HPE's Q2 FY2026 results blew past Wall Street estimates by nearly $1 billion, powered by explosive growth in AI servers and networking, forcing the company to pull its 2028 financial targets two years forward.
The AI infrastructure trade finally showed up on HPE's bottom line in a way that could not be argued away. On June 1, shares of Hewlett Packard Enterprise surged roughly 37% in a single session after the company posted record second-quarter revenue of $10.68 billion, up 40% year over year and nearly $900 million ahead of the $9.79 billion analysts had penciled in. It was the company's biggest earnings beat since 2018, and the stock move was one of the sharpest single-session gains in HPE's post-spin history. The market was not pricing in this outcome.
Two product lines drove the outperformance. Server revenue climbed 32.7% to $5.5 billion as enterprises and government customers continued pulling forward AI infrastructure buildouts. Networking was the more striking story: that segment posted 148% revenue growth to $2.7 billion, supercharged by HPE's integration of Juniper Networks, which the company closed in July 2025. Juniper's QFX data center switches, PTX routers, and AI-native operations software are becoming the connective tissue inside enterprise AI deployments, and HPE is the vendor capturing that spend.
The company reported a total AI backlog of $6.3 billion, spanning both AI systems and networks for AI. Critically, 61% of that backlog is secured from government and large enterprise clients rather than hyperscalers. That customer mix matters. When hyperscalers dominate an AI hardware vendor's order book, concentration risk is real and cycles can be violent. HPE's skew toward sovereign and institutional buyers signals more durable, contract-backed revenue than the lumpy spot-market dynamics that can whipsaw competitors.
Management did not just beat the quarter. It pulled its 2028 long-term financial targets two years forward, projecting them achievable by 2026. Full-year revenue growth guidance jumped to 29-33%, up from an already ambitious 17-22%. Annual EPS guidance moved to $3.35-$3.45, a dollar-per-share increase over prior guidance of $2.30-$2.50. Those are not incremental revisions. They reflect a fundamental reassessment of demand duration, and they force institutional investors to recalculate valuation models on the fly.
Networking margins are expected to hold in the low 20% range for the fiscal year, while Cloud and AI margins sit in the 7-9% band. The margin profile here tells an important story about the composition of HPE's AI revenue. Server and AI system sales carry thinner margins by nature; networking and software-attached services carry more. As HPE's Juniper-powered networking revenue grows and the backlog converts, the question for the next several quarters is whether the mix shift improves blended margins or whether competitive pressure forces pricing concessions in the server segment.
How HPE Stacks Up Against Dell and Super Micro
HPE's results arrive one week after Dell Technologies surged 33% on its own AI server blowout, where Dell reported AI-optimized server revenue of $16.13 billion in its most recent quarter, up 757% year over year. Dell is operating at a different scale in the AI server market and growing faster in raw unit terms. Super Micro, meanwhile, posted Q3 FY2026 revenue of $10.24 billion, up 123% year over year but below the $12.45 billion the market expected, causing its stock to underperform the hardware cohort.
The three companies represent distinct bets within the infrastructure stack. Super Micro is the high-velocity, direct-to-hyperscaler GPU server assembler. Dell competes across servers, storage, and networking at enterprise scale, leveraging its formidable direct sales force. HPE's differentiated angle is networking. No other pure-play infrastructure vendor has Juniper's depth in data center switching and routing inside the same sales motion as AI compute servers. That integration is either a margin catalyst or a complexity liability, depending on execution. So far, execution is validating the thesis.
Structural Demand or Episodic Boom
Skeptics of AI infrastructure spending have argued for two years that enterprise capex cycles would eventually pause, that hyperscaler commitments were one-time catch-up spending, and that hardware vendors were benefiting from a bubble rather than a sustained upgrade cycle. The argument has been getting harder to sustain. Alphabet, SoftBank, and Microsoft have all accelerated data center commitments in 2026. HPE's backlog composition, weighted toward enterprises and governments that have longer procurement cycles and contractual obligations, suggests the current demand environment is structural rather than episodic. These are not speculative orders from startups experimenting with LLMs. They are infrastructure commitments from institutions that have already deployed AI at scale and are expanding capacity.
For investors tracking AI infrastructure as an asset class, HPE's move raises a timing question. After a 37% single-session jump, the stock is no longer cheap relative to its historical multiple. The more interesting signal may be what this result implies about the rest of the hardware food chain: optical interconnects, power distribution, liquid cooling, and storage vendors tied to the same data center buildouts. HPE's backlog conversion in the next two to three quarters will serve as one of the clearest real-time reads on whether enterprise AI infrastructure spending maintains its velocity into the second half of 2026. Watch the networking segment's order growth rate, not just revenue, because that leading indicator will tell you whether the backlog runway is extending or compressing.
Also read: Alphabet is raising $80 billion in equity to win the AI infrastructure arms race and Warren Buffett is buying in • Anthropic's Claude Opus 4.8 outpaces GPT-5.5 on benchmarks while making AI agents smarter and more honest • Meta AI chatbot exploit allows hackers to hijack high profile Instagram accounts
