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

Good afternoon, everybody. I'm Sami Badri, and I oversee Investor Relations here at Cisco. With me are Chuck Robbins, who serves as our Chair and CEO, and Scott Herren, our CFO. You can find Cisco's earnings press release and supplemental information, which includes reconciliations from GAAP to non-GAAP, over on our Investor Relations website. Once this call wraps up, the recorded webcast and the slides will also be posted there for you. During today's call, we'll be covering financial results on both a GAAP and a non-GAAP basis. When we talk about product results, we'll do so in terms of revenue, while geographic and customer results will be discussed in terms of product orders, unless we say otherwise. Every comparison we make will be on a year-over-year basis. I want to flag that our discussions today will contain forward-looking statements, among them our guidance for the second quarter and for fiscal year 2025. Such statements carry risks and uncertainties that are laid out in our SEC filings, in particular our most recent 10-K report, which spells out key risk factors that might lead actual results to differ materially from what our forward-looking statements contain. Regarding guidance, please also refer to the slides and the press release accompanying this call for additional details. Cisco won't be commenting on its financial guidance at any point during the quarter unless that happens by way of an explicit public disclosure. I'll now hand things over to Chuck.

Appreciate it, Sami, and thanks to everyone for being here. Cisco kicked off fiscal '25 on a solid note, posting $13.8 billion in first-quarter revenue, which landed at the top end of the range we had guided to. Underpinning that were healthy gains in annualized recurring revenue, remaining performance obligations, and subscription revenue, the last of which now makes up 57% of total revenue. Our recurring revenue base is a big part of what's powering profitability. Non-GAAP EPS came in at $0.91, above the top of our guidance range, helped along by the best non-GAAP gross margin we've recorded in over 20 years, which Splunk pushed even higher. Bringing Cisco and Splunk together, plus the terrific execution from our teams, is what's producing such strong operating leverage. Given how well we operated and the free cash flow we generated in Q1, we returned $3.6 billion to shareholders via buybacks and dividends, which keeps us on track with our pledge of consistent capital returns. Following the rising demand we observed as fiscal year '24 wrapped up, product orders climbed 20% year-over-year in Q1, and 9% on an organic basis. That marks an acceleration from the 14% product order growth posted just 3 months earlier, and it's a clear indicator that demand for Cisco's innovation is normalizing; now I'll walk through a few specifics of our first quarter, beginning with the demand backdrop. Enterprise product orders rose 33%, powered by especially robust results in the Americas and EMEA across a wide customer base. Momentum stayed strong in service provider and cloud, where product orders increased 28%, lifted by triple-digit growth in webscale. During Q1, our webscale customers placed AI infrastructure orders totaling more than $300 million. Our AI pipeline remains robust. We've secured additional design wins and feel confident we'll surpass our $1 billion target for AI orders this fiscal year from webscale customers. Coming off an exceptionally strong Q4 last quarter, public sector orders rose 2% year-over-year in Q1. U.S. federal spending declined year-over-year because of continuing resolutions discussions, the effects of the Fiscal Responsibility Act, and funds being redirected within U.S. federal; worth noting is that we view the deals in play with U.S. federal customers as merely delayed rather than gone; in EMEA and APJC alike, public sector order growth year-over-year was strong. Let me now offer some data on demand from a product standpoint for additional color. Our networking portfolio posted double-digit product order growth overall, led by switching, wireless, and Internet infrastructure. Turning specifically to data center switching, we've now seen 3 straight quarters of double-digit order growth, with acceleration from Q4 into Q1. That demonstrates our competitive strength in this crucial market, the strength of our Nexus brand, and our customers' build-out of private cloud infrastructure. We anticipate that this momentum will carry forward, since customers are expressing considerable interest in our 400-gig and 800-gig switches that are built on Silicon One. Orders in security increased by more than two times compared with a year earlier, fueled by