Cisco (CSCO.US) Q1 conference call: CEO says AI order fluctuations are normal and reaffirms that the multi-year network supercycle is in the very early stages
The Zhitong Finance App learned that Cisco (CSCO.US) announced the results for the first quarter of fiscal year 2027 (ending October 24, 2026) after the market on November 12, EST. Revenue for the quarter was US$18.52 billion, up 24% year on year, the fastest growth rate in more than 15 years, higher than LSEG's agreed expectation of US$18.1 billion; non-GAAP earnings per share were 1.40 US dollars, up 40% year over year, higher than market expectations of 1.32 to 1.33 US dollars, and the non-GAAP operating margin rose to a record high of 37.8%. The company simultaneously raised the 2027 revenue guide to 74.8 billion US dollars to 75.8 billion US dollars, and raised the AI infrastructure revenue forecast for hyperscale cloud service providers from 7.5 billion US dollars to 9 billion US dollars, more than double the actual level of the previous fiscal year. After the results were announced, the stock rose 8% after the market, then fell back down to about 1%; when the regular trading period closed, it rose 1.24% to $102.01. At the conference call, in response to the slight month-on-month decline in AI orders from UBS's large-scale customers, Chairman and CEO Chuck Robbins responded that AI orders are still growing strongly year over year. The business is naturally “lumpy” (lumpy) and fluctuates greatly between quarters, which is why the company changed to disclose AI revenue caliber; he reiterated that the company is still in the very early stages of a multi-year, multi-billion dollar network supercycle driven by AI deployment, and enterprise adoption has only just begun.
Focus and response: AI orders are cooling down month-on-month, management says “blockage” is normal
What is most noteworthy at this conference is the marginal change in AI order momentum: product orders for the quarter increased 21% year over year, down from 35% in the previous fiscal quarter; UBS analysts also pointed out that AI orders from hyperscale customers declined month-on-month.
Robbins responded that AI orders continued to grow strongly year over year. The first quarter of last year was only “a little over 1 billion US dollars”, and the total amount reached 9.3 billion US dollars for the full fiscal year. “This business has no normal linear relationship at all. It will be blocky. Some quarters will be big, and some quarters will be smaller”. This is the reason why the company decided to disclose AI revenue instead of order caliber. At the same time, he revealed that the G300 has begun to be shipped and entered the laboratory; 3 new AI designs won bids this quarter (18 in the past 5 quarters), of which the 1.6 terabit (T) optical module was the company's first client optical bid for this hyperscale customer.
CFO Mark Patterson added that of the $2.5 billion increase in annual revenue guidelines, 1.5 billion came from hyperscale businesses and 1 billion from core businesses.
Performance: The strongest growth rate in 15 years
Revenue of US$18.52 billion for the quarter reached a record high of US$18.52 billion, up 24% year on year; GAAP net profit of US$4.55 billion, up 59% year on year; GAAP diluted earnings per share of US$1.14; non-GAAP earnings per share were US$1.40, up 40% year over year. The non-GAAP operating margin rose to 37.8%. Revenue, operating margin, and EPS all exceeded the upper limits of the company's guidelines, and the profit growth rate was significantly faster than the revenue growth rate. Operating cash flow was US$4.4 billion, up 36% year over year.
Annual guidelines raised
The company raised its revenue guidance for fiscal year 2027 from US$72.2 billion to US$74.8 billion to US$75.8 billion, corresponding growth rate of about 19% (12% in FY2026); the non-GAAP earnings per share guide was raised from US$5.05 to US$5.11 to US$5.47 to US$5.53, corresponding growth rate of about 27% (14% for fiscal year 2026). The second-fiscal quarter led revenue of $18 billion to $18.2 billion and non-GAAP earnings of $1.34 to $1.36 per share, all higher than market expectations of approximately $17.88 billion and $1.23.
AI infrastructure revenue guidance doubles to $9 billion
The company disclosed that AI infrastructure orders for hyperscale customers in fiscal year 2026 totaled US$9.3 billion, about 4.5 times that of the previous fiscal year. Of these, about 60% were switching systems based on self-developed Silicon One chips and 40% were Acacia high-speed optical interconnection products. Acacia's single quarterly orders exceeded US$1 billion. Based on this, the company raised the 2027 hyperscale AI infrastructure revenue forecast from 7.5 billion US dollars to 9 billion US dollars, more than double the previous fiscal year, and the share of the company's total revenue will rise from 6% to about 12%.
Gross margin under pressure: AI hardware dilution continues
The non-GAAP gross profit margin for the quarter was 65.6%, down 250 basis points from the previous year; the product gross profit margin was 63.7%, down 350 basis points, mainly due to hardware growth far exceeding the increase in software, services, and memory costs, partly offset by productivity improvements and price increases; the gross profit margin for services was 72.9%, up 220 basis points. The company expects gross margin of slightly less than 65% for the full year. This is also the logic behind Morgan Stanley's logic of downgrading Cisco at the beginning of this month on the grounds that “AI hardware instantly recognizes revenue and dilutes gross profit before software revenue arrives.” However, due to scale effects, the non-GAAP operating margin still rose to 37.8%.
