Palo Alto Networks Inc (PANW) Q3 2023 Earnings Call Transcript

Welcome to the earnings call on March 23, 2023, at 1:30 PM Pacific Time. Joining me today are Nikesh Arora, our Chairman and CEO, and Dipak Golechha, our CFO. After the prepared remarks, Lee Klarich, our Chief Product Officer, will participate in the Q&A session. You can find the press release and additional information relevant to today's discussion on our website at investors.paloaltonetworks.com. Please click on the link for Events and Presentations to access the investor presentation and supplementary information. During this call, we will be making forward-looking statements regarding the company's business and financial performance, which are subject to risks and uncertainties. We do not have an obligation to update these statements. Please review the press release and our recent SEC filings to understand these risks and uncertainties. We will also discuss non-GAAP financial measures, which should not replace GAAP financial measures. The most directly comparable GAAP metrics and their reconciliations can be found in the press release and the appendix of the investor presentation. Unless stated otherwise, all results and comparisons are based on fiscal year-over-year data. Additionally, management will be participating in the Bank of America Global Technology Conference on June 6th. I will now hand the call over to Nikesh.

Thank you for being here today. Good afternoon, everyone. I appreciate you joining us for our earnings call. Once again, our teams have achieved a balanced quarter regarding both top and bottom-line performance in the current economic landscape. In the third quarter, our billings increased by 26% compared to last year, while revenue grew by 24%, and remaining performance obligations outpaced these figures with a 35% increase. Our non-GAAP operating income and adjusted free cash flow over the past twelve months both grew by around 60% year-over-year, and we maintained profitability based on GAAP for the fourth consecutive quarter. Let's discuss the macro environment. Trends of cautious spending, deal scrutiny, and cost consciousness remain prevalent and are increasingly widespread among our customers. Despite this, we are proactively addressing challenges and refining our focus on business value and security outcomes for our customers. From a technology perspective, there have not been significant changes. Trends in cloud adoption, automation, and hybrid work continue with slight variations. Although network transformations have long cycles, they are still being pursued as they offer cost savings and fit into most customers' modernization strategies, driving sustained demand for SASE and firewall products. The theme of consolidation around platforms continues to emerge, and we are in a strong position to provide relevant solutions. Recently, interest in AI, particularly through advancements like ChatGPT and Generative AI, has returned to the forefront. While AI can facilitate superior security outcomes in real time; it also presents potential threats as it can be exploited to generate attacks. We are actively addressing these issues and will delve deeper into them today. On the hardware front, growth in the industry remains limited. Though the supply chain crisis has mostly subsided, we are seeing a shift towards software and cloud-based solutions amid CapEx constraints. Meanwhile, threats remain active and current trends indicate ongoing security challenges, as highlighted in our recent Cloud Threat Report, which reveals that security teams take, on average, around six days to resolve alerts, while threat actors exploit vulnerabilities within hours. Regulation continues to garner attention, with various governments increasing mandates around cybersecurity. This trend reflects a growing focus on organizational accountability for cybersecurity, necessitating further investments from our customers. Economic conditions are prompting some customers to seek efficiencies within their businesses, including in cybersecurity processes. When I previously mentioned the importance of platform architectures, there was some resistance, but our three-platform approach has gained traction, allowing us to