Cisco Systems, Inc. (CSCO) Earnings Call Transcript & Summary

Good afternoon, everyone. Sami Badri here, Cisco's Head of Investor Relations, and with me are Chuck Robbins, our Chair and CEO, and Mark Patterson, our CFO. You can find Cisco's earnings press release and supplemental information, which includes GAAP to non-GAAP reconciliations, on our Investor Relations website. This call is also being live streamed on YouTube, LinkedIn and X. After the call, the recorded webcast and slides will be posted on the website as well. During today's call, we'll reference both GAAP and non-GAAP financial results. Product results will be discussed in terms of revenue, while geographic and customer results will be discussed in terms of product orders, unless stated otherwise. All comparisons will be year-on-year unless stated otherwise. Please note that today's discussion will include forward-looking statements, including our guidance for the second quarter and fiscal year 2027. These statements carry risks and uncertainties detailed in our SEC filings, particularly our most recent 10-K report, which identifies important risk factors that could cause actual results to differ materially from those contained in our forward-looking statements. Regarding guidance, please also see the slides and press release accompanying this call for further details. Cisco will not comment on its financial guidance during the quarter unless it is done through an explicit public disclosure. Now I'll turn it over to Chuck.

Appreciate that, Sami, and thank you all for joining today. Cisco, I'm glad to report, opened fiscal '27 on a strong note. Our strategy is being executed, and we're seeing demand climb as customers direct spending toward the infrastructure that the AI or SI era requires. My confidence comes not just from the results we're delivering right now but equally from the opportunities ahead of us. In Q1, revenue hit a quarterly record of $18.5 billion, up 24% year-over-year, and non-GAAP earnings per share also set a record at $1.40, a 40% year-over-year increase; those are the strongest top and bottom line growth rates we've seen in over 15 years. We're delivering strong operating leverage and more earning capability, and earnings are climbing at a quicker pace than revenue. Given that performance, we're lifting our fiscal year '27 guidance substantially, and Mark will walk everyone through the details shortly. The growth we expect generates healthy cash flow and supports steady capital returns for shareholders; in Q1 we gave back $3.7 billion through dividends and buybacks, representing 95% of free cash flow. Our solid start to fiscal year '27 demonstrates how essential networking and security are when combined as agentic AI accelerates. With customers trying to manage growing traffic costs and threats, they're turning to Cisco for our proven expertise, our security and observability solutions, and our services. Customers count on our help to scale, secure, and optimize their environments for the coming agentic AI era. Because our portfolio is both deep and broad, we can support customers no matter how they choose to deploy AI — in the public cloud, via neo clouds or sovereign clouds, on-premise, or at the edge. Our view is that the networking super cycle fueled by AI deployments is still in its early stages, and enterprise adoption has only just begun. Q1 brought strong demand: total product orders climbed 21% year-over-year, building on 13% growth a year earlier. Demand was also strong outside hyperscalers, with product orders up 14% year-over-year on top of 9% growth a year ago. We observed double-digit expansion in every geography and customer market as well. First-quarter enterprise product orders rose 15% compared with a year earlier, building on 4% growth in the prior-year period, and that came with double-digit gains across all geographies plus networking and security. Product orders increased 13% year-over-year, on top of 12% growth a year ago; public sector growth reached double digits in the Americas and APJC, as well as in networking, security and collaboration. Orders from service provider and cloud customers continue to be broad-based, climbing 47% versus a year ago; that follows 44% growth in the year-earlier period, including triple-digit growth in APJC and double-digit growth in the Americas and EMEA. Demand from telcos continues to look robust—Q1 orders rose roughly 20% compared with a year earlier, following 24% growth in the prior-year period. Now, looking at demand by product category: networking product orders increased 28% in Q1, marking our ninth straight quarter of double-digit expansion. In fact, every product family posted double-digit growth—service