The network business and region are fully blossoming
Network revenue for the quarter was US$11.13 billion, up 43% year on year. Among them, AI infrastructure achieved three-digit growth; overall product revenue increased 32%, and service revenue increased 1%. Product orders increased 21% year over year, the Americas increased 18%, EMEA increased 12%, and APJC surged 57%; by customer market, enterprise orders increased 15%, public sector increased 13%, service providers and cloud customers grew 47%, and telecommunications increased by about 20%. Nexus AI data center switch orders increased 47% year over year.
Park renewal and safety business are both picking up
Park network orders increased 14%. The company said that next-generation combination products only accounted for about 9% of the switches sold this quarter, and the refresh cycle “is still very early”; orders for industrial IoT increased by about 50%, and have been growing by double digits for 10 consecutive quarters. After excluding Splunk, core security orders grew 15%, the best in three years; firewall orders increased by more than 30% for the third consecutive quarter; AI Defense protects more than 2 trillion tokens per month, and nearly 500 customers.
Supply is secure, shareholder returns remain high
The CFO said that the supply of silicon wafers, PCBs, power supplies, memory and optics is in good condition. A new multi-year memory agreement was signed this quarter. The biggest shortcoming was optics, which is expected to improve next quarter, and the supply capacity is sufficient to cover the guidelines after the increase. Shareholder returns for the quarter were 3.7 billion US dollars (dividends of 1.7 billion US dollars+repurchases of 2 billion US dollars), accounting for 95% of free cash flow. The remaining amount of the repurchase plan was 6.1 billion US dollars; 15.8 billion US dollars in cash and investments at the end of the period.
Full minutes of the meeting
Meeting time: November 12, 2026 (US stocks after market)
Company attendance: Ahmed Sami Badri (Head of Investor Relations), Chuck Robbins (Chairman and CEO), Mark Patterson (Chief Financial Officer)
I. Opening
Moderator: Welcome to the Cisco FY2027 First Quarter Earnings Conference Call. At the request of Cisco, this meeting will be recorded. Sami Badri, head of investor relations at Cisco, is now being asked.
Ahmed Sami Badri (Head of Investor Relations): Good afternoon everyone. I'm Sami Badri, head of investor relations at Cisco, and I was joined by Chairman and CEO Chuck Robbins and Chief Financial Officer Mark Patterson. Cisco's earnings announcements and supplementary materials (including GAAP and non-GAAP reconciliation tables) have been posted on the Investor Relations website. The call was also livestreamed on YouTube, LinkedIn, and X. Today's discussion will reference both GAAP and non-GAAP financial results. Unless otherwise specified, product results are discussed in terms of revenue, and regional and customer results are discussed in terms of product orders; all comparisons are made on a year-over-year basis. Today's discussion includes forward-looking statements, including guidance for the second fiscal quarter and fiscal year 2027. The relevant risks and uncertainties are detailed in our SEC documents, in particular the latest 10-K Annual Report. For more details on the guidelines, please refer to the presentation materials and press release that accompanied this call. Other than clear public disclosures, Cisco will not comment on financial guidance during the quarter. Now it's up to Chuck.
II. Chuck Robbins speech
Chuck Robbins (Chairman and CEO): Thank you Sami, and thank you all for attending today. I'm happy to report that Cisco made an impressive start to the 2027 fiscal year. We are firmly implementing our strategy and are seeing growing demand from customers to invest in the infrastructure needed for the AI era. What gives me the greatest confidence is not only today's performance, but also the opportunities ahead of us.
In the first quarter, we achieved a single-quarter revenue record of US$18.5 billion, up 24% year over year; non-GAAP earnings per share record of US$1.40, up 40% year over year. This is our strongest revenue and profit growth rate in over 15 years. We have shown strong operating leverage and stronger profitability, and the profit growth rate is faster than the revenue growth rate. Based on this strong performance, we are drastically raising our FY2027 guidance, which Mark will discuss in detail later.
Strong growth brings abundant cash flow to support our continued returns to shareholders. In the first quarter, we gave back $3.7 billion to shareholders through dividends and repurchases, accounting for 95% of free cash flow.
The strong start of the 2027 fiscal year confirms the importance of combining network and security with the accelerated implementation of intelligent AI (agentic AI). Customers are turning to Cisco to leverage our expertise, security, and observability solutions and services as they manage rising traffic costs and threats. They trust us to help them expand, secure, and optimize their operating environment for the smart AI era.
The depth and breadth of our product portfolio allows us to support our customers to deploy AI in any way — whether it's a public cloud, a new cloud (neocloud), a sovereign cloud, on-premise, or the edge. We believe we are still in the early stages of the AI deployment-driven cyber supercycle, and enterprise-side adoption has only just begun.