offer simplicity and enhanced security. Our customers are increasingly scrutinizing deals, leading to proactive discussions involving CFOs and procurement teams. As discussions regarding financing become more routine, we are leveraging our strong balance sheet to better accommodate customer needs. This fiscal year, we have focused on operating more efficiently. Our teams have adapted well, and alongside the easing supply chain challenges, we've significantly shifted our operating model. We have optimized hiring practices and streamlined our go-to-market strategies while still building teams in key innovation areas. This has resulted in improved operating margins and guidance for increased free cash flow margins. We are witnessing a trend towards platformization in cybersecurity driven by our commitment to providing superior outcomes for customers through a cohesive, robust portfolio. We continue to see success in driving larger platform transactions, with notable year-over-year increases in booking amounts. The value for our largest customers has grown steadily, and more customers are engaging with all three of our platforms, reinforcing our belief in the opportunity to shape outcomes effectively. Now, regarding our network security platform, we have been recognized as a leader in Gartner's recent Magic Quadrants for both Security Service Edge and SD-WAN, showcasing our innovative capabilities and market leadership. Our SASE segment is one of the fastest-growing areas in cybersecurity, with over 4,200 customers. Recently, we've secured significant transactions, including a comprehensive transformation for a global beverage company and notable deals with tech firms to modernize their network security strategies. We also combined our SASE sales organization with our core sales efforts, allowing us to engage more effectively as we saw demand grow. This quarter also marked the launch of our AI-powered SASE solution, aiming to improve IT operations. In the broader network security landscape, we foresee an ongoing shift towards software solutions. Our VM-Series firewall offerings have been particularly successful, reflecting a clear preference for software deployments amidst budget constraints. This year, VM-Series bookings are significantly up year-over-year, and we are experiencing a shift in the revenue contribution from software, which now accounts for 30% of our product revenue. AI advancements are encouraging a rapid transformation across our security products. Our integration of AI and machine learning over the years has positioned us as a market leader, enhancing detection and prevention capabilities across our product lines. We continue to invest in AI applications within our network security stack, analyzing millions of unique telemetry objects daily and leveraging our data to fend off billions of attacks. In Prisma Cloud, usage and customer demand for platform solutions are increasing, as evidenced by customers adopting foundational security modules. We are poised to meet industry needs with an expanding offering of cloud security capabilities and certifications. Our Cortex segment has reached significant milestones, achieving $1 billion in bookings in the last twelve months, reflecting our growth and success in SOC automation and innovation. We're excited about our XSIAM platform, which aims to revolutionize SOC approaches. It has shown early success and is set to grow rapidly during its launch phase, signaling a new focus on AI-driven security transformations. We are fortunate to be operating in a resilient part of the technology market, and despite the challenges, our team continues to execute effectively. Our strategy of consolidating capabilities and simplifying cybersecurity architectures has proven successful. Looking forward, we anticipate integrating generative AI into our products, which will enhance detection efficacy, improve customer engagement through intuitive experiences, and drive operational efficiencies within our organization. As we continue to execute our plans, we see potential for greater efficiency and operating profitability, laying the groundwork for future growth. I will now hand the call over to Dipak to discuss our Q3 performance and guidance.

