Neither Line Is Being Killed. You Can Stop Waiting to Find Out.
Since HPE closed the Juniper acquisition, campus and data centre refresh decisions have been sitting frozen on one question: if I standardise on Aruba, does HPE kill it for Mist — or the reverse? A year of announcements has answered it, and the answer is neither. HPE's stated approach is "build once, deploy twice": capabilities developed once are deployed to both Aruba Central and Juniper Mist, hardware convergence begins with access points that run on either platform, and existing customers are told to keep using what they have and converge at refresh. Waiting for a winner means waiting for something that is not coming.
That does not mean the decision is free of risk — it means the risk moved. It is no longer discontinuation. It is roadmap timing, feature parity between two management planes, and what "converge sometime in the future" means when written into a five-year estate plan. This page is about that, not about the merger news. For choosing within each line, see the Aruba selection guide and the Juniper selection guide.
What HPE has actually committed to
- Both management platforms continue. Aruba Central and Juniper Mist are being merged through shared features rather than consolidated into one. HPE's own framing is "build once, deploy twice" — a capability is written once and shipped to both.
- Capability flows in both directions. Aruba's Agentic Mesh technology is coming to Mist, improving issue detection and root-cause work there; Aruba Central receives the Marvis framework, extending AI-driven actions across the portfolio. This is not one platform absorbing the other.
- Hardware convergence starts at the access point. The first genuinely combined hardware is Wi-Fi 7 access points that can run on either Aruba or Juniper.
- Existing customers are explicitly told to carry on. HPE's position is that Aruba and Juniper customers continue with the products they prefer and, as they refresh, arrive at the same point in future.
- The switching lines both continue. Juniper QFX has been extended for AI data centre work with new models, and Mist support has been extended toward HPE Networking CX switches. Neither switching family is being retired.
Read that last commitment carefully, because it is the one that matters commercially. "Arrive at the same point sometime in the future" is a direction, not a date. It is a reasonable thing for a vendor to say a year into an integration. It is not something you can put in a business case without asking for specifics.
What the DOJ settlement tells a buyer
The competition remedy is usually reported as merger trivia. Two parts of it are worth a buyer's attention:
- HPE had to divest the Instant On campus and branch WLAN business — assets, intellectual property, R&D staff and customer relationships — to a department-approved buyer within 180 days. Instant On is small-business gear rather than enterprise, so it rarely touches an enterprise bill of materials directly. Its significance is as a signal: the wireless overlap was considered real enough to require a divestiture. Confirm the current ownership before quoting anything from that line.
- Juniper's AI Ops for Mist source code had to be licensed by auction, on perpetual and non-exclusive terms, with optional transitional support and personnel transfers. That means a competitor can ship products built on the same AIOps engine that underpins Mist. If your reason for choosing Mist is the AI operations layer specifically, note that the layer is no longer exclusive to HPE — which is good for your negotiating position and worth knowing before you treat it as a lock-in.
So which line do you standardise on
With discontinuation off the table, the decision returns to ordinary technical and commercial fit. The honest version:
| Your situation | Reasonable answer |
|---|---|
| Existing Aruba estate, refreshing part of it | Stay Aruba. Operational knowledge, ClearPass policy, existing Central tenancy and spares all carry forward. A mid-refresh platform switch buys you retraining and a migration you were not asked to do |
| Existing Juniper estate, refreshing part of it | Stay Juniper. Same reasoning in reverse — Junos operational practice and Mist tenancy are real assets |
| Greenfield campus, wireless-led | Either. Decide on the wireless and policy requirements and let switching follow, since the AP hardware is the part converging first |
| Greenfield data centre fabric | Juniper. QFX with Apstra is the more developed data centre story, and it is where HPE has been investing for AI fabrics |
| Mixed estate today, both present | Standardise per site or per domain, not across the whole estate. Two management planes are a fact for now; splitting them cleanly by boundary is cheaper than a migration |
| Strict policy or NAC requirement | Check ClearPass against your requirement first — it is frequently the deciding component and it sits on the Aruba side |
Notice what is absent from that table: "whichever HPE keeps". That column no longer exists, and building a decision around it produces the worst outcome available — deferral. Campus and fabric hardware ages on a schedule that does not pause for a vendor integration, and the estates that froze in 2025 are now a year further into their support cliff without having bought anything.
The two management planes, honestly
Running both is the part that has a real cost, and it is worth naming rather than glossing:
- Two consoles, two operational models, two sets of skills for as long as convergence takes. For a small network team that is a genuine burden.
- Feature parity is a moving target. "Build once, deploy twice" means a capability lands in both, but not necessarily on the same day. Ask when a specific feature you depend on reaches the platform you are on.
- Licensing and subscription terms differ between the platforms. Compare them on the same term length and the same device count before concluding either is cheaper — the same discipline that applies to the compute-side management licensing, where a subscription and a perpetual entitlement can deliver the same capability at very different total cost.
The counterweight is equally real: one vendor relationship, one support organisation, one commercial negotiation covering both lines. If you were running Aruba and Juniper as two separate vendor relationships before, the consolidation is on your side of the ledger.
What to get in writing before you commit
The risk is timing, so the questions are about timing. Before a multi-site standardisation decision:
- Published end-of-sale and end-of-support dates for the specific models you intend to buy — not the family, the model. A product line continuing does not mean every SKU in it continues.
- Which management platform your chosen hardware is supported on today, and which it is committed to reach, with a date rather than a direction.
- Feature parity for the specific capabilities you depend on. Name them; do not accept a general assurance of convergence.
