HPE earnings preview: Juniper makes networking the AI story

Hewlett Packard Enterprise reports fiscal Q2 results on June 1, and the lazy version of the preview is easy: AI servers, backlog, margins, guidance. That misses the better story. HPE is no longer just asking investors to believe it can sell more AI systems. It is asking them to believe Juniper has turned networking into the more valuable AI layer.

That changes the way the quarter should be read. The most important number may not be the headline server line. It may be whether HPE Networking can keep order momentum, defend margins, and prove that data-center switching, routing, campus automation and AI-native operations are becoming the connective tissue of enterprise AI spending.

The catalyst is official. HPE Investor Relations lists the Q2 Fiscal Year 2026 earnings call for June 1, 2026, with Antonio Neri, Marie Myers and Paul Glaser scheduled for the webcast. TECHi’s AI stocks earnings watch also lists HPE for June 1 after the close, with an EPS estimate of $0.44.

Article Brief

Key Takeaways

4 points24s read

1. 01The real story–HPE reports June 1, but the clean AI read is now Juniper-led networking rather than just AI servers.
2. 02Q1 proof–Networking produced $2.7 billion of revenue, grew 151.5% year over year and carried a 23.7% operating margin.
3. 03Server caveat–Cloud & AI revenue fell 2.7% in Q1 and server revenue was down 2.7%, so the AI-server narrative still has to catch up.
4. 04June 1 test–Watch Networks for AI orders, Juniper revenue synergies, Networking margin, AI backlog conversion and leverage reduction.

The stock already knows something changed

HPE is not entering the report as a sleepy enterprise hardware name. TECHi’s AI stocks universe shows HPE in the AI infrastructure bucket at $44.49 after-hours, up 16.44%, with a Hold signal, a 49 score and a negative 1.6% one-year model read. That is the useful tension. The market has rewarded the Juniper story quickly, but TECHi’s scoring model is still asking whether the fundamentals justify chasing the move.

The quote subpages tell the same story from different angles. The HPE forecast page shows a Buy consensus across 23 analysts, but the average target in that snapshot sits below the live quote. The technicals page shows an RSI of 79.61, an overbought reading, and spot sitting 39.14% above the 50-day moving average in the May 26 snapshot. In plain English: the narrative has improved, but the tape is stretched.

Q1 explained why Juniper matters

HPE’s Q1 FY26 earnings release was the first clean proof point. Revenue rose 18% year over year to $9.3 billion. Non-GAAP gross margin reached 36.6%. Free cash flow was $708 million. HPE also raised fiscal 2026 EPS and free-cash-flow guidance.

But the segment split is the real story. Networking revenue was $2.7 billion, up 151.5% year over year, with a 23.7% operating margin. HPE said the segment incorporates the former Intelligent Edge business and Juniper Networks. Cloud & AI revenue, by contrast, was $6.3 billion, down 2.7% year over year. Server revenue was $4.2 billion, also down 2.7%.

That is why the June 1 setup is not simply “did AI servers grow?” If servers alone were the story, HPE would still look like a cyclical infrastructure vendor trying to catch Dell, Supermicro and the GPU-rack crowd. With Juniper, HPE has a different pitch: the AI data center needs switching, routing, WAN connectivity, observability and AIOps as much as it needs boxes full of accelerators.

The network is becoming the AI bottleneck

The strongest evidence is buried in HPE’s Q1 presentation. Networking represented nearly 30% of total revenue and more than half of total non-GAAP operating profit. HPE said record order growth was driven by wireless, data-center switching and routing demand, and raised its fiscal 2026 target for cumulative Networks for AI orders to $1.7 billion to $1.9 billion.

The same presentation said AI backlog was above $5 billion, with 64% of cumulative order mix booked in enterprise and sovereign customer segments. It also called out strong Networks for AI orders driven by Juniper QFX data-center switches and Juniper PTX/MX routing products. That is the core of the angle: HPE’s AI opportunity is no longer confined to the server rack. It runs through the fabric that makes the rack usable.

What Juniper changed

HPE closed the Juniper acquisition on July 2, 2025. The company said the deal doubled the size of its networking business, moved the portfolio toward higher-margin growth areas, and gave it a full networking stack from silicon and hardware to operating system, security, software and services. Rami Rahim, Juniper’s former CEO, now leads the combined HPE Networking business.

