build the engagements, but that almost always generates channel conflict. To me, the main thing to draw from this is that HPE wants partners who can push the symbiosis with Juniper, ensuring some positive near-term sales benefits to the bottom line, a move that can cover longer-term strategic plans.
“Can” is the operative word. I don’t know of many companies who have pushed a strategic positioning through a channel without a major marketing and PR initiative behind it. There are a number of potential initiatives HPE might undertake with Juniper, but I don’t know whether they’re committed to any, which is one concern I have.
The HPE wait-a-bit strategy for AI is another, but I’m not sure I believe the statement can be taken at face value. Remember the deal might not close until 2025, the UK just said its regulator is probing the deal, and AI is a moving target for vendors. I’d bet HPE has people working on integrating Juniper’s AI into the full HPE line, but till they get some idea of what they want there, and where AI is heading, it might be wise to be careful, and even a bit slow. But I think Juniper has a good lead in the application of AI to netops, and I think it’s a mistake not to be thinking about how to leverage it. I also think Juniper’s Apstra offering could be a game-changer in self-hosting AI, and that’s a space that could develop very quickly, and get competitive just as quickly.
HPE’s just-announced deal with NVIDIA makes AI an in-the-present-and-in-your-face issue. It’s inconceivable that HPE could pursue this without a thought for the symbiosis with or impact on Juniper’s AI, and the NVIDIA AI Enterprise software stack should logically become the toolkit for any generative AI work on the Juniper side, particularly given that HPE Greenlake management and observability may be a target, and HPE’s OpsRamp operations copilot tool is similar to what Juniper does in the networking space. A combined visibility and generative documentation copilot is logical, but it will require careful planning to align it with user views (a topic I’ll address tomorrow).
HPE notes the SD-WAN/SASE opportunity, which in my view is both linked to security (or should be) and has been under-exploited by Juniper (they got their technology from 128 Technology via M&A). I think that’s an area where security melds with data center networking policy, and both these Juniper technologies got an AI feature boost just a couple weeks ago. It would seem to me that prepping the channel for this stuff would be important to HPE, since it’s an untapped opportunity space that HPE could easily leverage.
Things are even more complicated in the service provider space and with regard to the cloud. I suspect that the hyperscalers and operators are almost always major account plays, with a dedicated sales team. Will HPE merge them, eliminating some redundancy? How will the team get a really integrated view of the products? Right now, this space is an opportunity desert from a capex perspective, and if HPE wants to fix that they can’t continue on as though the M&A had never occurred, and just integrating the sales team won’t fix that. We can’t expect insight in this area from a channel program head.
From a strategy perspective I think these two companies need to get this deal done as quickly as possible. There’s too much going on in spaces where the M&A can be expected to have a major impact to sit around, but for either company to do much in anticipation of the deal is a risk. Not because of a risk it won’t go through as much as because there’s always a risk of a leak from someone who feels the proposed combined strategy will put them at a personal disadvantage.