Thank you, Nikesh, and good afternoon, everyone. For Q3, revenue was $1.72 billion and grew 24%. Product revenue grew 10%, total service revenue grew 29% with subscription revenue of $838 million, growing 31% and support revenue of $495 million, growing 25%. Moving on to geographies, we saw revenue growth across all theaters with the Americas growing 24%, EMEA up 23%, and JPAC growing 24%. The strength of our next-generation security capabilities continues to drive our results. With NGS ARR of $2.6 billion growing 60%. We saw strength across all three platforms: network security, cloud security, and security operations. We delivered total billings of $2.26 billion, up 26% and above the high end of our guidance range. Total deferred revenue in Q3 was $8.1 billion, an increase of 38%. Remaining performance obligation or RPO was $9.2 billion, increasing 35% with current RPO just under half of our RPO. Our non-GAAP earnings per share was significantly ahead of our guidance, growing 83% year-over-year. We again delivered strong cash flow in Q3 with trailing 12-month adjusted free cash flow of $2.8 billion, growing 68% year-over-year. Moving on to the rest of the financial highlights. Non-GAAP gross margin of 76.1% was up 320 basis points year-over-year, driven mainly by a higher software mix, reduced supply chain costs, and some efficiencies in customer support. Our non-GAAP operating margin of 23.6% increased 540 basis points year-over-year. In addition to improving gross margins, slower headcount additions contributed to our operating leverage. Based on our performance in Q3, we are raising our fiscal year '23 non-GAAP operating margin guidance. Non-GAAP net income for the third quarter grew 86% to $359 million or $1.10 per diluted share. Our non-GAAP effective tax rate was 22%. We again delivered GAAP profitability in Q3 with GAAP net income of $108 million or $0.31 per diluted share. Now turning to the balance sheet and cash flow statement. We ended Q3 with cash equivalents and investments of $6.7 billion. It is worth reminding investors that our 2023 convertible note will mature on July 1, 2023, and we expect to settle the principal obligation with cash on our balance sheet of $1.7 billion. The excess will be settled in shares. These shares have previously been accounted for in our non-GAAP diluted shares outstanding. Q3 cash flow from operations was $432 million with total adjusted free cash flow of $401 million this quarter. Stock-based compensation declined by 90 basis points as a percentage of revenue sequentially, and on a year-over-year basis, stock-based compensation was down 220 basis points as a percentage of revenue. As we look forward, we remain focused on profitable growth. At our Analyst Day in 2021, we outlined plans to drive 50 basis points to 100 basis points of margin expansion annually in fiscal year 2023 and fiscal year 2024. In the months leading up to this profitability commitment, we focused in-depth on optimally balancing investments in our business and opportunities to capture efficiencies and benefit from our growing scale. As a result, we came out of this effort with significant conviction in meaningful operating leverage. In fiscal '22, we started influencing these plans but faced supply chain challenges that unexpectedly drove higher costs. While the supply chain was uncertain as we entered fiscal year 2023, we also saw signs of the changing macroeconomic environment. As such, it was the right time to accelerate our efficiency plans. We focused our headcount additions in sales and R&D to fuel our medium-term growth prospects. Outside of these critical investment areas, we've leveraged our scale and employed technology to accommodate our growth in other business areas. Additionally, supply chain challenges have continued to abate at an increasing pace, helping to improve our gross margins. The result has been a significant acceleration in operating margin expansion through the first three quarters of fiscal year 2023 and also increases to our operating and free cash flow margin guidance throughout the year. As you see with our guidance for non-GAAP operating margin in fiscal year 2023, we were nearly 300 basis points ahead of the midpoint of our fiscal year 2024 range, that we implied back in 2021. We now see our fiscal year 2023 non-GAAP operating margins as a baseline to build on in the future. Moving on to guidance. For the fourth fiscal quarter 2023, we expect billings to be in the range of $3.15 billion to $3.20 billion, an increase of 17% to 19%. We expect revenue to be in the range of $1.937 billion to $1.967 billion, an increase of 25% to 27%. We expect non-GAAP EPS to be in the range of $126 to $130, an increase of 58% to 63%. For fiscal year 2023, we expect billings to be in the range of $9.18 billion to $9.23 billion, an increase of 23% to 24%. We expect NGS ARR to be in the range of $2.80 billion to $2.85 billion, an increase of 48% to 51%. We expect revenue to be in the range of $6.88 billion to $6.91 billion, an increase of 25% to 26%. We expect product revenue growth in the range of 15% to 16% in fiscal year '23 as we see supply chain challenges normalize as we exit fiscal year '22. In fiscal year '23, we expect operating margins to be in the range of 23% to 23.25%. We expect non-GAAP EPS to be in the range of $4.24 to $4.29, an increase of 69% to 70%. We expect our adjusted free cash flow margins to be 37.5% to 38.5%, and we expect to be GAAP profitable for fiscal year 2023, including in Q4. Additionally, please consider the following modeling points: We expect our non-GAAP tax rate to remain at 22% for Q4 '23 and fiscal year '23, subject to the outcome of future tax legislation. For Q4 '23, we expect net interest and other income of $50 million to $55 million. We expect Q4 diluted shares outstanding of $326 million to $332 million. We expect fiscal year diluted shares outstanding of $322 million to $324 million, and we expect Q4 capital expenditures of $35 million to $40 million. With that, I will turn the call back over to Walter for the Q&A portion of the call.