Demand was strong in the first quarter, and total product orders increased 21% year over year (based on 13% for the same period last year). Excluding hyperscale customers, product orders increased 14% year over year (9% in the same period last year). All regions and all customer markets experienced double-digit growth. Enterprise product orders increased 15% year over year (4% in the same period last year), and all regions and network and security businesses saw double-digit growth. Public sector orders increased 13% year over year (12% in the same period last year), and the American and APJC public sector, as well as networks, security, and collaboration, all achieved double-digit growth. Orders from service providers and cloud customers continued to grow broadly, surging 47% year over year (44% in the same period last year), with APJC growing by three digits and the Americas and EMEA growing by double digits. Telecom customer demand was healthy, and orders increased by about 20% year over year (24% in the same period last year).
By product category, orders for online products increased by 28%, achieving the ninth consecutive quarter of double-digit growth. In fact, all product lines have achieved double-digit growth — service provider routing, computing, enterprise routing, wireless, data center switching, and campus switching. These broad results support our judgment: we are in the early stages of a multi-year, multi-billion dollar cyber supercycle. We are executing with focus and urgency to seize this incremental opportunity.
The overall hyperscale business remains strong, and AI infrastructure revenue is expected to grow to $9 billion in fiscal year 2027, which is more than double the level of fiscal year 2026. Hyperscale vendors are currently deploying our scale-across network architectures in production environments to expand collaborative AI workloads to multiple data centers. The architecture combines a Silicon One P200-based deep buffered (deep buffered) system, market-leading Acacia coherent optical modules, and an open line system. A multi-rail (multi-rail) optical system will soon be launched. Together, these products provide a unified compute fabric (compute fabric) with deep programmability and built-in security capabilities.
In the first quarter, we won bids for 3 new AI designs: two from P200-based systems for data center interconnect (DCI) and scale-across scenarios; the third is a 1.6 terabit (T) optical module for scale-out scenarios.
In addition to hyperscale demand, we are also seeing an acceleration in AI infrastructure orders from new cloud, sovereign, and enterprise customers, with Nexus data center switch orders for AI deployments rising 47% year over year.
This quarter, we announced a new partnership with Supermicro to expand our “secure AI factory with NVIDIA built in.” By integrating SuperMicro's high-density liquid-cooled and air-cooled computing power into a full-stack architecture that complies with NVIDIA Cloud Partner standards, we provide a more efficient end-to-end infrastructure for large-scale AI workloads.
As smart AI expands, security threats have become more complex, and discussions surrounding sovereignty are heating up. Our focus is on providing our customers with trusted sovereign solutions to help them confidently advance data center construction. This quarter, we jointly announced with AMD and HUMAIN that AI infrastructure based on AMD GPUs and built on Silicon One systems has been put into production in Saudi Arabia, enabling HUMAIN to provide GPU-as-a-service for training and inference scenarios. A year ago, we launched a portfolio of sovereign critical infrastructure products in Europe, covering networking, security, computing, collaboration and observability, and can be configured as locally deployed, air-gapped (air-gapped) environments. We expanded to Canada this quarter and will cover more countries in the future, providing government agencies and critical infrastructure providers with more control, choice, security, and resilience.
Customers are also transforming workplaces for people and smart bodies. Demand for campus network solutions continued to be strong, and product orders increased 14% year over year in the first quarter. Our next-generation product portfolio is climbing faster than previous product launches, and we believe this refresh cycle is still very early. Industrial IoT product portfolio orders increased by about 50% year-on-year in the first quarter, and have maintained double-digit growth for 10 consecutive quarters.
Switch to a secure business. Orders excluding Splunk achieved ten-digit mid-teens (mid-teens) growth. Within the security platform, firewall orders increased more than 30% for the third consecutive quarter. Cisco AI Defense is helping enterprises prevent unregulated AI use within the enterprise, protecting more than 2 trillion tokens every month. Our collaboration with Workday Agent Passport, ServiceNow, and (transcribed as “DefensecLaw with NVIDIA OpenShell”, to be reviewed) shows how we can expand the ecosystem and embed trust into the day-to-day operations of intelligence agents. Earlier this week, we announced the integration of Anthropic's Claude Mythos cybersecurity model into Cisco AI Defense for adaptive AI Red Team testing, providing automated and continuous testing across enterprise systems, enabling defenders to stress test AI systems using cutting-edge advers' reasoning methods.
About Splunk. The platform is helping organizations solve some of the most pressing challenges of the AI era. A Splunk study found that the urgency of deploying agents is outpacing the ability of enterprises to monitor and secure them. With Splunk, we provide a unified data platform that enables customers to monitor and fence agent behavior while helping customers understand and manage the tokenomics (tokenomics) economy (tokenomics). This gives customers greater confidence and predictability when optimizing the business value and cost of AI. We're working with a leading hyperscale customer to explore how AI agents can help security teams investigate threats and coordinate responses more efficiently while keeping analysts in control. This will further reinforce our strategy to strengthen security operations as threats become more complex. We'll have more to share later this month.