Splunk's advanced threat intelligence capabilities. If we set Splunk aside, growth in security product orders came from our refreshed security strategy and our new product pipeline, which customers and analysts keep responding to favorably. Momentum persists around our newer security offerings such as XDR, Secure Access and Multicloud Defense, with more than 1,000 customers having deployed these products so far. During Q1, we landed a marquee win worth over 75,000 cloud security seats with a multinational IT services and consultancy company. That customer, for the most part, swapped out an SSE competitor for our solution and went from planning to implementation within 120 days, which shows how quickly our teams can roll out our solutions at scale. We also closed our first 7-figure Hypershield deal in Q1. Orders for collaboration products rose by double digits once more, propelled by demand for devices along with our cloud Webex Suite. Last, in observability, orders climbed by high single digits, lifted by our network assurance solution, ThousandEyes, and Splunk Observability. I'd also like to call attention to some of the progress achieved in bringing Splunk together with our business. At this point we have a dozen updated data integrations linking Cisco and Splunk throughout our security and networking portfolio, covering Secure Firewall, Catalyst Center, SD-WAN and ICE. Our joint selling motions between Cisco and Splunk carry on, including secure — Cisco Secure Network Analytics and XDR together with Splunk's SIEM, delivering enhanced capabilities to the security operations center. We're also still building our market-leading observability solutions to speed up full stack observability for the enterprise. On the whole, our Q1 results demonstrate sustained strong demand for Cisco technologies, pushed by the need for modern resilient networks as AI starts to scale. Results from our new global AI partner study indicate that IT partners worldwide foresee a transformative wave of AI technology demand, powered by infrastructure, cybersecurity and customer experience, which they believe will generate most of their revenue across the coming 4 to 5 years. Given how broad our portfolio is, we hold a unique position to take advantage of this AI technology demand while customers put money into their critical infrastructure to get ready for AI. Indeed, 3 distinct yet connected pillars make up the AI networking opportunity for Cisco, which we've laid out on a slide included in our quarterly materials. First, AI training infrastructure for webscale customers. AI networks require scalable, programmable, low-power switches featuring advanced load balancing and observability. These traits in our products are prompting hyperscalers to roll out the Cisco 8000 with Cisco's Silicon One G200 for maximum power efficiency in their back-end AI training networks. As additional evidence, we also secured a new super spine AI networking use case with one of our webscale customers during this quarter. Second, when it comes to AI network connectivity, during this quarter we closed additional platform deals with multinational enterprise clients. These customers are using our technology platforms to update and automate how their networks run, getting ready for connectivity at a large scale to AI applications. By drawing on breakthrough innovations such as AI-driven robotics and supply chain visibility that has no equal, our technology platforms spanning switching, routing, security, and observability will assist in enabling our customers. Third, AI network inference and enterprise clouds. A majority of the CIOs and technology leaders with whom we speak indicate that within the next 2 years, their organizations intend to adopt gen AI fully. According to the Cisco AI Readiness Index, however, just 13% of organizations report that their infrastructure is prepared for AI at present. To assist our customers in getting ready, we have introduced 2 new additions to our data center infrastructure portfolio: a new NVIDIA-based AI server along with AI PODs, which include NVIDIA's AI enterprise cloud-native software platform and are managed via Cisco Intersight, in order to make AI infrastructure deployment simpler and lower its risk. Our new AI server should start shipping next month, while our AI PODs can be ordered this month. Furthermore, Hyperfabric, our fabric-as-a-service offering, which should become available in early calendar 2025, makes infrastructure deployment and management even simpler and delivers real-time visibility into how networks perform. Through these solutions, we are delivering the strength of open hyperscale AI networking to the enterprise. We also keep integrating AI capabilities into our products.