To allow for broad participation, I would ask that each person ask only one question. Our first question will come from Saket Kalia of Barclays with Hamza Fodderwala from Morgan Stanley on deck. Saket, you're muted. All right. Why don't we go to Hamza?

Okay. Can you hear me now?

Go ahead.

Sorry, I didn't unmute. Thanks so much for taking the question here and a nice job to the team executing in a very challenging environment. Nikesh, maybe a lot of good things to talk about, but I'd love to just double-click on the operating margin improvement here that you've seen and really a new baseline that the team is creating going into next year. Maybe the question is, can you and Dipak maybe talk about what areas the team is finding efficiency and what are the opportunities for efficiency maybe going forward as well? Thanks.

Yeah. Look, I’ll preface that as Dipak highlighted, the supply chain crisis is all but over and there were some adverse impacts to gross margins driven by hardware. I think the product mix is in our favor. As we go from hardware to software, our gross margins are way better than software than they generally are on hardware given the software firewalls are much, much more profitable for us. Coupled with that, I think what Dipak really has been driving for the last year as we flipped into the new macroeconomic environment has been a real focus on resource utilization, ROI as well as making sure we are focused on hiring only on stuff where it's important. He also talked about streamlining the sales force. If you remember, Saket, we have the conversation around making sure our SASE team has integrated with our core, which saved us hundreds of heads in terms of efficiency as well as driving more outcome and output from a SASE perspective. So generally, those have been some of the key drivers but, Dipak, did you want to add something?

No, I think you covered it all. I think Saket, we've talked this before on cloud. We scale well as a company, right? And I think that's across all the different elements of our P&L. I think Nikesh has talked about the supply chain, he talked about the OpEx, I'll just also mention cloud-hosting and cloud consumption as we get bigger and we can see more, we have the ability to go back to our service providers and negotiate better contracts. So I think across all the areas of the P&L, we scale pretty well as a company.

And I think to your question in terms of where this goes, as Dipak said, this is a new baseline. We think there is continued opportunity from here and we haven't even factored in the potential impact of generative AI. As you've been hearing all the conversation in the industry, we're still working on it, we're understanding it, we're really looking at processes, but we believe there is a there, there. We think there will be an opportunity in the future to get more efficiency from generative AI as we go ahead and implement some of the capabilities through our organization. So I think there is upside both in the continued efforts of what Dipak has been driving for the last nine months and there is the icing on the top, which is the potential application of generative AI as we continue to grow the business over the next few years.

Got it.

Thanks, Saket.

Well done. Thank you. The next question is from Hamza Fodderwala from Morgan Stanley with Brian Essex from JP Morgan on deck. Hamza, go ahead.

Hey, guys. Good evening. I hope you can hear me okay? Maybe a question for Nikesh and Lee Klarich if he is around. Nikesh, on AI you've clearly been thinking about this a lot based on what I can tell from your Twitter. But we were at RSA last month, and while there's a lot of opportunity around AI, there seem to be a lot of risks around data security, around sort of the data that these models are trained on. So I'm curious as you have the AI-based conversations with your customers, how are you getting them comfortable around that to really leverage the full capabilities of AI to automate their SOCs?

I believe there are two aspects to consider. One aspect is how we are already integrating AI into our products, which we have been doing for some time to analyze patterns and provide real-time data insights. We utilize over 1,000 AI models to examine what has occurred in XSIAM, all of which is proprietary to us. In this case, we are not using a general large language model; rather, we have developed a proprietary AI model specifically for Palo Alto Networks to address security needs. Additionally, as we plan to incorporate conversational AI into our models, we are collaborating with various public and open-source models to discover how we can create solutions that leverage our proprietary data. Lee, do you have anything to add on that?

Yes, it's still early in the adoption of large language models, and there are several risks associated with them, especially in enterprise scenarios. We've observed instances where data has been included in large language models without proper understanding of its usage, leading to confidential information becoming publicly available. This highlights the need for caution. Additionally, there are security concerns such as prompt injection attacks and data poisoning that must be addressed. Therefore, the enterprise applications of LLMs need to develop more carefully and methodically, considering these security challenges. However, it's also crucial to acknowledge the significant potential they offer. As mentioned earlier, they can assist in guiding product adoption and usage, enhancing security capabilities, and improving overall business efficiency.

Yeah. I think to cap it off, I think there is no doubt we will continue to deploy our proprietary AI models for XSIAM or for our network security use case as I highlighted. We believe in our preliminary analysis over the last three months and driving a lot of these work streams internally that there is a dare there with generative AI. So we believe that we will be deploying generative AI over the course of the next few months, and we'll talk more about it at a later event. But we think that has an opportunity both to significantly improve our customer efficiency and the efficacy of our products, at the same time also to drive efficiencies within the way we run Palo Alto Networks. I think last but not the least, which is something you didn't ask, but I'll say, separately, Lee and his team have been working hard to see and look at the adverse impact that generative AI could have in terms of adversaries using generative AI to build new malware, to try and attack our customers. And there's a lot of work we're doing as well to make sure we are able to protect our customers against any such activity that is conducted using generative AI.