Talk about innovation. We focus on four directions that often overlap: first, simplifying network operations; second, expanding infrastructure for intelligent AI; third, reimagining security and trust for intelligent AI; and fourth, providing an intelligent workplace experience.
Last quarter, we highlighted Cisco Cloud Control's breakthrough in streamlining operations. Since launch, more than 6,000 customers have registered to access multiple Cisco and partner products through a single interface, improving productivity. We're building integrations within the platform, including AI Canvas, Cisco iQ, and the recently launched Cisco Model Foundry. Cisco IQ currently has more than 17,000 customers, nearly doubling from month to month, with 23 million observable devices, reducing bug tracking time from days to minutes.
In order to expand infrastructure and improve the efficiency of AI factories, we recently announced plans to jointly develop a USFP ultra-compact pluggable optical module with NVIDIA. We expect USFP to provide 4 times the bandwidth density and 1.3 times the energy efficiency of existing products, and plan to establish open industry USFP specifications to drive adoption.
Advances in security and trust continue, and we continue to introduce resilient and quantum-resistant innovations: we've expanded Live Protect's support for more campus networking products, extended quantum readiness capabilities to WiFi 7 access points, and plan to provide quantum-resistant secure communications across most of our core product portfolio by the end of this calendar year.
We are also promoting intelligent workplace experiences, and recently launched the Cloud Management Agent and Dialogue, an agent framework in Webex. Our team continues to use AI tools to drive innovation, productivity, and engagement. At the beginning of fiscal year 2027, we equipped each employee with an AI agent, and we expect to continue to accelerate the application of AI to help employees improve their execution. Now it's up to Mark to share the details of this quarter's results and our outlook.
3. Mark Patterson's speech
Mark Patterson (CFO): Thanks Chuck. I am proud to report that we have achieved another record quarter, with revenue, operating margin, and EPS above the guideline limits, while operating cash flow is strong.
Total revenue for the quarter was US$18.5 billion, a record high, up 24% year over year. Non-GAAP operating margin rose to 37.8%, and non-GAAP earnings per share were $1.40, up 40% year over year, showing continued operating leverage.
Product revenue was US$14.7 billion, up 32% year on year; service revenue was US$3.9 billion, up 1% year on year. Product revenue growth was led by the network business, which accelerated to 43% year over year. Various product lines were generally strong. Among them, AI infrastructure achieved three-digit growth, and data center switching and campus networks achieved double-digit growth. The security business grew 7%, reflecting the growth of Splunk, Cybersecurity, and SASE. The collaboration business grew 8%, and devices achieved double-digit growth.
In terms of recurring revenue. The total RPO was 46.2 billion US dollars, up 8% year on year, with product RPO increasing 7%. The total ARR was US$32.7 billion, up 4% year on year, and product ARR also increased 4%. Subscription revenue accounts for 44% of total revenue. Software revenue was $6.1 billion, up 8% year over year.
Product orders increased 21% year over year in the first quarter, and markets in all regions, technical fields and customers were generally strong. All regions experienced double-digit growth: the Americas grew 18%, EMEA grew 12%, and APJC grew 57%. In terms of customer markets, service providers and cloud led the way in 47% growth; enterprises and the public sector grew 15% and 13%, respectively.
Non-GAAP gross margin was 65.6%, slightly higher than the guideline median value, down 250 basis points year over year because hardware revenue grew significantly faster than software and services revenue. The gross margin of non-GAAP products was 63.7%, a decrease of 350 basis points, mainly affected by rising hardware share and rising memory costs, partly offset by improved productivity and price increases. The gross margin for non-GAAP services was 72.9%, an increase of 220 basis points.
As the share of hardware in our overall revenue structure continues to rise, and hyperscale revenue is expected to more than double year-on-year in FY2027, we are driving record scale and profitability. In fiscal year 2027, we expect AI hyperscale revenue to account for about 12% of total revenue, up from 6% in fiscal 2026. At the same time, the total number of employees fell by a low single-digit percentage. Based on the revised FY2027 guidelines, per capita revenue is expected to increase by about 19%.
The accelerated realization of these huge growth opportunities, balanced with financial discipline, enabled us to increase our operating margin from 34.8% in FY2026 to 36% of our forecast for FY2027. The non-GAAP tax rate for the quarter was 18.5%.
On the balance sheet side. Our total cash, cash equivalents and investments at the end of the first quarter was US$15.8 billion. Operating cash flow was US$4.4 billion, up 36% year over year, driven by strong growth in revenue and net profit.
In terms of capital allocation, we gave back US$3.7 billion to shareholders this quarter, of which US$1.7 billion was quarterly cash dividends and US$2 billion was repurchases. The remaining balance of the buyback program is $6.1 billion.