provider routing, compute, enterprise routing, wireless, data center switching and campus switching. These broad-based results reinforce our view, which I noted earlier, that we are still early in a multiyear, multibillion-dollar networking super cycle, and we are chasing that additional opportunity with focus and urgency. Overall, momentum in our hyperscaler segment remains strong. We anticipate AI infrastructure revenue hitting $9 billion in fiscal year '27—over double what we expect for fiscal year '26. As hyperscalers distribute coordinated AI workloads throughout their data centers, they're starting to deploy our scale in networking architecture and production environments. This architecture combines our Silicon One P200 deep buffered systems, our market-leading Acacia coherent optics, open line systems, and multi-rail optical systems that are coming soon. Together, these offerings provide a single unified compute fabric with deep programmability and built-in security. During the first quarter, we secured three fresh AI design wins. Two of those involve systems built on our P200, along with data center interconnect and scaling across various use cases. The remaining one involved 1.6 Terabit Optics serving a scale-out use case. In addition to demand from hyperscalers, Q1 brought quicker expansion in AI infrastructure orders from neocloud, sovereign and enterprise clients, and Nexus data center switches earmarked for AI deployments climbed 47% year-over-year. In that same quarter, we introduced a new partnership with Supermicro that broadens our secure AI factory with NVIDIA. By integrating Supermicro's high-density compute — both liquid and air cooled — into an NVIDIA cloud partner compliant full stack architecture, we deliver more efficient end-to-end infrastructure for enormous AI workloads. With agentic AI expanding, security risks are becoming more intricate and sovereignty debates are intensifying. Our goal is to provide trusted sovereign offerings that assist customers in confidently expanding their day center build-outs. This quarter, together with AMD and HUMAIN, we revealed that AI infrastructure using AMD GPUs, running on Silicon One based systems, is now live in Saudi Arabia. That enables HUMAIN to deliver GPU as a service for both training and inference scenarios. Responding to shifting customer requirements, we introduced our sovereign critical infrastructure portfolio in Europe twelve months ago. It spans networking, security, compute, collaboration and observability, and can be deployed for on-premise, air-gapped environments. During the first quarter, we made it available in Canada, and additional nations will follow, which provides government bodies and operators of essential infrastructure with more options, improved control, stronger security, and better resilience. In the same way, customers are focusing on upgrading their work environments for both people and agents. Demand for our campus networking offerings remained robust in Q1, as product orders grew 14% compared with a year earlier. Our newest portfolio continues to gain traction more quickly than previous launches, and we think this refresh cycle remains at a very early phase. Industrial IoT portfolio orders climbed about 50% year over year in Q1 and have now risen at double-digit rates across 10 straight quarters. Moving on to security: orders excluding Splunk increased in the mid-teens. On our security platform, firewall orders rose by more than 30% in the third quarter in a row. Cisco AI Defense helps customers protect against unmanaged AI usage throughout the enterprise, and it safeguards over 2 trillion tokens each month. Through our collaboration on Workday Agent Passport, with ServiceNow, and on DefenseClaw with NVIDIA OpenShell, we show how the ecosystem keeps expanding and how trust gets built into agents used in everyday operations. Earlier this week, we said we're bringing Anthropics Claude Mythos, a cybersecurity model, into Cisco AI Defense to enable adaptive AI red teaming. That gives defenders automated, ongoing testing across enterprise systems, letting them stress test AI systems using frontier adversary reasoning. Now to Splunk. The platform helps address some of the toughest challenges organizations confront in the AI era. One Splunk study discovered that the push to deploy agents is moving faster than enterprises can monitor and secure them. Through Splunk, we provide a common data platform that lets customers watch over agent behavior and enforce guardrails, while also assisting them in grasping and managing the tokenomics. This leaves them more confident and better able to predict outcomes as they seek to optimize the business value of AI along with its cost.