Thank you.

Thanks for your question, Hamza. Next question is from Brian Essex at JPMorgan, followed by Brad Zelnick from Deutsche Bank. Brian, go ahead.

Good afternoon, and thank you for the question. Following up on Saket’s comments, it's encouraging to see progress in operating margin and an increase in cash flow margin guidance. Could you elaborate on the key drivers behind the improvement in cash flow margin and its sustainability? We've noticed that CapEx seems to be lower than your previous guidance, so we're curious about how sustainable this metric might be as we consider it for valuation. As we project operating margins in the future, should we expect the gap between operating margins and cash flow margins to remain fairly consistent?

Yeah. So Brian, thanks for the question. Let me just start off with like the biggest driver over the long term is really just to strength in your bookings. At least your billings and then comes down. Then the foundation really is your operating margins that then makes up the base that you can do on your cash. There are multiple other factors, but do recognize that when we came into the year, the interest rates were at a different level. We have had the benefit of higher interest rates. We've deployed a lot of our cash that we earn interest income. We're not predictors of interest rates, but fundamentally, we believe that, that will continue to be a tailwind for our cash generation. And then last but not least, we do have PanFS. We have a certain amount of our business that we do structural and financing. Frankly, that's been broadly in line with what we assumed at the beginning of the year, but those are really the drivers, and we feel pretty comfortable on what we're able to do with those different drivers and delivering on our numbers?

Great. Thank you.

Great. Thanks, Brian. Next question from Brad Zelnick at Deutsche Bank, followed by Andrew Nowinski at Wells Fargo. Go ahead, Brad.

Great. Thanks so much for the question and nice job, both to Nikesh, Dipak, and the entire team. Nikesh, my question is about M&A, which I feel like typically comes later in the call, but I feel like it's such a great opportunity right now. What's the hurdle to doing a large deal and can you remind us how you think about transformative M&A? And just related to that, your competitors naturally knock you on having grown through required innovation. Just to set the record straight, can you talk about how much of a priority and a focus it is to have a deeply integrated product?

Yeah, Brian. I think, first of all, I'm amused that you're asking for transformational M&A. I think I feel like somehow we at Palo Alto Networks have been going through a transformation already for the last five years. Let me talk about it in two different parts. One, and I'd like to bust a myth of the notion that we've grown our innovation through M&A because pretty much the entire XSIAM product that we've built, which is now going to be one of the fastest platforms of Palo Alto Networks is homegrown. It was built by our team internally. It was designed, built, and delivered by the Cortex team. So I think it's a disservice to them to say that some of the fastest-growing platforms being built at Palo Alto Networks was acquired. Similarly, our next-generation firewalls or our SASE product for the most part is entirely homegrown, driven by the security capabilities that we built using our firewalls as well as our virtual firewall business. So I think the majority of our M&A has been focused on building our cloud security portfolio where we felt where we needed to be assertive and be out there in the front. And I would say auxiliary capabilities, whether it's in automation with XSOAR or auxiliary capabilities around attack purpose management. So bottom line, we're very comfortable with the three platforms that we have and what we need to get done. I think we've been very clear about from an acquisition perspective, we look for product capability, where we can take product capability and attach that and make sure we can solve more problems for our customers that they're looking at. So from that perspective, my view on M&A is consistent that we find something interesting, an industry trend, which is added incremental tech capability, we will do it. I think from a transformational M&A, I think we can transform this company and have continued to transform it to where it is based on our innovation and our balance of execution. I think we will continue to do that. I don't think the market is particularly cheap yet. If you were to try and look for transformation M&A, and I think it's kind of a dual double-edge situation. One, I think we continue to get stronger as we get execution under our belt, and we continue to grow in value as Palo Alto Networks. And if some of the large players out there end up committing missteps, then we'll go take a look at it for now. I feel very comfortable with the position Palo Alto has in the industry. I feel very, very comfortable with the amount of cash we have on our balance sheet. And I believe it is our job to keep our heads down and keep executing because it's a tough market. And I think one of the things that was brought up just a minute ago, I think the opportunities from AI have not been fully comprehended by most enterprise businesses. I think we are going to undergo a transformation both at Palo Alto Networks as well as generally in the enterprise software industry over the next 12 months to 24 months as we embrace generative AI. I think that's the real opportunity and challenge in front of us, and I think half of the people out there will get it wrong. And hopefully, we're on the right side of history.