We continue to focus on strategic investments in organic innovation to seize immediate significant growth opportunities, as well as complementary M&A opportunities. In the first quarter, we completed the acquisition of Widefield Security to accelerate the Splunk agent SoC and help stop identity threats at machine speed. These investments continue to be supported by strict expenditure management.
Overall, we achieved record quarterly performance, with revenue and profit exceeding expectations. This reflects the accelerated innovation, strong momentum, and strict cost management of our product portfolio. Looking ahead, we will continue to focus on achieving sustainable growth, maintaining financial discipline, and driving operating leverage and sustained return on capital.
Steering guide. The guidelines for the second fiscal quarter are as follows: revenue of $18 billion to $18.2 billion; non-GAAP gross margin of 65% to 66%; non-GAAP operating margin of 36% to 37%; non-GAAP earnings per share of $1.34 to $1.36; non-GAAP effective tax rate of approximately 18.5%.
The revised guidance for fiscal year 2027 is as follows: revenue of $74.8 billion to $75.8 billion; non-GAAP earnings per share of $5.47 to $5.53. Sami, on to the Q&A session.
Q & A session
Amit Daryanani (Evercore ISI): Congratulations for such impressive numbers. First question: You raised your AI revenue target for the 2027 fiscal year from $7.5 billion in the previous quarter to $9 billion. Can you talk about how things have changed in these 90 days? Are existing customers deploying larger scale, or do you have better visibility into TSMC's quotas? What's getting better? Is there an inclination between Silicon One and the optical module? My question to Mark: The second half of the year suggests that revenue will accelerate to a quarterly operating level of about $19.3 billion, while the second fiscal quarter was $18.1 billion, but the increase in EPS in the second half of the year was relatively flat, at just over $1.38. Can you talk about revenue side considerations for the second half of the year?
Chuck Robbins: I'm Chuck, answer first. Thank you. After 90 days, our visibility has improved. Orders grew steadily in the first quarter, and this year has been even longer. I think nothing gives us more confidence than the continued strong demand from these customers, plus 90 days of visibility, to raise the target from 7.5 billion to 9 billion. Mark, second question?
Mark Patterson: OK, Amit. Let's talk about color in a nutshell. Obviously, our comparison base for the second half of the year was higher, and our visibility was not as good as the second quarter. Overall, however, FY2026 was a record year for both our revenue and EPS, with revenue growth of 12% and profit growth of 14%. According to our revised guidelines, both have accelerated significantly: revenue growth increased from 12% to 19%, and profit growth increased from 14% to 27%. So this is a very strong operating leverage and acceleration on the basis of a record year.
Ben Reitzes (Melius Research): Good results. Chuck, can you talk about your exposure to the new cloud (neocloud) field? Is it included in the AI revenue you're directing? Some think this part may slow down as their funding situation becomes clear this year. Also, regarding the park refresh cycle: Where are we? Given all the safety concerns, why is the park upgrade slowing down here?
Chuck Robbins: Thanks Ben. I think our new cloud exposure is relatively significant (significant material). We do have a business in this area, and have always been very conservative when it comes to credit issues. To answer the second question: New cloud customers don't count towards our AI revenue — that only includes hyperscale customers. Regarding the park cycle, I think you're right. The overall performance of enterprise networks was strong this quarter, and all network categories were in double digits. In terms of campus switching, only about 9% of the switches we sold this quarter came from a new combination of products. So we think this is still very early. I think your assumptions about the impact and upgrade of Mythos are pretty accurate. We don't expect any major changes; it's probably just more customers modernizing for bandwidth reasons to match their smart AI deployments and the security modernization requirements of the “post-Mythos” era.
Tal Liani (Bank of America Research): Chuck, you said in preparation for your statement that we are at the beginning of a multi-year cycle. The question is, what makes you confident to say that? Another view is: this is just a cycle, and we are now at the peak of hyperscale vendor spending. What makes us believe this is a multi-year cycle and not the top of the cycle? Second question: Everyone is talking about the park, but can you talk about how the enterprise network is preparing or deploying AI - are you seeing enterprises preparing to upgrade the network? What are your expectations in this regard?