Working with a top hyperscaler, Splunk is looking into ways AI agents could assist security teams in investigating threats and coordinating responses with greater efficiency, while analysts stay in control. That effort extends our push to reinforce secure operations as threats become more sophisticated. There will be more to share later this month. Moving on to innovation. Our focus covers 4 vectors that frequently overlap: first, making network operations simpler; second, scaling infrastructure to support agentic AI; third, rethinking security and trust for agentic AI; and fourth, delivering intelligent workplace experiences. During the last quarter, we called out Cisco Cloud Control as a breakthrough in simplifying operations. Since it launched, over 6,000 customers have signed up, and they can boost productivity by reaching multiple Cisco and partner products via one user interface. We're building integrations inside the platform, among them AI canvas, Cisco IQ and the recently launched Cisco Model Foundry. Cisco IQ now counts more than 17,000 customers, nearly doubling quarter-over-quarter, with visibility into 23 million devices, speeding vulnerability tracking from days down to minutes. To scale infrastructure and raise the efficiency of AI factories, we recently announced plans with NVIDIA to develop the ultra dense small form factor pluggable, or USFP, optical module. We expect USFP to deliver 4x the bandwidth density and 1.3x of power efficiency of current models. We also intend to create an open industry USFP specification to drive adoption. Our progress in security and trust continues as we keep introducing resilient and quantum-safe innovations. We've expanded support for Live Protect across more of our campus networking portfolio, extended quantum readiness to WiFi 7 access points, and plan to provide quantum-safe communications across most of our core portfolio by the end of this calendar year. We keep developing intelligent workplace experiences, recently launching cloud-managed agents and Dialogue, an agentic harness in webex. Our teams keep driving this work forward through innovation, productivity and a commitment to leveraging AI tools. At the start of fiscal '27, we provided an AI agent for every employee; we expect to keep accelerating AI adoption, helping our employees execute; now I'll hand it over to Mark for more detail on the quarter and our outlook.

Thanks, Chuck. I'm pleased to share that we posted yet another record quarter, beating the top end of our guidance on revenue, operating margin and EPS, and also generating robust operating cash flow. For the quarter, revenue came in at $18.5 billion, an all-time high and up 24% year-on-year. Non-GAAP operating margin climbed to 37.8%, and non-GAAP earnings per share reached $1.40, a 40% increase year-over-year, which shows our operating leverage continues. Product revenue totaled $14.7 billion, up 32% year-over-year, while services revenue was $3.9 billion, up 1% year-over-year. On the product side, networking drove the gains, climbing 43% compared with a year earlier. We saw strength broadly across that portfolio, with AI infrastructure growing in the triple digits and data center switching plus campus networking each up by double digits. Security delivered a solid quarter, rising 7%, helped by Splunk, network security and SASE. Collaboration came in 8% higher, with devices up double digits and growth in < >. Now for recurring revenue: total RPO reached $46.2 billion, an 8% increase, and product RPO was up 7%. Total ARR closed the quarter at $32.7 billion, 4% higher, and product ARR also grew 4%. Subscriptions made up 44% of Cisco's overall revenue. Software revenue totaled $6.1 billion, up 8%. For Q1, product orders rose $0.21 compared with a year earlier, and the strength was broad across geographies, technologies and customer markets. Every geography posted double-digit growth: the Americas gained 18%, EMEA 12% and APJC 57%. By customer market, service provider and cloud led the way with a 47% increase, while enterprise and public sector also performed well, rising 15% and 13%, respectively. Total non-GAAP gross margin was 65.6%, a bit above the midpoint of our guidance and 250 basis points lower than a year ago, because hardware revenue grew much faster than software and services revenue. On a non-GAAP basis, product gross margin came in at 63.7%, a decline of 350 basis points, mainly because of a richer hardware mix and rising memory costs, partly offset by productivity gains and price increases. Non-GAAP services gross margin reached 72.9%, an increase of 220 basis points. With hardware making up a growing share of our total revenue, and hyperscaler revenue projected to more than double year-over-year in FY '27, we are achieving record scale and profitability across our business. For FY '27, we anticipate AI hyperscale revenue accounting for 12% of total revenue, compared with 6% in FY '26. Meanwhile, our overall workforce is down by a low single-digit percentage, which, on the basis of our revised FY '27 guidance, should lift revenue per employee by roughly 19%. These exceptionally large growth opportunities, pursued faster while still holding to financial discipline, let us raise operating margin to an expected 36% in FY '27 from 34.8% in FY '26. For the quarter, our non-GAAP tax rate came in at 18.5%. Turning to the balance sheet: we closed Q1 with cash, cash equivalents and investments totaling $15.8 billion. Operating cash flow reached $4.4 billion, a 36% increase, powered by robust revenue growth and net income. When you look at how we allocated capital, we gave back $3.7 billion to shareholders in the quarter. That breaks down as $1.7 billion in our quarterly cash dividend plus $2 billion of share repurchases. Our share repurchase program still has $6.1 billion available. Our focus stays on strategic investments in organic innovation so we can seize the meaningful growth opportunities we see ahead, along with M&A opportunities that are complementary. During Q1, we completed the WideField Security acquisition, which speeds up our Splunk agent SOC and helps halt identity threats at machine speed. Underpinning these investments is our dedication to disciplined spend management. That powerful combination continues to strengthen how solidly our business can generate earnings, underpins healthy cash flow, and enables us to give back meaningful value to shareholders. In summary, we posted record results for the quarter, with both revenue and profit exceeding what we had anticipated. These results reflect quicker innovation across all of our portfolio, good momentum, and careful cost control. Looking ahead, you can count on us to remain focused on lasting growth and financial discipline, while also improving operating leverage and sustaining capital turns. Turning now to guidance. For fiscal Q2, this is what we anticipate: revenue between $18 billion and $18.2 billion. Non-GAAP gross margin between 65% and 66%. Our non-GAAP operating margin should land between 36% and 37%, with non-GAAP earnings per share of $1.34 to $1.36. For the non-GAAP effective tax rate, we're using roughly 18.5%. Turning to our increased outlook for fiscal year '27: revenue between $74.8 billion and $75.8 billion, and non-GAAP earnings per share of $5.47 to $5.53. Sami, let's go ahead and open it up for questions.