You're doing a great job; keep it up. Thank you, Nikesh.

Thanks for the question, Brad. The next question is from Andy Nowinski from Wells Fargo, followed by Matt Hedberg from RBC. Andy, go ahead.

Okay. Thank you. And congrats on a great quarter. So nearly every single vendor and nearly every single reseller we talked to says they're seeing an elongation of sales cycles, yet you seem to defy those headwinds with massive growth in large deals and customer spending $5 million and $10 million with you. I guess would you view this as an important inflection point as it relates to sort of consolidation in that if you can drive large deals in this macro constrained environment, you could potentially see an acceleration of those consolidation trends when the macro improves?

Are you predicting a macro improvement, Andy?

I certainly hope so.

Well, look, I think first and foremost, I don't want to leave you with any impression that the macro is not hard. It is hard out there. I think everything you're hearing from resellers, from other people in the industry is true. Customers are spending more time paying attention to deals. Customers are taking longer; some are rightsizing deals; some are focusing on things that are important. Some are looking for financing; some want to pay annually. So all the effects that you talked about are true in the industry. And we recognize this towards the end of our first quarter. And I'll tell you what, we've been working at double time, like literally, the day Dipak shut the doors and us being able to book anything this quarter, we are out there hunting for next quarter. We have a big number to hit this quarter. We're out there in the field. We're executing; our teams are out there. So as you probably appreciate, there is no magic in the world around the fact that our quarter ended July 31. There's no budget year-end for any part of the world on July 31. It's a date that's been created at Palo Alto finishes the year Q4 July 31, which means we have to run as hard as we can to get business done by July 31. We know that at the end of our year; we know that we see end of our quarter; our customers know that. So what we're doing is we're getting ahead of it. We're hoping that us getting ahead of it and continuing to rigorously execute is going to allow us to be able to improve our conversion rate. Our conversion rates on our pipeline are down; guess what? You dug up more pipeline, therefore, your conversion rate that's down still allows you to make the number that you promised the Street. That's what we've been trying to do. And as I've said, the macro is hard, and we're going to keep trying to keep our heads down and execute.

Thanks, Nikesh. Keep up the good work.

Great. Thanks, Andy. The next question is from Matt Hedberg at RBC followed by Gabriela Borges at Goldman. Go ahead, Matt.

Thanks, Walter. Nick, congrats again to the team, outstanding results. I guess, Nikesh or Lee, on the success you've seen thus far with XSIAM, you noted you essentially have full access to SIM budgets right now. I'm curious with some of the large deals you're seeing, are these generally replacing legacy SIM vendors? Or are you actually generating new TAM that didn't exist previously?

So Matt, I'll let Lee jump in and talk about some of the specifics, but I'll tell you what, every one of these deals is a replacement of a legacy SIM or a data store. In addition, we do not sell XSIAM without our endpoint products. So you have to buy Palo Alto Cortex XDR to deploy XSIAM because we believe the only way to have normalized good source — single source of truth data is to deploy our endpoint products. And then we use that, as I showed in the AI funnel of how we can go cross-correlate that and drive great security outcomes. So in every case, we are replacing an existing vendor. But I will tell you, the SOC industry is upside down. It was designed so far to go understand when a breach happens, how the breach happened, and trying to figure out how to remediate it. And those remediation times, as I highlighted, are six days, and now most modern attacks are in and out in under 12 hours. So if you've got a SOC infrastructure where it allows you to come up with what happened to you after six days, the bad actors have gone in and out in 12 hours; you have a mismatch. That is a problem. But Lee, can you highlight some of the key use cases that we've seen in the first 30 plus customers that we have; what's driven some of this transformation?

XSIAM is not only replacing traditional SIM but also other tools within the Security Operations Center. This transformation occurs through three main components. Firstly, the platform processes 3.5 petabytes of data daily, which is crucial for effective AI, as XSIAM is built to handle vast amounts of data from various sources into an AI data lake. Secondly, we utilize AI-driven analytics to identify attacks in real-time, a capability that traditional SIMs were not equipped to provide, which enhances our detection times. Lastly, XSIAM integrates automation to significantly reduce remediation times from several days to mere hours or even minutes. The results we are witnessing from our XSIAM deployments are impressive and mirror the success we experienced in our own Security Operations Center after implementing an operationalized version of XSIAM.