Hypershield is our first genuinely distributed cybersecurity solution that's AI-native and woven into the network fabric, and it demonstrates how we deliver secure networking while making it simpler for customers to guard against threats that keep evolving. Another illustration of how we supply insights that help customers grasp their security maturity and show compliance with regulations that continue to change is Splunk's asset and risk intelligence offering. When it comes to our collaboration portfolio, we lately rolled out the Webex AI Agent, AI Agent Studio, and Cisco AI Assistant capabilities for Webex Contact Center. Advanced conversational intelligence and automation are what these solutions rely on to boost overall customer satisfaction; initial interest from customers has been strong, and we'll keep reinforcing our collaboration products by adding more AI capabilities. We're also boosting our own productivity through AI use within our services and customer experience organization. AI-enabled automation, supported by our expertise for quicker resolutions, now touches more than 60% of our Technical Assistance Center support cases, and that's driving higher customer satisfaction scores. On top of that, our internal AI Assistant for TAC collaborates with our engineers to raise quality and shorten time to resolution while freeing our people to tackle more complex and creative problem solving. In summary, when we examine what's happening with AI, 3 key things are taking place. First, there's substantial investment in back-end AI networks, with hyperscalers concentrating on training. Second, as enterprises seek to adopt and deploy AI, they must modernize and secure their infrastructure so they're ready for pervasive deployment of AI applications. Finally, mature back-end models paired with enterprise AI application deployment will produce greater capacity requirements on both private and public front-end cloud networks. Across all 3 of these major opportunities, Cisco already plays a significant role, and with the breadth of our product portfolio plus our trusted customer relationships, we are uniquely positioned to win. I'll now hand things over to Scott for more detail on the quarter and our outlook.

Appreciate that, Chuck. We had a solid Q1, with good momentum on orders, margins, and operating cash flow. For the quarter, total revenue landed at the top end of our guidance range, coming in at $13.8 billion, which is a 6% decline year-over-year. Keep in mind that Q1 of fiscal '24 was the final quarter with significantly elevated backlog shipments. Non-GAAP net income came to $3.7 billion, and non-GAAP earnings per share finished above the top end of our guidance range at $0.91. Let's dig into Q1 revenue a bit more. Total product revenue was $10.1 billion, down 9%. Service revenue came in at $3.7 billion, up 6%. Networking declined 23%, mainly because of the elevated shipment level we had a year ago. Like Chuck mentioned, we saw solid order growth across our networking products as customers burned through inventory and brought online the networking products we shipped to them last year. Security rose 100%, mainly fueled by growth in our threat intelligence, detection and response offerings, which include the offerings from Splunk, network security and SASE. If you set Splunk aside, security grew 2%, and our security portfolio took a particular hit from delays in U.S. fed spending. Collaboration fell 3%, due to declines in our on-prem Webex suite and collaboration devices, partly offset by growth in our contact center and CPaaS offerings. Observability climbed 36%, driven mostly by Observability Suite and growth in network assurance offerings. With Splunk excluded, observability grew 1% for the quarter. Now to our recurring metrics. ARR finished the quarter at $29.9 billion, up 22%, with product ARR growth of 42%. Total subscription revenue rose 21% to $7.8 billion, which works out to 57% of Cisco's total revenue. Total software revenue increased 24% to $5.5 billion, with software subscription revenue up 35%. Total RPO was $40 billion, a 15% increase year-over-year. Product RPO grew 24%, and total short-term RPO rose 15% to $20.3 billion. Q1 product orders increased 20% year-over-year. With Splunk excluded, product orders rose 9% year-over-year. Now for product orders by geographic segment. The Americas grew 17%, EMEA grew 26% and APJC grew 25%; within our customer markets, enterprise is up 33%. Service provider and cloud rose 28% and public sector was up 2%. Total non-GAAP gross margin was 69.3%, up 220 basis points year-over-year, beating the top end of our guidance range and hitting the highest level in over 20 years. Product gross margin was 68.9%, up 240 basis points, helped by Splunk, favorable product mix, productivity improvements and the onetime benefit of duty drawback, partly offset by price. Services gross margin was 70.3%, up 130 basis points. We stayed focused on profitability and financial discipline, with non-GAAP operating margin above the top end of our guidance range at 34.1%. On the bottom line, Q1 non-GAAP net income was $3.7 billion, and earnings per share were $0.91. Turning to the balance sheet. We closed Q1 with total cash, cash equivalents and investments of $18.7 billion. Cash from operations came in at $3.7 billion, which is a 54% increase. When it comes to how we allocate capital, we gave back $3.6 billion to shareholders over the quarter — that breaks down into $1.6 billion paid out as our quarterly cash dividend and $2 billion spent on repurchasing shares. We're still putting money to work both organically and inorganically in our innovation pipeline. In Q1, we completed 2 software acquisitions: DeepFactor, which speeds up innovation within the Cisco Security Cloud and strengthens our secure access offerings; and Robust Intelligence, which delivers improved protection for AI models across their entire life cycle, from development through production. On top of that, we shared our plan to acquire Deeper Insights AI in order to grow the technology footprint and engineering talent of our customer experience team, speeding up innovation and momentum behind Cisco's CX AI capabilities and services. These moves complement our internal R&D extremely well and fit with our approach of bolstering our standing in security and AI through focused strategic M&A. To wrap up, we kicked off the fiscal year with a solid quarter, marked by strong order growth, margins and non-GAAP EPS. Our attention stays on making strategic investments and speeding up innovation throughout our business so we can best seize the significant growth opportunities we see ahead, with disciplined expense management as the foundation for all of it. That powerful combination is what keeps driving our strong cash flow generation and our capacity to deliver significant value back to shareholders. Now to our financial guidance. For fiscal quarter Q2, here's what we're guiding. Revenue should land between $13.75 billion and $13.95 billion; non-GAAP gross margins should come in between 68% and 69%. Non-GAAP operating margin should fall between 33.5% and 34.5%. Non-GAAP earnings per share should land between $0.89 and $0.91. For fiscal year '25, we're guiding revenue between $55.3 billion and $56.3 billion. Non-GAAP earnings per share guidance should come in between $3.60 and $3.66. Across both our Q2 and full year guidance, we're using an assumed non-GAAP effective tax rate of roughly 19%. Sami, let's go ahead and open it up for Q&A.

