Palo Alto Networks (TICKER: PANW) Palo Alto Networks Inc Panw Ceo Nikesh Arora Presents Barclays Global Technology Media And

Palo Alto Networks, Inc. (NASDAQ:PANW) Barclays Global Technology, Media and Telecommunications Conference Call December 7, 2022 3:45 PM ET Company Participants Nikesh Arora - Chairman and CEO Conference Call Participants Saket Kalia - Barclays Saket Kalia Excellent. Well, hey, good afternoon, everyone, and welcome to Day 1 of the Barclays TMT Conference. It's an honor to have with us the team from Palo Alto. Of course, we've got Nikesh Arora, Chief Executive Officer. We also have Walter Pritchard, Head of IR, around here somewhere. So, we've got about 30 minutes together. Maybe I'll take the first 20 or 25 minutes and do some fireside chat here with the Nikesh, which I know is going to be fun. And then, maybe in the last five or ten minutes, let's make this interactive. So, if anyone has a question, I think we've got some mic runners around here. Just pop up your hand, and we'll make sure we get your question in. So, with all of that as a framework, Nikesh, thanks so much for being with us here today. Nikesh Arora Thank you for having me. Question-and-Answer Session Q - Saket Kalia No, absolutely. Maybe just to start off, make sure we're all on the same page, a lot to recap from last quarter. But maybe to level set for all of us, what were some of the things that you were particularly proud of in Q1 to help maybe frame the discussion? Nikesh Arora Well, Saket, as you can see, we've had a flurry of reporting in the security industry. I think the key highlights of Q1 was there have been reports from many of the security companies dipping on the margin, slightly off from what you might have expected. But across the board, as an industry sector, it still is more resilient than most subsectors of technology. That was kind of obvious in Q1. Business is not going away, at least that's not what anyone is telling you, that's not what we're seeing. You're seeing as per people ramp deals or business deals elongating, which is something you contend with. All that means is things are getting more scrutiny, because Jay Powell is being hammered. Everybody is worried that he's going to be successful, in which case, there will be some slowdown of economic growth and everybody is kind of preparing for it, be it us or our customers. And you're seeing the impact of that in everybody's narrative. I don't think we're seeing as much of it just yet, but everybody is apprehensive, and hence, expecting that's going to happen. Now, from our perspective, I was excited that our teams are still able to nail the number. They're still able to go out and get it. We saw it a bit earlier in the quarter and we decided to accelerate our go-to-market efforts and get people out in the field and on the payment much faster. We went ahead and accelerated hiring on key sales roles, because we expect that we need more feet on the street, we need people fully trained out there, because we think if this trend continues over the next six to nine months, which means you can contend -- contended with sort of more activity upfront. So, generally, across the board, there are some increasing signs of customer attention to budgets. We can hopefully execute our way through it. As I have said before, there was a silver lining, the consolidation conversations, we-can-save-you-money conversations by having lesser vendors and lower total cost of ownership, those things are now beginning to work in our favor. Saket Kalia Yes, absolutely. And certainly, what I would add to that last quarter, and I'm sure we're going to talk about later on is while nailing the number, we were still able to raise the margin guide for the year despite everything else. Nikesh Arora Saket, there's a time and place for everything. You've been asking me and people have been asking me what about your margins, and I always maintain that you've got to balance growth and profitability. Growth was our key focus because we were trying to build new product categories. We're trying to make sure we establish our capability across various categories. We feel comfortable that we have repeatable go-to-market motions in our cloud security business, in our network security business, and our cortex business. As we told you at the end of Q4, we are converging our network security teams, trying to extract more efficiency out of those teams. And we just put all that into play. We're beginning to see the benefits. I think in long term, if you want to grow north of 20% in our business, you should be able to sustain north of 20% operating margins. Saket Kalia I think, that's fair. I think one of the benefits that Palo Alto Networks has is the broadest portfolio in security. Maybe notwithstanding some of the macro backdrop that we mentioned, the question is, how is that broad portfolio helping in closing business? And maybe where I'm going there is, you've talked about building this muscle internally for selling the broad platform. Where do you think the company is in that evolution of platform selling? Nikesh Arora So, the first two, three years at Palo Alto, I had to spend 70% of my time with the product teams, making sure we have this extensible platform. And it's kind of clear that there are approximately three buying centers. There's a whole network security team, which is focused on Zero Trust, which is dealing with the impact of the cloud, how do I take my data center infrastructure, figure out a hybrid infrastructure between the data center and the cloud, and now people working from home. So, they're all busy being network transformation projects and we have the ability to come in and say, "I can do Zero Trust for you." And I use this explanation the other day when somebody asked me a question if firewall is going to be dead. And I was particularly delighted with the way we talked about this. So, if all of you who are worried about firewalls and network security, what firewalls are, they are a mechanism to inspect traffic, okay? That's what they do. They look at traffic to see if there's malware, bad URLs, bad DNS, they're just inspecting traffic, making sure traffic flows without security flaws. Traffic is exploding in a big way, whether it's traffic from the campus