Evercore ISI downgrades HP Enterprise stock rating on valuation

Evercore ISI downgraded Hewlett Packard Enterprise to In Line from Outperform while retaining its $65 target, citing a 158.5% year-to-date share gain to $62.08 and limited near-term room for further multiple expansion. HPE trades at 13x fiscal 2027 P/E versus a five-year average of 8x, while networking revenue growth of 10% lagged Cisco's 28% and Arista's 38%. Offsetting the valuation concerns, HPE reported fiscal Q3 2026 EPS of $1.11, raised fiscal-year revenue-growth guidance to 34%-37%, projected non-GAAP EPS of $3.75-$3.85, and expects more than $3.75B of free cash flow, supported by Cloud and AI demand.

Analysis

HPE has transitioned from a turnaround/Juniper-synergy rerating into an execution-sensitive compounder, leaving little room for a further multiple bid without evidence that acquired networking assets are producing incremental margins rather than merely revenue scale. The key market disagreement is not AI order demand but conversion quality: Cloud/AI systems typically carry lower gross margin, greater working-capital needs, and more volatile delivery timing than proprietary networking. That mix shift can leave EPS intact while reducing FCF conversion and suppressing the valuation premium investors are currently underwriting.

ANET remains the cleanest public-market expression of AI-networking demand because its higher-growth profile and software/content mix support superior incremental margins; however, its valuation makes it vulnerable if hyperscaler capex rotates from front-end training clusters toward inference, where architectures and purchasing cadence are less favorable. CSCO is a more indirect beneficiary: constrained supply or integration disruption at HPE can push enterprise refresh demand toward Cisco, but its upside is likely limited without a broader campus/networking recovery. EXTR is unlikely to capture meaningful AI spillover given its lower-end enterprise exposure.

Over the next 1-3 months, investor focus should shift from bookings to backlog aging, networking gross-margin progression, and cash conversion. A failure to show sequential margin improvement after integration costs normalize would likely trigger a derating toward mature infrastructure peers; conversely, a material lift in networking profitability or faster backlog conversion would invalidate the short thesis. Over 6-18 months, the relevant structural question is whether HPE can monetize a full-stack AI architecture rather than remain a lower-margin hardware aggregator.

Key Decisions for Investors

• Initiate a tactical HPE short or underweight versus ANET over the next 1-3 months; target 10-15% relative downside if HPE's next results show AI-led revenue growth without networking-margin expansion. Cover if networking EBIT margin improves materially for two consecutive quarters or FCF conversion exceeds guidance despite elevated AI mix.
• For lower-beta expression, buy ANET / short HPE in equal dollar amounts through the next earnings cycle. The pair isolates HPE's integration and mix risk from continued AI-networking demand; reassess if ANET reports a hyperscaler digestion signal or HPE demonstrates credible Juniper synergy realization.
• Do not add outright HPE longs solely on order commentary. Upgrade the thesis only if disclosed backlog conversion accelerates, receivables/inventory remain controlled, and segment margin gains confirm that revenue growth is translating into durable free cash flow.
• Monitor CSCO as a substitution beneficiary rather than a primary AI trade. A meaningful HPE supply or integration miss could create a near-term enterprise-routing share-gain catalyst for CSCO, but use post-earnings guidance and order-growth inflection as confirmation before initiating.