Thanks, Scott. <<operator_script>> Operator, let's take the first question from the analyst queue.

I'll put both of my questions out there at once. So on cloud orders — cloud and service providers were up 22% excluding Splunk. In the prior quarter, that figure was just 2%. So what's behind such enormous order growth? And could you go deeper on the segments where you play in cloud — more detail on switching, routing, optical, and so on — where your strengths lie and how broad that is, meaning is it the established cloud providers, the cloud titans? Or do you, like some other vendors we see, concentrate on particular cloud providers rather than others? Then my second question concerns margins. Gross margin was very strong, and operating margin was very strong this quarter. Please discuss what's driving those margins and whether they can hold.

All right. Thank you, Tal. Appreciate you joining us, and appreciate the question; I'll address the first one, covering service provider and in particular cloud orders. If we take a step back and examine our service provider business, we felt good about it. First and foremost, I'd say, beginning with cable, pressure there persists, yet at the same time it's the smallest piece of the business and matters less as it shrinks. Turning to the telco side, we actually recorded what was likely low single-digit growth in telco, and it was good to see that. We observed some genuine — some activity taking place in Europe as well as in APJC, which was also encouraging. Then, shifting to cloud, or the webscale customers, that's the area where our order growth came in above 100%. What I'd share with you is that 4 among the 6 largest players were each individually above 100%. So this wasn't concentrated in a single customer. It was in fact 4 separate customers, each delivering triple-digit growth for us. Regarding the areas where we play, you brought up routing, switching, optical, and so on, and we do play in every one of those. I did note that we booked orders in Q1 for AI and webscale above $300 million, which keeps us firmly on track for the $1 billion and to surpass the $1 billion for the year. We participate — let me begin with AI. We participate — when we discuss AI, it's mainly — it's systems underpinning GPUs and back-end networks. I mentioned a super spine win, which refers to systems that in fact interconnect clusters in the back end. Beyond that, we also supply both systems and optics that deliver that same connectivity, along with the connectivity needed from the back end to the front end. So we hold a number of design wins in this space. We feel good about how we're progressing. And I'd simply say this is among the big benefits of our strategy, which is essentially to sell to our customers in whatever manner they prefer to consume the technology, whether through our silicon or through our systems. And I'd say that owning our own silicon is turning out to be a genuine edge over rivals. So that's something we feel good about. That said, I'd note that AI infrastructure accounted for only under 50% of the overall business we won from the webscale customers. So the conventional cloud side was still strong too, and we're happy with how that performed. Scott, would you like to handle the margin question?

