HPE Stock Q3 Earnings Preview: Can AI and Juniper Keep the Growth Story Alive?

Hewlett Packard Enterprise has entered its fiscal third-quarter earnings with investors expecting another strong quarter from its AI and networking businesses. The stock closed at $50.85 on September 1, 2026, after gaining roughly 113% so far in 2026.

The problem is that expectations have moved just as quickly as the stock. HPE's Q3 guidance called for revenue of $11.5 billion to $12.1 billion and non-GAAP EPS of $0.88 to $0.93, while analyst expectations had moved toward roughly $12 billion of revenue and $0.93 of adjusted EPS.

Let's break down what Wall Street is expecting, why AI and Juniper matter, what numbers could move HPE stock, and what investors should watch beyond the headline beat.

HPE Q3 Earnings Expectations

The latest Wall Street consensus available on September 2 was about $11.97 billion of revenue and $0.93 of adjusted EPS. The table below compares these estimates with HPE’s guidance and last year’s result.

The revenue estimate is only 1.1% below the top of HPE’s range, while the EPS estimate is at the top. A result that only matches consensus may not impress investors.

There is also a comparison problem. HPE completed the Juniper acquisition on July 2, 2025. The year-ago quarter included only one month of Juniper, while Q3 FY2026 includes three months. Reported growth will therefore look stronger than the real growth of the combined business.

Management expects:

• Networking revenue growth of 73% to 78% as reported, but only around 10% on a like-for-like basis.
• Cloud & AI revenue growth in the high teens.
• Cloud & AI operating margin in the low-to-mid teens.

How to Judge the HPE Earnings Beat

The easiest way to judge the result is through three tests: Mix, Margin and Cash.

For example, $12.2 billion of revenue with stronger underlying Networking growth and a margin above 22% would be a good beat. The same revenue with weak unit sales and falling margins would be much less valuable.

Juniper Integration Faces Its Real Test

Phase one was buying Juniper. Phase two is proving that Juniper can become more valuable inside HPE than it was on its own.

Networking revenue reached $2.7 billion in Q2, up 148.2% from a year earlier. Most of that jump came from adding Juniper. On a like-for-like basis, Networking revenue grew about 10%.

The like-for-like order data were encouraging:

• Campus and branch orders grew in the high-20% range.
• Data-centre switching orders rose nearly 20%.
• Security orders grew in the mid-teens.
• Routing orders increased nearly 30%.
• HPE raised its FY2026 Networks for AI order target to at least $2 billion.

Cost integration is moving faster than planned. HPE expects to exceed its target of $200 million in annualised cost savings by the end of FY2026. The networking sales teams have also been combined.

Q3 must now show sales benefits, not only cost cuts. Investors should look for larger deals that combine servers, storage and Juniper networking. New customer wins and bigger deal sizes will matter more than another acquisition-led growth number.

Networking Margin May Decide the Reaction

Networking’s operating margin was 21.6% in Q2, down from 25.0% a year earlier. HPE partly blamed higher performance-linked employee costs. Even so, Networking has become an important profit engine.

The numbers show why this margin matters:

• Networking produced about $583 million of operating profit in Q2, based on $2.7 billion of revenue and a 21.6% margin.
• It generated roughly 38% of the combined operating profit of Networking and Cloud & AI, despite contributing only about 25% of HPE’s revenue.
• Every one percentage point of Networking margin is worth about $27 million of quarterly operating profit at the Q2 revenue level.

HPE expects Networking margin to reach the mid-to-high-20% range in FY2027. A Q3 margin above 21.6% would support that target. A fall below roughly 20% would make it look much harder to reach.

AI Backlog Must Become Revenue

HPE entered Q3 with an AI systems backlog of $5.9 billion, up nearly 20% from Q1. It received $1.8 billion of new AI systems orders in Q2, taking cumulative AI systems bookings to $16.4 billion.

The backlog mainly comes from companies and government-backed sovereign AI projects. These orders can be more profitable than basic server deals, but large sovereign projects may take longer to install and turn into revenue.

Backlog is not the same as revenue or profit. HPE still needs to get the required parts, ship the systems, receive payment and protect its margins.

Company and sovereign orders can include storage, networking, software and services. This gives HPE more ways to earn money from the same AI project.

Cloud and AI Margin Is the Server Test

Cloud & AI revenue rose 22.9% in Q2 to $7.7 billion. More importantly, its operating margin increased to 12.4% from 6.6% a year earlier.

That improvement more than doubled estimated operating profit from the segment, from about $414 million to $955 million. This was a much stronger signal than revenue growth alone.

However, HPE’s filing said Q2 growth came mainly from higher selling prices. Traditional-server unit sales increased only slightly. Part of the revenue growth therefore came from passing higher DRAM and NAND costs to customers.

This makes Q3 a price-versus-volume test. Higher revenue is less useful if HPE sells few extra units or earns a lower margin on each sale.

Management expects Cloud & AI revenue to grow in the high teens, with a low-to-mid-teens margin. A margin near or above Q2’s 12.4% would be healthy. A fall below 11% would be a warning sign.

Traditional Servers Are Not the Same as AI Servers

Server revenue rose 32.7% in Q2 to $5.5 billion, but investors should not label all of this as AI growth.

HPE said traditional-server orders more than doubled as customers replaced old systems and prepared to use AI. Training an AI model often needs large GPU systems. Running a trained model, known as inference, can also require high-memory servers in company offices and data centres.

That creates three different growth drivers.

1. Dedicated AI systems for training and large-scale inference.
2. Traditional server replacements as companies modernise ageing infrastructure.
3. Higher selling prices caused by more memory per system and more expensive components.