Chuck Robbins: Thanks Tal. We do think this is a multi-year cycle, and the scale is clearly multi-billion dollar. Let me give what I think are the key drivers underpinning the sustainability of this growth. The first is the success of the hyperscale business: we won 18 AI design bids in the past 5 quarters, and we raised our AI guidelines to 9 billion US dollars, more than double that of last year. We expect this trend to continue. The second is the renewal and modernization of the park network: we have seen double-digit order growth over multiple quarters, the conversion progress is less than 10%, and it is still in the very early stages. Third, related to the second point, but broader in scope, is the “after Mythos” opportunity: we're starting to hear customers attribute orders to Mythos' early impact. For example, in Europe, service providers needed to upgrade their CPE for managed services. Customers want to be quick, but if they want to do things right, they will spend the right amount of time planning and designing the right architecture, because if you want to do it, you have to do the right architecture. Fourth is sovereign infrastructure: countries around the world — starting with Europe — are seeking infrastructure where they can take control of their operations. We're fortunate to be one of the few companies that maintain a strong on-premise product portfolio. We announced our sovereign critical infrastructure product for Europe a year ago, then expanded and released the Canadian version. I recently pre-released it in India, and we will continue to promote it globally, in addition to collaborating with sovereign cloud providers such as HUMAIN and G42. The fifth and final one is the acceleration of hybrid data center architectures for enterprise customers: they are not only running AI applications in the cloud, but are also investing heavily in their own data centers. They want to run open weights and open source models in their own data centers and train with their own data. That's why in 9 of the past 11 quarters, our enterprise data center networking business orders grew by double digits. Each of these five fields is multi-year, multi-billion dollar, and they all feel like they're in a relatively early stage. That's where my confidence in sustainability comes from. Regarding companies' readiness for AI, I think this is largely what I just described, plus the real emphasis on smart security, smart identity, and agent observability, and the new security architecture needed in the AI era. A few hours ago, I was at a major bank in New York and had a long discussion on this topic. I think this is roughly what we saw.
Meta Marshall (Morgan Stanley Research): A few questions. Asia Pacific orders were strong this quarter, I guess mainly from Sovereign Cloud. Can you give me some more details? Also, there's the security business, and Splunk in particular — I'm aware of the license to subscription conversion. Can you update on the progress and its impact on security growth?
Chuck Robbins: Thanks Meta. APJC has indeed shown impressive growth and is very balanced: over 30% for enterprises, 70% for the public sector, and triple digits for service providers and cloud. Also, all regions, all architectures, and all customer segments are good. I think as you said, this is an excellent performance of sovereign requirements, public sector requirements, telecom investments, and campus networks. We're also seeing opportunities for customers to focus on “last day of support” (last day of support) and early “post-Mythos” refreshes. Regarding safety — Mark talks about Splunk's data, but we've seen some very good signs in terms of safety this quarter. First, core Cisco security orders after excluding Splunk grew 15%, which is probably the best level in three years. Through the team's efforts, the firewall business was refreshed and became competitive again, achieving more than 30% growth for the third consecutive quarter. Our Zero Trust secure access portfolio — products that didn't exist three years ago — now has 5,400 customers, with nearly 800 new customers added this quarter. At the AI Defense level, we now protect 2 trillion tokens every month, and nearly 500 customers use AI Defense. We're using Splunk as an AI-secure platform. We've also seen customers really appreciate the combination of security and networking in technologies such as monitoring workloads, east-west traffic segmentation, Hypershield, and Live Protect. There is a lot of positive momentum in terms of security. Mark, talk about Splunk's data.
Mark Patterson: Okay. Let me add an Asia-Pacific problem: We have seen strong momentum in many Asian countries — Japan is very strong, China, Singapore, South Korea, and Australia are included, all of which have had a very good quarter. The overall breadth of growth is excellent. In terms of safety, when we gave our outlook for this year at the end of last year, we said that revenue for the first quarter is expected to grow in low single digits, Splunk will correct and improve over time, and overall safety data will improve quarterly. So the 7% we announced this quarter is, I think, slightly higher than the expectations we set at the beginning of the year. We are satisfied that both businesses have begun to take a turn and will continue to improve throughout the year.
Joseph Cardoso (J.P. Morgan Chase Research): The first question is to understand the state of the supply chain. You've been quite confident about securing supply over the past few quarters. Now that you've been watching for a few more months, has anything changed? In particular, silicon wafers and memory, have they been tightened? The second question followed Amit's question: The gross margin performance was much better than expected, and even better than the guidance for the second fiscal quarter. Can you help break down the price, cost, and structural contributions? Also, considering higher AI goals, how will gross margin evolve throughout the 2027 fiscal year, and how should we think about withdrawing from operating levels?
Mark Patterson: I'll answer the first one. We feel very good about the supply. Using our financial strength, the team did extensive work to investigate up and down the supply chain—silicon, PCB, power, memory, and optics—and sign long-term strategic supply agreements where needed. You've seen us sign a new multi-year memory agreement this quarter, and our pre-purchase commitments have increased dramatically as a result. Overall, we didn't have any supply issues. The biggest shortcoming is probably optics, which we expect to improve next quarter. Overall, I'm relieved about the supply and ability to meet the revised guidelines, and if demand is stronger, we should be able to meet it. Second question --
Chuck Robbins: What he asked was that gross margins are more resilient than expected; what are the supporting factors.
Mark Patterson: Right. We are still facing the same headwinds: rising component prices, and structural issues — overall hardware growth far exceeds software growth. Looking at the full year, most expectations are for gross margins of slightly less than 65%, which is reassuring.