Sure. Tal, we saw really robust margins. Naturally, our non-GAAP total gross margins came in at 69.3%, which, like Chuck mentioned before, is the best figure we've posted in 20 years. On the product side, non-GAAP gross margin reached 68.9%, which is an increase of 240 basis points. Several factors are behind this. Clearly, the boost compared to last year reflects bringing Splunk into the fold, which helps on the gross margin side, naturally. Beyond that, we also benefited from a more favorable product mix. The team keeps doing excellent work on what I'd describe as efficiency improvements; meaning they keep finding ways to lower costs within the cost of goods sold. Our supply chain group has performed really well there. That part is lasting; however, we did receive a onetime gain during Q1 within the cost of goods sold, tied to a multi-year initiative involving duty drawback; specifically, these are parts we brought in that faced a tariff. They went into finished goods and were later sent abroad. When that occurs, naturally, you become eligible to recover that money; we've been striving to reclaim the tariff paid when those items first entered the country. This effort has been underway for quite a while. In the quarter, the submissions got approved; as a result, we recorded a onetime gain that's pretty significant in Q1. That said, given the strong performance from the team, when I think about the remaining 3 quarters this year, and you can see this reflected in our Q2 guidance, I'd anticipate gross margins will keep settling within the 68% to 69% bracket for the entire year. And naturally, that will provide a boost to operating margin too.

My main question concerns the macro picture. I was hoping, Chuck, you might describe how conditions on the macro side shifted across the past 3 months. Then, as a brief follow-up, I'd like to dig deeper into that. Regarding the election, I'm curious whether any reluctance or effect on purchasing patterns has shown up in various verticals, given the uncertainty surrounding U.S. elections during the quarter.

Thank you, Matt. Turning to the macro picture, first and foremost, Scott brought up the U.S. federal challenges, and that connects to your second question. I'll address part of it there. The continuing resolution pressure hit the U.S. federal business, and so did the Fiscal Responsibility Act, which runs out at the close of this calendar year. Those two things weighed on federal. I believe we've heard the same from certain peers when they reported earnings. If you strip out Splunk, or if you actually examine our product orders excluding Splunk — meaning organic product order growth — and then take out U.S. federal on a year-over-year basis, the rest of the world rose in the mid- to high teens. That indicates we enjoyed quite balanced strength globally. By geography, Europe performed well, and Asia performed well. Excluding U.S. fed, the United States performed well. We experienced a very robust recovery in enterprise. That pleases us. I've already covered the service provider results from the quarter. So the sole genuine vertical I'd flag is the U.S. federal matter I just described. At this stage, I don't think we've observed anything else tied to the elections.

No. And Matt, what you have to keep in mind regarding those delays in U.S. fed is that's exactly what they are — delays. These aren't deals that got away because of it. What's creating the pressure is how the continuing resolution is being funded under the Fiscal Responsibility Act. So once we move beyond those continuing resolutions, we do see a portion of them returning; it appears the Republicans will win, taking both Houses of Congress along with the White House. As a result, my expectation is that a budget will be put in place fairly soon. So yes, I believe we will see — none of those are transactions that have been lost. And to drive home what Chuck just pointed out, total product order growth on an organic basis, and leaving out the U.S. fed impact, is up mid to high teens; so that lines up, I think, with what you'd have been expecting.

Yes, aside from that, everything was strong.

Let me ask about security. Hearing about the first Hypershield deal worth 7 figures during the quarter was encouraging. Could you expand on that a bit, what kind of deal it was? My understanding is it became available not long ago, so it seems quite early. But perhaps you could discuss how it's been received. Then second, I believe you said security orders had doubled compared with a year earlier, Splunk included. Would you share the organic figure too? And some qualitative color on the trends would be useful.