I think the last — sorry, Matt, the only thing I'll add on this is that over the last 15 years, what has happened is the cost and value equation in existing SOCs has diverged tremendously. So people are spending a lot of money collecting data in large data stores, and they're not getting adequate value out of it, and they're not getting adequate security outcomes out of it. So I think that is a big gap, and that gap is something we've been — we've built this product to try and fill — and now it really is very early days for us. I think the fact that we'll get to $100 million in the time spent that you thought was aggressive less than that. I think tells us there's a huge potential out there, which means we have to keep our heads down, again, keep building, keep executing, and keep trying to solve the problems that our customers are presenting in front of us, but I have a good feeling about it.

Certainly seems that way. Thanks.

Thanks, Matt. Our next question from Gabriela Borges at Goldman Sachs with Adam Tindle from Raymond James on deck. Gabriela, you go ahead.

Good afternoon. Thank you. Either for Lee or Nikesh. I wanted to ask about your cloud security strategy in Prisma, specifically with respect to how you think about the right balance of incentives that you give customers upfront to catalyze adoption? And then also how you think about the balance of top-down growth versus product-led growth, given that DevSecOps, DevOps, some of those tools seem to be driven by product line growth as well? Thank you.

Yeah. Lee, go ahead and answer that question.

One of the challenges we aimed to tackle with Prisma Cloud is the issue of numerous point products in enterprise cybersecurity. Whenever there's a new security requirement, a new product is introduced, and customers end up acting as system integrators for all these different solutions. This often leads to them spending more time managing these integrations than actually gaining value from the products. With Prisma Cloud, we've taken a different approach by creating a platform that offers many integrated capabilities from a single source. Simultaneously, we've simplified the adoption process through a single credit system for Prisma Cloud, making it easier for customers to purchase a certain capacity and utilize the platform as needed. This strategy allows us to concentrate more on customer engagement and product adoption, guiding users to leverage additional features as required, rather than turning each module into a separate transaction. As highlighted by Nikesh, we've seen a 44% increase in year-over-year credit usage, and the number of customers utilizing multiple modules, particularly those using four or more, has nearly doubled year-over-year, demonstrating the effectiveness of our approach.

Great. Thanks, Gabriela. Next up, Adam Tindle, Raymond James; followed by Gregg Moskowitz, Mizuho. Adam, go ahead.

Okay. Thanks. Good afternoon. I want to start by just acknowledging the progression in operating margin is really impressive, and commitment to that being a baseline is a really important point. If I'm thinking about tomorrow, some of the distracting questions that might come up would be around product revenue. I think you grew 10% year-over-year in Q3, and you had previously guided the fiscal year to 10%. But if I saw in the slides correctly, I think you're now raising that to 15% to 16%. So what's driving that increase in product revenue and the acceleration in Q4 despite the cautionary comments? And anything we can think about in terms of puts and takes to product revenue as we think about fiscal '24, so we don't get ahead of ourselves? Thanks.

Yeah, Adam. I think there are two parts to it. One is, as you will appreciate, we highlighted that software has become 30% of our product revenue. So we — when you book a hardware firewall, you get a dollar for dollar for revenue. In software, you don't get a dollar for dollar for revenue; there is some part of an amortized value we get from our software firewalls and some part of our SD-WAN, which becomes part of our product revenue. So we have to run harder on billings to be able to deliver product revenue in the context of software. But as I mentioned, our virtual firewalls grew at 55% this quarter. They grew at 40% for the year so far. This is a tailwind we had not expected. At the same time, the hardware, as I mentioned, is not as strong as we’d expected. So they balance each other out. But in balance, it is in favor of software for now, coming off a low base of last year. So as a result, we have been able to improve our product revenue guidance. So obviously, it comes at the cost of services revenue because some of our software has now had to work triple time to be able to deliver product revenue. So I think that's the context in which you should think about it overall, where there's been a draw from one side and a partial give on the other side on the product revenue. However, given our RPO is growing way ahead of revenue, it just means we are saving up a lot of revenue for a future rainy day.