Thanks, Mark. Before we begin the Q&A part of the call, I want to remind analysts to ask 1 question along with just a single follow-up question at the same time. Operator, could I go to the first analyst in the queue.

Congratulations on some truly impressive numbers. I suppose to kick things off, you're lifting your fiscal '27 AI revenue target to $9 billion, up from $7.5 billion last quarter. Could you take a little time to explain what has changed over those 90 days? Is it simply bigger deployments among existing customers, or do you now have improved visibility with TSMC regarding allocation? I'd really like to grasp what is getting better. Is there any skew between Silicon One and optics that you could shed light on? And my follow-up, really for you, Mark, the back half guide suggests revenues will in fact accelerate quite strongly to something like a $19.3 billion run rate, compared with the $18.1 billion in Q2. The EPS uptick in the back half is somewhat more muted, at above $1.38. Could you just walk through the puts and takes in the back half on the revenue line? That would be really helpful.

This is Chuck. I'll take the first one. Thank you, and thanks for the kind words. The teams are doing a fantastic job. On the $9 billion, I think it's a matter of being 90 days in and having a little more clarity. We had solid order growth in Q1. And we're simply further along in the year. So I think it's nothing more than continued strength in the flow from those customers, plus 90 days more visibility, which gave us the confidence to raise it from 7.5 to 9. Mark, on the second one?

Yes, absolutely, Amit. I'll add a bit of context here. Clearly, in the second half we're facing a tougher comparison, and visibility is somewhat lower than what we'd typically have in Q2. That said, taking a broader view, FY '26 was a record year for us on revenue as well as EPS, with the top line growing 12% and the bottom line 14%; and with our updated guide, both of those show significant acceleration. The top line goes from 12% all the way up to 19%, while the bottom line moves from 14% growth to 27% growth. So we're seeing really strong operating leverage and really good acceleration off what was a record year last year.

Good numbers. Chuck, can you speak to how much exposure you have to the neo cloud space, and is that included in the AI revenues you're guiding to? Some people think it could slow down as those players figure out their funding situation over the course of the year. So I'm curious what your exposure there is. And then more broadly, on the campus cycle. My follow-up is about where we stand in it. Is there — why would it slow down here when it comes to campus upgrades, given all the security concerns out there? A bit more detail on where we are in that cycle would be great.

Thanks, Ben. I'd describe our neo cloud exposure as relatively material. We definitely have business there. We've stayed very conservative when it comes to credit situations. And for the second part of that question, the neo clouds are not included in our AI revenue — that's simply the hyperscalers. Second, on the campus cycle, I think you're right. We had very strong results across the enterprise networking portfolio this quarter. I think double digits across all of the networking, every networking category. On the campus switching side, I believe we are — in a quarter, of the switching we sold, only 9% really came from the new portfolio. So again, we see this as very early. And I think your assumption about the impact of Mythos and some of the upgrade stuff is quite accurate. We don't expect any meaningful change, and likely just more customers looking at modernization for bandwidth reasons related to their rollout of agentic as well as just security issues and modernization related to the post Mythos era.