Simon Leopold (Raymond James): First question: Your progress on the AI backend network is impressive. In terms of system value, can you give an approximate ratio between top of rack (ToR) exchange, spine (spine) exchange, and scale-cross? Second question: Can you update the general split between price increases and shipments in growth?
Chuck Robbins: I'll answer the first one, Mark the second. We don't have this precise breakdown of technical aspects. Think of it this way: we won 18 design bids in the past 5 quarters, covering scale-out architectures (systems), scale-across (systems), and optics—as stated in the preparation statement, 2 P200 won bids for Scale-across, and 1 1.6t scale-out optical won the bid. So we're involved in all of the links we mentioned, but the numbers aren't broken down to that level. mark?
Mark Patterson: Again, we've talked about the roughly 50-50 split of optics and systems before. There will be fluctuations in any quarter, but this is generally the current state of affairs. As for prices and shipments, Simon, thanks for mentioning that this is an imprecise science. The estimate we gave a quarter ago was that the annual revenue was affected by the price by about 5 points. I think this is still a reasonable estimate for the whole year.
David Vogt (UBS Research): Chuck, first question. I know you don't want to be chased about AI orders. But by simple calculation, AI orders from hyperscale customers seem to have declined month-on-month this quarter. I know that the industry is somewhat blocky and not linear. Can you talk about how you see this trend developing over the year and how best you are communicating with customers? Potential bid wins for projects other than P200, G200, G300, etc. were mentioned later in the presentation materials. How are these conversations progressing? Ask Mark: Your AI revenue guide shows that the incremental growth rate from AI is about 2 points faster than we expected 90 days ago, and the price is relatively unchanged. Can it be understood that the park is just an incremental driver to supplement the revenue guidelines after the increase?
Chuck Robbins: David, when it comes to AI orders, we have achieved very good growth year over year. Looking back at last year, I can't remember the exact figure, but the first quarter was “a little over 1 billion dollars,” then it climbed quarter by quarter to reach 9.3 billion US dollars for the whole year. I don't think this business has a normal linear relationship. You're right, it will be blocky, some quarters will be big, and some quarters will be smaller. That's why we decided to switch to a revenue caliber — it's cleaner, and we don't have to wrestle with fluctuations like “how many orders did you sign in week 7 of the first quarter”. Regarding different technologies: The P200 is clearly very successful; in the scale-out field, the G200 has already won a large number of bids; I want to say that the G300 has now begun shipping and entering the laboratory, which is a good sign for us. We feel very good about this; the number of winning bids in the field of optics continues to increase. The winning bid for the 1.6T optical design this quarter was a client-side (client-side) bid, and our first client-side optical bid for this particular hyperscale customer. We are pleased with the continued progress in this area.
Mark Patterson: David, back to your numbers problem. We raised the guideline to 19% revenue growth: about 8 percentage points were due to hyperscale businesses, and 11 percentage points to core businesses excluding oversized businesses. Dismantling the overall increase: We raised $2.5 billion, of which 1.5 billion came from large-scale, and 1 billion from core non-hyperscale businesses. As for the drivers, you mentioned Canvas, but it's far more than just the park. If you look at all product families within the network, each achieved double-digit growth this quarter; all regions and all customer markets were also double-digit. So this is a broad foundation, and the core business continues to perform quite well.
Samuel Feldman (Karl Ackerman, BNP Paribas Research): We are Sam Feldman asking questions on behalf of Karl Ackerman. We learned that Silicon One is mainly used for AI products, and enterprise products still rely on external silicon. Has there been any change in the company's product integration schedule (the fiscal year cannot be clearly transcribed in the record)? Also, what does $9 billion in AI sales mean as a share of internal silicon?
Chuck Robbins: I answered the first one and the second one I'm not sure I know the answer. First: Silicon One has begun to be implemented in some products in our enterprise network product portfolio, and the target fiscal year of converting all products to self-developed silicon is still progressing according to plan. So we feel like we're headed well. Regarding the split between Silicon One revenue and non-Silicon One revenue: In the hyperscale sector, with the exception of optics, 100% of the $9 billion is based on the Silicon One platform.
Gianmarco Conti (Deutsche Bank Research): Congratulations on this quarter's results. About P200 and Scale-Across: You won bids for P200 scale-across designs with several hyperscale customers. What is Silicon One's catalyst for winning these positions? As the model gets bigger, how big will scale-cross get?
Chuck Robbins: Thank you. I think the key is performance and power consumption, that is, performance and power consumption. Add to that the field's focus on internal silicon diversity. If we want to talk about the size of this market, we have done a study and believe that bandwidth demand in this area will increase 14 times. So TAM is quite large and continues to grow. We are optimistic and satisfied with the number of winning bids for the P200.
Mark Patterson: One more thing: look at our product portfolio, which should be the most complete in the industry—silicon, Acacia optics, and multi-rail launched in the 2027 calendar year. I think we're a pretty good match on this point.