Yes. Thank you, Michael. Regarding your first question on Hypershield, the customer was a particularly big financial institution involved in early adopter planning, and that matches what we're observing. This solution has very high impact and proves quite effective in some of the most advanced data centers globally, I'd say. So we're seeing very large customers putting it into production. I think we'll see a gradual uptick in adoption among customers. At present we have something of a white glove program that we're using to assist these customers in getting up and running; so seeing that makes us quite pleased, because a commitment of that kind from a large bank signals that the technology is genuine. And often, they'll lead the rest of the enterprise when it comes to actually deploying these new security technologies. So that's an encouraging sign. On security orders, we ran into somewhat the same issue we faced with U.S. federal regarding overall product order growth. U.S. federal is our largest — it includes our biggest customer globally and likely our biggest security customer globally; and so our security orders rose in the low to mid-single digits excluding Splunk, the organic ones. And then — yet excluding U.S. federal, it matched the product data we provided a few minutes ago. So organic security orders, if you normalize away the federal year-over-year impact from the continuing resolution and the federal — the Fiscal Responsibility Act, they were also up in the mid- to high teens. We passed 1,000 customers having deployed these new technologies that we introduced over the past 12 to 15 months. So we're happy with the adoption that keeps happening there. And I think if you look — if you listen to the analyst or follow the industry, I think there's a widespread belief that our innovation pipeline and the technologies we've been bringing to bear will be highly impactful for our customers, and I'd say there's more to come; so stay tuned.

To start, if I could. You said — Chuck, I believe you said that 50% of the webscale orders involved AI, while the remaining 50% were more along the lines of conventional front end. Is that roughly what produced the enterprise order growth? Or did that figure include some campus recovery from a product standpoint? Then I'll hand the second one to Scott. Regarding gross margin, I appreciate the detail on a few of the effects during the quarter; but looking at Splunk, was there anything additional in those numbers? We're attempting to reason through, say, the contribution coming from Splunk subscription as opposed to term licenses; and how should we view that as we go through the rest of fiscal '25?

Thanks, David. If I'm reading your first question correctly, I believe you're asking whether the portion of cloud spend outside AI was what accounted for the enterprise growth we recorded. So I'll attempt to address that; when we refer to the enterprise segment, nothing sold to webscale customers is included, no matter what it is. The enterprise segment is therefore just enterprise customers purchasing whatever they purchase from us. That growth we recorded in enterprise came entirely from them. I'd add as well that the bulk of all webscale purchases made from us fell within the Internet infrastructure space. There was a small scattering on the non-AI side that you could attribute to, say, classic enterprise technologies, but relative to what they bought from us, it was quite minor. Internet infrastructure systems, optical, routing, switching, and so on made up most of it. The second one?

Yes. Regarding gross margin, David, that connects to what I mentioned before. We benefited from a favorable product mix, and Splunk contributed to that, which helped push margins up. In Q1, there was a onetime gain I just referenced related to duty drawback. This is a project we've had running for a while, recovering essentially import duties — tariffs we'd paid on components brought in, incorporated into goods, and subsequently shipped abroad. You're permitted to reclaim those. It took time to gather all the documentation, get it approved, and finally receive the refund. That occurred in Q1. So we had a onetime gain. It's a catch-up covering a couple of years of those. It won't recur for the remainder of the year. A small portion carries over — a bit in Q2 and somewhat more in Q3 and Q4. The team has also performed really well in continuously removing costs everywhere. That's another factor behind this — our current expectation for gross margins staying roughly in the 68% to 69% range through year-end; so consider Splunk as a benefit. That isn't new information. What I believe drove the better-than-expected result this quarter was far more the onetime effect from the duty drawback project.

Let me ask both of my questions at once. Chuck, during your prepared remarks you noted that the data center switching portfolio posted its third straight quarter of robust order growth. I'm just curious how you read that — does it mean enterprises are already beginning to prepare for their AI deployments? Or do you still think that's mostly ahead of you? What exactly is that order momentum telling you? Then the second one, on AI orders, you sound a bit more upbeat about beating the $1 billion target. But does that necessarily imply anything about revenue recognition this year? Is revenue recognition possibly somewhat higher than you originally expected when it comes to timing?