- Migration commitments if HPE later asks you to move platforms — who pays, and what tooling exists.
- Subscription terms and what lapses at the end of a term, on whichever platform you land on.
These are ordinary lifecycle questions. The merger did not create them; it made people forget to ask them because they were busy asking a bigger and less answerable one.
Where this changes a refresh plan
Two practical consequences for anyone with a network refresh in the budget:
Unfreeze the decision. If a campus or fabric refresh has been on hold pending clarity, the clarity has arrived and it is "both lines continue". The cost of another year of waiting is a year of support cliff, deferred capacity and, frequently, a worse price than was available during the uncertainty.
Buy the hardware on fit and treat the management plane as a separate decision. Access points are converging first, so AP hardware is the least risky place to commit. Switching is where operational practice is stickiest and where staying on your existing line has the most value. If the two point in different directions for you, that is a legitimate reason to split by domain rather than force one answer.
Where a refresh spans several countries, the availability and support questions matter more than the platform question — commitments differ by market, and that is worth settling per site, as covered in HPE support tiers and site eligibility. And whichever line you land on, the channel checks are unchanged: regional SKU, serials before payment, deal registration — see sourcing HPE.
Get the refresh priced on both lines
Send this and we price it — no questions back:
- Site count, and access point plus switch port count per site
- What you run today — Aruba, Juniper, or both — and which management platform your team already operates
- Wireless standard required (Wi-Fi 6E or Wi-Fi 7) and any policy or NAC requirement such as ClearPass
- Uplink speeds and whether the data centre fabric is in scope as well as campus
- Destination country per site, and target date
You get firm pricing, availability and delivered lead time within one business day.
Still deciding between the two? Send site count, port and access point counts, what you run today, destination country and target date — we come back with both lines specified and priced, not a list of questions.
Get Aruba and Juniper priced side by side Prefer email? sales@haink.org
We quote both lines against the same site plan, with the management subscription shown separately, so the platform decision is visible rather than buried in a bundle.
Frequently asked questions
Is HPE discontinuing Aruba?
No. HPE's stated approach is to develop capabilities once and deploy them to both Aruba Central and Juniper Mist, with Aruba Central receiving the Marvis framework and Aruba's own Agentic Mesh technology going the other way into Mist. Existing Aruba customers are told to continue with the products they have and converge at refresh.
Is HPE discontinuing Juniper Mist?
No, and Juniper's position within HPE is arguably stronger than a simple acquisition would suggest — the QFX switching line has been extended for AI data centre work and Mist support has been extended toward HPE Networking CX switches. Both management platforms continue.
Should I wait for the platforms to merge before buying?
No. There is no announced date on which one platform replaces the other, because that is not the plan. Waiting costs you a year of support runway and capacity while producing no additional information. Buy on technical and commercial fit, and ask for dated commitments on the specific features you depend on.
We run both Aruba and Juniper. Should we consolidate onto one?
Usually not as a whole-estate migration. Standardising per site or per domain — campus on one, data centre on the other, for example — captures most of the operational benefit without paying for a migration you were not asked to make. Revisit at the next natural refresh boundary rather than forcing it now.
What happened to Instant On?
HPE was required to divest the global Instant On campus and branch WLAN business — assets, intellectual property, R&D personnel and customer relationships — to a department-approved buyer within 180 days of the settlement. It is small-business rather than enterprise equipment. Confirm current ownership and support arrangements before quoting anything from that line.
Does the DOJ remedy affect what I buy?
Indirectly, and in your favour. Juniper's AI Ops for Mist source code had to be licensed by auction on perpetual, non-exclusive terms, so the AIOps engine behind Mist is no longer exclusive to HPE. If that layer is your main reason for choosing Mist, it is worth knowing it may appear in a competitor's product — which weakens the lock-in argument and strengthens your negotiating position.
Does this change how we should buy the hardware itself?
Not at all. Regional SKU suffix matched to the country of installation, serial numbers verified before payment, deal registration under a named partner ID, and support level confirmed per installation address. A vendor integration changes roadmaps, not the ways a quote can be wrong.
Related
- How to choose HPE Aruba Networking · How to choose Juniper Networks — selection within each line
- Cisco vs Juniper switching — the comparison that matters if you are considering leaving both
- Choosing HPE ProLiant servers and storage · sourcing HPE — the channel checks are the same across the combined portfolio
- HPE support tiers and site eligibility · management licensing — subscription versus perpetual, the same question in a different domain
- Vendor partner programs · deal registration — how a merged vendor's programmes affect a bid
- Buying at project scale · stock vs BTO/CTO
- Hong Kong · Dubai — regional availability and support coverage
Sources
- US Department of Justice — Divestitures and licensing commitments in HPE's acquisition of Juniper Networks, 28 June 2025 (divestiture of the global Instant On campus and branch WLAN business "including all assets, intellectual property, R&D personnel, and customer relationships, to a DOJ-approved buyer within 180 days"; auction to license Juniper's AI Ops for Mist source code, "perpetual, non-exclusive, and include optional transitional support and personnel transfers")
- SiliconANGLE — Aruba meets Juniper Mist at HPE Discover ("build once, deploy twice"; Aruba's Agentic Mesh coming to Mist; first combined hardware as Wi-Fi 7 access points running on either platform; customers continue with preferred products and converge at refresh)
- NAND Research — HPE Discover 2026 recap (QFX extended with new models for AI data centre work; Mist support extended toward HPE Networking CX switches; Marvis framework embedded in Aruba Central; Mist Data Center Assurance integrated with Compute Ops Management and GreenLake)