Adrienne Colby (Citif Malik, Citi Research): I'm Adrienne Colby, asking questions on behalf of Atif Malik. I want to ask about the optical business. In the past, you've talked about Acacia's order operation level of about 1 billion US dollars or slightly higher. Can you talk about whether it is expanding? As well as timelines for multi-rail line systems and USFP products.
Mark Patterson: Thanks Adrienne. In terms of optics, yes, we think this is quite an opportunity for us to grow rapidly. Our quarterly revenue for Optics reached $1 billion last quarter, which is a significant increase. The first quarter was also a record first fiscal quarter, and we saw demand continue to rise.
Chuck Robbins: The USFP platform will take a few years, probably a few years from now. The multi-rail system, as Mark previously mentioned, is in the first few months of the 2027 calendar year.
Ben Bollin (Cleveland Research): Can you talk about what kind of expectations should a service business have? It has recovered from its low point, but it hasn't really accelerated. When and how can it be improved? Another big problem: there seem to be concerns about production capacity on the raw material side, but there may be uncertainty about the number of data center shells that will be launched in 2027, 2028, and beyond. How do you think there may be a risk of “having production capacity and no place to put it” after 2028?
Mark Patterson: I'll answer the first one. Ben, thank you. In terms of service business, we saw a significant increase in hardware revenue — hardware revenue actually increased by more than 50% in the first quarter. The service business is characterized by on-time confirmations (ratable), and our hyperscale customers don't have much additional service purchases. But you'll see this business continue to improve: it's been corrected this quarter, and I expect it to continue improving throughout the year.
Chuck Robbins: I'm actually not worried about the demand side, especially data center issues in 2028 and beyond. I think the geopolitical issues will be solved, and given the needs we see today, these data centers will be built. So we don't think there's a major risk in terms of resources—anything that could hinder data center construction over the next two years.
Michael Ng (Goldman Sachs Research): First, ask Mark: Can you give me some more profit margin forecasts for the second fiscal quarter and the whole year? This quarter's profit margin was better than expected, given the rising share of hardware and hyperscale — what's improving? Chuck, question: Given the midterm election results and the risk that it will be difficult to build new production capacity in the US, are you seeing early signs that demand for data center AI networks is shifting outside of the US?
Mark Patterson: Michael, I'll answer first. The two major factors affecting gross margin are that we have seen component costs rise — we have dealt with it a lot, including raising prices and signing long-term agreements to lock in supply (and in some cases also lock in prices); the second is structure. Hedging these are a number of productivity improvements (including memory), as well as the scale and efficiency we have achieved far higher than our previous revenue scale. I actually think gross margin is basically on target: we are slightly above the guideline median value for the first quarter, and most people expect gross margin slightly below 65% for the whole year. I think this is a reasonable expectation; we should fall right there.
Chuck Robbins: Michael, on the question of data center demand migration: We haven't seen any signs of demand moving from the US abroad, and frankly, I don't expect this to happen. I think, on the one hand, there are still many states that have not announced the suspension of construction; of course, those states must resolve energy issues; we have also seen at least one governor publicly welcome data center construction, as long as it is carried out in the right way, responds to community concerns, and promises to maintain energy costs. I think this matter is too important to our country, too important to this industry, and the problem will be solved. I think the last thing I want to see is that the data center will end up being built outside of the US, and I think things will move in this direction. Of course, there are a lot of political games right now. Thanks for the question.
Chuck Robbins (Concluding Remarks): Thank you all for attending today. I want to emphasize that I am confident about Cisco's future and, quite frankly, the sustainability of the results we have seen. We have executed our strategy very well, the team has performed well, and the results are strong. The investments we have made over the past few years — and in some cases ten years — are turning into tangible momentum. I would also like to point out that the efficiency and scale of this business is greater than any time I have seen: if we calculate revenue according to the guidelines we have given today, we would have added $19 billion in revenue compared to two years ago, and the number of employees is lower than two years ago. This efficiency has never been seen before, and the team has done a fantastic job scaling the business and increasing productivity. We are confident in the message our customers and partners are delivering. We are indeed in the very early stages of the multi-year networking supercycle, and we are capable of continuing to seize the opportunities it brings to Cisco. I believe we have the right strategy, innovation, and team to focus on turning this into continued profitable growth. Thank you to the team for their continued execution and for the trust of customers, partners, and shareholders. Sami, give it back to you.
Ahmed Sami Badri (Head of Investor Relations): Thanks Chuck. The next conference call will announce the results for the second quarter of fiscal year 2027 at 1:30 p.m. Pacific time and 4:30 p.m. ET on February 10, 2027. This concludes our meeting for today. For playback, call (US and Canada) (800) 839-2232, (outside the US) (203) 369-3662. Thank you all for participating, and see you soon.
Moderator: Thank you all for participating in today's conference call. This is the end of the meeting, everyone can disconnect.