There was also a regulatory cost. The Justice Department settlement required HPE to divest its global Instant On campus and branch WLAN business and to license parts of Juniper’s Mist AIOps source code through an auction process. That matters because Mist is one of the clearest AI-native assets in the portfolio. The settlement did not kill the thesis, but it means investors should listen for how HPE protects differentiation while honoring the remedy.

Since the close, HPE has pushed the networking-for-AI message hard. In December, it introduced a broader AI-native networking portfolio, including the HPE Juniper Networking QFX5250 switch built on Broadcom Tomahawk 6 silicon with 102.4 Tbps bandwidth, plus the MX301 multiservice edge router. In February, HPE added PTX routing products for AI network fabrics and service-provider modernization, including 800G-capable platforms aimed at low-latency, high-capacity traffic growth.

The June 1 questions

The first question is whether Q2 clears HPE’s own guide. Management guided for revenue of $9.6 billion to $10.0 billion, GAAP diluted EPS of $0.09 to $0.13, and non-GAAP diluted EPS of $0.51 to $0.55. A simple beat matters, but it will not be enough if the mix is weak.

The second question is whether Networking can keep its margin profile. HPE guided Networking to the low-20% operating margin range for fiscal 2026, and Q1 came in at 23.7%. If that margin holds while revenue grows, investors can treat Juniper as more than acquired revenue. If margin slips quickly, the market will ask whether the Q1 strength was front-loaded integration math.

The third question is whether AI backlog converts without sacrificing profitability. HPE said the majority of AI backlog would be recognized in the second half and beyond, with Q3 expected to be the largest AI revenue quarter. That makes June 1 a bridge quarter, but not a throwaway. The call needs to explain whether the second-half ramp is locked, slipping or getting better.

The bull case is cleaner than it used to be

The bull case is that HPE has found a better AI lane than chasing every hyperscale server deal. Networking is higher value, stickier and more strategically important as AI workloads move from training clusters into distributed inference, edge routing, secure campus networks, sovereign AI and service-provider fabrics. Juniper gives HPE credible hardware, software and AIOps assets exactly where the bottleneck is moving.

That is not just a product claim. It also shows up in financial architecture. HPE is targeting at least $2.0 billion of fiscal 2026 free cash flow, says it is committed to an investment-grade balance sheet, and showed $4.8 billion of cash and equivalents at the end of Q1. The company also put net leverage at 2.6x and said it remains on track for roughly 2.0x by fiscal 2027.

The bear case is not dead

The bear case is that investors may be over-upgrading HPE because the segment labels changed faster than the core business. Cloud & AI was still down in Q1. Server revenue was still down. AI server revenue is expected to rise sequentially, but a lot of the payoff sits in the second half. That leaves room for commodity pressure, supply timing, backlog conversion risk and integration friction.

There is also valuation risk after the move. A stock can have a better story and still be too stretched into an event. TECHi’s Hold score, the overbought technical read and the analyst target gap all point to the same discipline: HPE needs a beat with evidence, not a beat with vague AI language.

Trading read and one-year view

For traders, June 1 is a networking catalyst. A strong print probably needs three ingredients: Q2 revenue near or above the high end of guide, durable Networking margin, and a more confident Networks for AI order commentary. If those land together, HPE can argue that Juniper is accelerating the AI infrastructure thesis. If one is missing, the stock may struggle because the rally already pulled forward some of the good news.

For one-year investors, the cleanest stance is selective patience. HPE is more interesting after Juniper because networking gives it a higher-quality AI angle than commodity servers alone. But the investable version of the thesis requires proof that Networking can grow on a normalized basis, keep low-20s margin, help convert AI backlog, and lower leverage without starving product investment.

Bottom line

HPE’s June 1 report is really a Juniper referendum. The company can still talk about AI systems, backlog and servers, but the differentiated story is now the network: the switches, routers, automation layer and data-center fabric that decide whether AI infrastructure performs after the GPUs are installed.

If HPE proves that Juniper is turning networking into a durable AI profit pool, the stock deserves to be read differently. If not, the rally risks looking like another hardware cycle dressed up as an AI platform story.