I’m hoping you’re able to hear me. There are 2 things I’d like to ask. First, Chuck, during the prepared remarks you mentioned that this is a multiyear — we are at the start of a multiyear cycle. So my question is what makes you confident enough to say that, when the alternative view would be that this is merely a cycle and we’re at the peak, et cetera. I’m just offering the alternative view; so what gives us confidence to believe this is a multiyear cycle and not simply a peak cycle of spending by hyperscalers? Then the second question is — we talk a lot about campus. But could you talk about how ready enterprise networks are for AI, or how they’re being deployed? Put another way, are you seeing enterprises preparing to upgrade their networks? Or what do you anticipate there?

Thanks, Tal. We do see this as a multiyear cycle, and we believe it clearly involves billions of dollars. Let me take you through the drivers that I think are key to why I’m quite confident this growth can be sustained. The first is success at hyperscale. Across the last 5 quarters, we’ve had 18 AI design wins. And obviously, we raised the AI guide to $9 billion, which is over twice what it was a year earlier. That one, we anticipate, will keep going. Next comes refreshing and modernizing campus networks. For several quarters now, we've recorded order growth in the double digits. We're still at a really early point — less than 10% through that transition. So that's driver number two. Driver three is broader and ties in somewhat: the opportunity after Mythos. Customers are beginning to tell us they're placing orders because of what early Mythos implications suggest. In Europe, for instance, we picked this up with service providers around the managed services CPE they must upgrade for their customers; yet although customers are eager to move fast, they also want to get it right. They're giving it the proper amount of time to plan and design the right architecture, since they understand that if they move forward with this, the architecture has to be exactly right — that's the third point. Number four is sovereign infrastructure. It's common knowledge that nations worldwide, starting with Europe, want infrastructure they believe they can exercise operational control over — and we're lucky to be among the few firms that have kept a strong on-premise portfolio. That means we can deliver. A year ago we announced sovereign critical infrastructure for Europe. We broadened that and brought it to Canada. Not long ago I gave advance notice of it for India, and we'll keep introducing it across the globe. Beyond what we're already doing alongside sovereign cloud players such as HUMAIN, G42 and additional sovereign countries, the fifth and last area is the pickup we're seeing in hybrid data center architecture with our enterprise customers. Those enterprise customers will run AI applications in the cloud, but at the same time they're making substantial investment in private data centers. They want open weight and open source models operating inside their own data center. They want to train on their own data within their own data centers; and that explains why, in 9 out of the last 11 quarters, we've recorded double-digit order growth in our enterprise data center networking business. We believe all five of these areas are multibillion-dollar, multiyear opportunities, and each one still seems fairly early in its development—that's the basis for our confidence in the durability. As for enterprise readiness for AI, I think it ties closely to what I just outlined, plus a strong emphasis on agentic security, agentic identity, and agentic observability, along with the new security architecture the AI era demands. In fact, a few hours before earnings I was at one of the big New York banks, where we spent a good while discussing exactly this subject. That's essentially the pattern we're observing overall.

I have just a few questions. First, regarding how strong the Asia Pac orders were this quarter — I assume the sovereign clouds account for most of that, but I wanted to see if you could give any more detail there. Then on the security business, what are you seeing with Splunk? I know we've gone through some license to subscription transitions, so if you could update us on where things stand there and how that's affecting security growth.

Thanks, Meta. I would say APJC showed incredible growth, and it was very balanced. If you look across the segments, enterprise came in above 30%, public sector above 70%, and service product and cloud grew by triple digits. So that was across every theater, every architecture, and every segment. And again, I think, to your point, it came from sovereign demand, public sector demand, telco investment, and strong results on the campus networking side. We also see them concentrating on the last day of support opportunity in some of the early activity following the Mythos refresh. As for — what was the second question again?

Security.