No, in a difficult situation. I would like to add that the supply chain dynamics discussed by Nikesh do play a role, but we have effectively managed to stay ahead of the supply chain challenges with our outstanding team. This may help clarify some of the fluctuations you are observing.

Great. Thank you, Adam. Next up, Gregg Moskowitz from Mizuho, followed by Shaul Eyal from Cowen.

Thank you. Can you hear me?

Yes.

All right. I have a follow-up for Lee or Nikesh on generative AI. So your comments on LLM were helpful, but do you think generative AI will tilt the scales in favor of Palo Alto and perhaps some other security vendors over time? Or is it ultimately more likely to cause an even faster game of cat and mouse between the vendors and the attackers? How do you see this playing out?

Well, I think, look, first and foremost, the benefit of generative AI so far is twofold, right? One is in its ability to summarize data and give you access to information much faster. Can I imagine a sales rep at Palo Alto having access to their fingertips about all Palo Alto information? Of course, I can. Can I imagine my customer support people having access to amazing amounts of information that's at the tip of their fingers so they can answer customer questions much faster? Can I imagine for showcasing that information directly to my customers as you're seeing the industry now suddenly a plethora of copilots start to emerge in every product. So I think that is going to become an obvious benefit of generative AI. Now don't forget, it relies on one principle called having a lot of data. But it's very important that whether you're using it for sharing your own information from your customers to your customers, you need a lot of that data. You have to clean all your data processes and have that. Secondly, if you're in the security business, it definitely helps if you have the largest data lake in the world of security data. So from that perspective, I think it favors the people who have a lot of data already as part of their strategy, and they have built a business on the back of a data-led strategy. I think not just specific to security but in any industry, especially consumer Internet, if you've been a UI company, you have something to worry about. If you're a travel booking operator or something that just takes other people's data and makes a better UI, you have something to worry about. So I think from that perspective, it favors companies which have tremendous amounts of data. I think the second thing is also important to understand, if I have 14,000 people, I spend thousands of billion dollars on customer support or more; there is leverage. I can go spend $30 million, $40 million, $50 million deploying an LLM and saving up my cost. If you're running a small company and your entire cost is $50 million, it probably doesn't behoove you to go out and create a LLM-based generative AI project to go out and pay and take away $12 million of cost. So I think it also benefits people of scale who are able to drive efficiencies using generative AI across the enterprise, allowing them to grow their business much faster with limited resources. Does that help?

It does. Thanks, Nikesh.

Great. Thanks, Gregg. And Shaul Eyal from Cowen, our last question.

Good afternoon. Congrats, team. Nikesh, I want to go back; actually, I know Brad was asking about M&A. I want to ask about the competitive landscape, but specifically with a focus maybe on the CNAPP front. So my question is, how do you think about it? Any change? Do you think that the product right now, as it stands, is comprehensive or anything you might be thinking of maybe augmenting specifically on the CNAPP front? Thank you for that.

That's by far the most comprehensive cloud-native application protection platform there is. That doesn't mean that we do everything, but we do far more than any other solution out there. There's a tremendous amount of focus on delivering capabilities that we've been building internally, organically amongst the team. We've seen the most recent one we delivered with secret scanning just a few months ago. We've seen very good early adoption of that. At the same time, we're also delivering on the latest acquisition of cyber security, where we expect that to become a new module in the next couple of months available to all of our Prisma Cloud customers. And so the — Nikesh talked about how we've leveraged M&A in the past to help build some of the key technology areas of Prisma Cloud, which is absolutely true. We have also shown an ability to deliver new cloud security capabilities organically and be very successful at that. And right now, I feel good about the balance of both those capabilities and how we're bringing them together and how we continue to deliver new innovations.

Thank you for the question. With that, we'll conclude the Q&A portion of the call, and I'd like to pass it back to Nikesh for his closing remarks.

Well, thank you very much again, everybody, for joining us. We look forward to seeing many of you at the upcoming investor events. I also want to once again take an opportunity to thank all of our employees who worked very hard in a very dedicated fashion, as you all know, to help us achieve the results. Not only that, a big thank you to all of our partners and our customers around the world. Have a wonderful day. Thank you.