Thanks. As far as switching for data centers goes, what I think it shows is that our customers are still split fairly evenly on the question of which workloads go to the public cloud versus which get moved into private infrastructure they build themselves. We're seeing customers who are, by and large, upgrading their infrastructure — we talked about this on the last call. And we see enterprise customers who, even if they don't yet know what AI applications they'll roll out 6, 9, or 12 months down the road, are certain they'll need modern infrastructure in place to be ready. So we're seeing them spend now, ahead of time, to prepare for that. Yet this is exactly the same trend we've been seeing for several quarters. So I'd say it's partly in anticipation of that. As for the last piece of your first question, my view is that on the enterprise side, the bulk of this build-out still lies ahead of us. Turning to your second question about AI orders and the fact that we appear poised to exceed the $1 billion mark, the one point I'd make about the orders is this — and I'll let Scott speak to what it means for revenue. In my view, these customers operate in a highly dynamic market, and their requirements shift quickly. When we spot these chances and secure a design win—which we've talked about—those wins eventually turn into bookings, and from there they turn into revenue. Often a design win signals their belief that we're capable of building the exact silicon they need, integrating it into a system, and then handing that system over. Because of this, how long it takes shifts based on what we're asking of them. That's the reason you'll notice these things appearing in our guidance, and you'll notice them appearing as we keep executing. You've also heard me mention how important it is that we keep delivering on multiple consecutive commitments in this area, and I'm genuinely pleased with what the team is accomplishing at the moment. That leads into your question regarding revenue, Scott — you...

Right. And regarding the revenue question, as Chuck mentioned, it’s somewhat fluid when it comes to shipments. Clearly, we don’t recognize any revenue until the product actually leaves the door. A number of elements play into that, and most of them are outside our control. Still, we do expect this to drive revenue, in line with what we’ve previously stated, Samik, driving revenue for us more during the second half of the year. Most of the revenue we receive from that will come in the second half.

Thanks. For the first one, I'd appreciate an update, or maybe a deeper dive, on the partnership with NVIDIA that you announced very early on — I'd like to get a feel for how it's advancing in terms of traction and where it sits within the pipeline. I missed that part. You said you have an AI server. I'm just curious whether that came out of that alignment; and my follow-up is simple. You previously discussed the Splunk opportunities involving new potential customers being brought to Splunk. If you could give us an update on how that's going and your visibility into getting some of those deals closed.

Yes. Thanks, Simon. On the NVIDIA partnership front, my sense is that we've only just begun when it comes to bringing these offerings to customers. A few quarters back, we rolled out Hyperfabric — that's our fabric-as-a-service product, designed to bring public cloud-style ease to enterprise AI work, and it's set to become available during the first part of calendar 2025. During this quarter, though, we unveiled 2 new platforms. One is our AI compute platform, whose initial stage runs on NVIDIA GPUs; the other is AI PODs, which are also built around NVIDIA GPUs. Down the road, you'll see us add support for additional GPUs whenever the market calls for it, yet things with that partnership continue to go well; we're still at an early point. My take is that actual adoption of several of these technologies by large companies will begin in 2025. Regarding the second point, about news on Splunk's customer base, my assessment is that things keep moving along nicely. When you look at our position with Splunk — the deal finished in March, as I recall, and now it's November — that works out to about eight months. Gary would say, if you asked him, that a typical sales cycle for them runs nine months. So we haven't yet reached — and on top of that, we only kicked off this effort once they'd been with us for a while, probably during the second quarter. My sense, then, is that we're still at a very early stage. Here's what I can share on that. The way we determine which customers belong on this list comes down to specific customer traits that signal a strong fit for Splunk. Since then, roughly 1,500 additional customers have been added. There's another encouraging development too: several hundred Cisco partners have now completed Splunk training and are standing up dedicated Splunk practices. Some of them have even pursued strategic acquisitions of Splunk partners to speed up integration into their operations. So when it comes to that area, we feel quite good about the momentum.

Right. The one thing I would tack on is that, looking at the Splunk integration generally, on the top line and especially in ARR, things are moving along just as we anticipated, and at this stage we're actually doing better than we expected when it comes to profitability. So all in all, Splunk is tracking quite well for us.

Scott, turning back to RPO for a moment — what growth did Splunk see in RPO and in products RPO? Then, Chuck, how are you all viewing not only the potential for enterprise refresh but also the campus refresh chance now that Wi-Fi 7 is available, which I believe you even announced this week? And would that Wi-Fi 7 also spur any campus switch refresh?