All right, security. Sorry about that. I'll have Mark go over the Splunk figures in a moment, but honestly we saw some genuinely encouraging signals on the security side this quarter. First off, our core Cisco Security orders, excluding Splunk, grew 15%, which is likely the strongest we've seen in 3 years. The firewall business is something the teams put real effort into refreshing and making competitive again. We posted a third straight quarter above 30% growth there. Then there's our Zero Trust secure access portfolio. That product didn't exist 3 years ago, and today we have 5,400 customers. We brought on close to 800 additional customers in Q1. It's a zero trust technology, and part of it also ties into AI defense. At the AI defense layer, we're now safeguarding 2 trillion tokens each month, and we're nearing 500 customers with that AI defense offering, so activity there looks good. We're leveraging Splunk as a platform for AI security. We're also seeing customers genuinely embrace security and networking converging around monitoring workloads, east-west traffic segmentation, and technologies such as Hyper Shield and Live Protect. So overall we see plenty of positive momentum on the security front. Mark, would you like to cover Splunk and the data behind it.

Yeah, and to give 1 more data point on the Asia Pac question you raised, we actually saw some really nice strength across a number of countries within Asia. Japan was really strong. China, Singapore, Korea, and Australia is in that mix too, and they had a really good quarter as well. So the breadth of overall growth was really good. And on the security side, I think when we exited last year and provided the outlook for this year, we said for Q1 to really expect low single-digit revenue growth, and that Splunk would turn positive and start improving as we move through the year, along with the overall security number really improving as we went through the year. So with the 7% we posted in Q1, I'd say we're a little bit above the expectations we set out for the year. And we feel good about turning the corner on both of those and beginning to see improved growth as we move through the year.

Starting with your first question, I was hoping you could give me a read on the supply situation as it stands today. Across the last two quarters or so, you sounded pretty confident in your ability to secure the volume you need, and then some. So I'm curious whether, with a few months of visibility now on the belt, anything has changed on that front — particularly around silicon and memory — and whether you're noticing any tightening. On the second one, perhaps piggybacking on what Amit raised a bit earlier. Gross margins are coming in significantly stronger than vision, even when measured against the 2 we're looking at in the 2Q guide. Could you help connect the pricing, cost and mix behind that? And then, regarding the bigger AI goal for the year, can you give any way to think about how gross margin moves through fiscal '27 and how we should view the exit run rate and the puts and takes.

Yes. I'll handle the first one, and then I might ask you to repeat the second one real quick. On the first one, just about supply — we feel really good. Management deserves credit for making the most of our strong balance sheet. We examined every stage of the supply chain — silicon, PCBs, power supplies, memory, and optics — and we're stepping in wherever necessary through long-term strategic supply deals. This quarter you saw us sign a new multiyear memory agreement. That has significantly raised our advance purchase commitments, and because of it we're encountering very few supply problems overall. Optics is likely the largest issue remaining, and we expect that to get better in the coming quarter. All in all, I'm feeling quite positive about our supply position and our capacity to fulfill not just the increased full-year guidance, but also, should demand come in hotter than expected, we'd likely be able to handle that too; and then the second item, apologies...

He was asking, and what he said was that gross margins are holding up better than maybe people had expected. What are the things contributing to that...

Yeah, I'd say the headwinds from component pricing remain the same, and mix is really the main factor when it comes to hardware growth overall, which is genuinely outpacing the software growth you're seeing. For the year, I'd note that most expectations called for a gross margin figure at just under — just shy of 65%, and we feel comfortable with that for the full year.

To start, I want to dig into the genuine headway you've made on AI back-end success. Inside systems value, can you give me a ballpark sense of how it breaks down between top of rack switching in the spine and the rallying user scale? I'm just looking for a rough order of magnitude. And then, as a follow-up, I'm curious whether you could bring us up to date on what portion of your growth has come from price changes as opposed to unit growth? I realize it isn't exact, but a rough picture there would be welcome as well.

Sure, I'll handle the first one, and Mark, you can field the second; we don't — I don't actually have the exact split between those technologies. My way of framing it would be that across the past 5 quarters, we've booked 18 design wins, spread over scale-out architectures — both systems — and also scale across, systems, and the optics we mentioned in the prepared remarks: 2 wins involving the P200 for scale across, plus a 1.6-terabit scale-out optics win. We do have operations in every market you listed—we simply haven't disclosed figures with that granularity. Mark, would you like to address gross margin?