These growth drivers are not equally valuable. A multi-year replacement cycle and wider business use of AI could support demand for longer. Temporary price increases or customers buying early may not. HPE said it had not seen double-booking, cancellations or early buying in Q2. Q3 order and unit trends will test that claim.

Memory Costs Could Weaken Margins

DRAM and NAND are memory and storage parts used inside servers. Tight supply has raised their prices. HPE has responded by increasing product prices, signing longer supply agreements and adjusting server configurations.

Management expects these cost and supply pressures to continue into FY2027. It also said that further FY2026 revenue upside would largely depend on getting more components.

HPE held $9.03 billion of inventory on April 30, up 42% from $6.35 billion at the end of October 2025. Some of this stock was purchased for second-half AI installations and to secure scarce parts.

Investors should watch four areas:

• Whether price increases cover higher memory costs.
• Whether customers are placing orders early to avoid further price increases.
• Whether gross and segment margins remain stable.
• Whether the large inventory balance turns into shipments and cash.

Can GreenLake Make Revenue More Recurring?

Hardware sales can rise and fall as customers replace old systems. GreenLake aims to make HPE’s revenue more predictable by allowing customers to pay for infrastructure and software as they use it.

The latest GreenLake numbers were:

• Around 50,000 customers in Q2.
• More than 6.7 million connected systems, up from 5.3 million a year earlier.
• Net retention near 110%, meaning existing customers were spending about 10% more after accounting for upgrades and lost customers.
• An FY2026 annualised revenue run-rate, or ARR, target of $3.5 billion, up about 9% from $3.2 billion at the end of FY2025.

ARR is not full-year reported revenue. It estimates recurring revenue at an annual rate using the latest quarter. HPE did not give a fresh ARR figure in Q2, so investors should look for an update in Q3.

HPE Discover 2026: Orders or Announcements?

HPE announced several products at its Discover event in June, including:

• Closer integration between Juniper and Aruba.
• Wider use of Mist and Marvis networking tools.
• New switches for AI data centres.
• More GreenLake automation.
• Expanded AI infrastructure with Nvidia.

Some products became available during Q3, but others will launch in Q4, 2027 or later. The combined HPE and Juniper partner programme starts on November 1. The best near-term proof would be more orders and early customer wins, not immediate revenue from every product.

What HPE Needs to Raise FY2026 Guidance

HPE raised its full-year outlook sharply after Q2.

The full-year guidance still depends on a strong Q4:

• Revenue: The current growth target implies FY2026 revenue of $44.2 billion to $45.6 billion. If Q3 matches the $11.97 billion consensus, Q4 revenue would need to reach about $12.3 billion to $13.6 billion.
• Adjusted EPS: HPE earned $1.44 in the first half. If Q3 delivers $0.93, Q4 EPS would need to be about $0.98 to $1.08.
• AI shipments: HPE expects AI systems revenue to peak in Q4. It must show that parts are available and large installations remain on schedule.

A further guidance increase would require a solid Q3 beat, stable margins, reliable component supply and confidence that AI orders will ship on time.

Free Cash Flow Must Catch Up

HPE generated $708 million of free cash flow in Q1 and $915 million in Q2. That adds up to $1.62 billion in the first half. It must generate another $1.88 billion across Q3 and Q4 to reach its target of at least $3.5 billion.

Accounts receivable rose 19% to $6.29 billion in Q2, meaning more cash was still waiting to be collected from customers. HPE’s cash cycle improved by two days, but mainly because it took longer to pay suppliers. Inventory and collection times both increased.

The cash-flow target is achievable if HPE ships its inventory and collects customer payments on time. A strong EPS beat matters less if more cash remains tied up in inventory and unpaid bills.

Is HPE Stock Expensive Before Earnings?

HPE closed at $50.85 on September 1, up about 113% in 2026. Investors are no longer treating it like a low-expectation stock.

The comparison uses market data available on September 2. HPE trades below Dell and Cisco on all three measures. It also trades at about 19 times its minimum FY2026 free-cash-flow target of $3.5 billion, based on a market value of $67.36 billion.

The discount is not automatically a bargain. HPE has lower margins than Cisco, meaningful debt and exposure to changing hardware costs. Its 13-times forward P/E is also within the low-to-mid-teens range that has often defined HPE’s valuation. The market still wants proof that recent growth can produce steady cash flow.

HPE Q3 Bull, Base and Bear Cases

The ranges below are an analytical framework, not company guidance or a prediction of the share-price move.

The bull case needs strength in more than one area. A revenue beat without healthy margins or cash flow would not be enough.

Five HPE Earnings Numbers to Watch

1. Networking growth: Compare reported growth with the roughly 10% like-for-like target.
2. Networking operating margin: Measure it against Q2’s 21.6% and the FY2027 goal of the mid-to-high-20% range.
3. Cloud & AI operating margin: Compare it with Q2’s 12.4% and management’s low-to-mid-teens Q3 outlook.
4. AI backlog and bookings: Start with the $5.9 billion Q2 backlog, then judge whether orders and revenue conversion remain healthy.
5. FY2026 guidance: Watch the $3.35 to $3.45 adjusted EPS range and the target of at least $3.5 billion of free cash flow.

HPE Q3 Earnings Preview: Final View

HPE enters Q3 with strong networking orders, faster-than-planned Juniper savings, a large AI backlog and better Cloud & AI profitability. But the stock’s 2026 rally means an ordinary result may not be enough.

The strongest result would combine three things: Networking margin above Q2’s 21.6%, strong AI order conversion without weaker Cloud & AI margins, and free cash flow that keeps the $3.5 billion target on track.

If revenue beats only because higher memory costs pushed up server prices, the headline may look better than the business. If mix, margins and cash flow all improve, HPE will have stronger evidence that it is becoming more than a traditional hardware company.