to the data center, from the data center to the cloud. So, generally, traffic is going up. That traffic must be inspected. You can inspect it with the hardware firewall or software firewall or a SASE product. Now, we're one of the very few people in the industry who actually have form factors in all three categories that work together. So, our network security people are working hard towards this platform notion of providing that capability. Cloud? Same thing. We have seven different modules we can go from code to deploy to run, so we, again, have a platform in cloud security, and that's the conversation that's beginning to happen and continues to happen. And then, with our most recent launch of XSIAM, we're finally putting our stamp in the idea that SOCs need to be reinvented, they need to be automated, and we have very good early signs that this is going to be an emerging big category. Saket Kalia Absolutely. A lot of fun stuff there that I want to dig into. But maybe we can just dig into the three main parts of the portfolio, maybe starting with Prisma Cloud. Nikesh Arora Sure. Saket Kalia I think Palo has been investing more in software supply chain security, most recently with the proposed acquisition of Cider, and of course, it started with Bridgecrew. I guess, with this now under the Prisma umbrella, can you just talk about how you view the software supply chain opportunity? And how Palo could maybe disrupt us with the portfolio that you're assembling? Nikesh Arora I think it's important for everybody here to understand what the challenge or the opportunity is, and I'm presuming most of you have not done coding in today's environment where -- and if you have ever coded in the past, in the past, when you coded, you wrote an application, there are application testing software, whether it's Checkmarx or Veracode, which are 15 years old, they checked your code, if they said everything is cool, you ship it to production and you're off to the races. And you wrote this stuff in a -- I remember programming in COBOL and FORTRAN. I'm sure there's a bunch of stuff that happened after that. But today's cloud world is very different. On average, every enterprise has about a 100 tools that they use, which are third-party tools, which are grabbed from open source, they grab widgets and images, and it's actually an assembly exercise. Software is assembled more today than it's coded. That's why it's got a supply chain problem, because you're assembling software together using a bunch of tools which could have a bunch of security flaws. So, in cloud, the point of insertion of malware, the point of insertion of security flaws is in the coding process. If you don't fix the security in the coding process, you will forever be playing whack-a-mole in deployment or production. So, you got to get ahead of the problem. The way you get ahead of the problem is you scan every piece of code. You scan every image. You look at every tool that's being used from a third-party perspective, that's where supply chain security comes in. We brought Bridgecrew that was doing all the scanning. We brought Cider, which is going to do all the CI/CD work and the supply chain stuff. So, we've got the left. Now, what we can do is we can investigate the left and then track you through the entire development process to see where the problem came from, so you can go fix it. There's nobody else in the industry. We think that generally, the industry is about two to two-and-a-half years behind where we are in terms of our thinking, our product capability and what we're working on. Saket Kalia Absolutely. I think you've talked about Prisma Cloud having more of a multi-product opportunity, I think with nine modules. And clearly, we've talked about software supply chain is one module. And of course, I think the cloud workload protection is also a popular module. Maybe the question for you, Nikesh, is what modules here do you think have the most room for adoption as you think about Prisma Cloud becoming a $1 billion business someday? Nikesh Arora Well, Prisma Cloud is going to be $1 billion business in the next 12 to 18 months, so not someday, that's pretty much in the near future. So -- and it depends what $1 billion business means. We'll do $1 billion in bookings in the next 12 to 18 months in Prisma Cloud. Saket Kalia Got it. Nikesh Arora So, we've got a worry about the $5 billion problem, not $1 billion problem, right? Saket Kalia Got it. Nikesh Arora I think about it differently, Saket. The way I think about it is, in our lifetimes, if you look at evolution in the next five to eight years, there'll be $1 trillion of public cloud being consumed every year, if you believe that. Now, I'll give you a data point. Palo Alto Network spends approximately $250 million a year in public cloud. And if I had to pay for my own cloud securities to my own products, I'd be paying somewhere between $10 million to $15 million, right? That's 5% to 7.5% of my cloud spend. That's how I benchmark the TAM for cloud security. So, if there's $1 trillion out there, there's got to be a $50 billion to $75 billion spend on cloud security in the future. If you believe that's the size of price, the question is, what can I honestly aspire to. I'm saying half of that will be taken by the public cloud providers. They will just take it away as part of the ELA business. There's still half up for grabs. So, there's $25 billion to $37.5 billion opportunity in cloud security in the next ten years. The current market is $2 billion. So, there's a 17x on the high end and a 10x on the low-end opportunity in ten years, right? So, I think people have to consume cloud security in its entirety. It's not one particular module that needs to be consumed. You have to secure the entire development process, deployment process and production process. So, it's an arms race. You got to rush fast to make sure you have the capability and to make sure your customers are deploying you in their infrastructure. Can I aspire to 25% market share? I got 29% market share in firewalls, and we were the last player in firewall to show up and the first player to show up in cloud security. That's how I think about it. Saket Kalia Interesting. I mean, maybe said another way, in that cloud security market, still a lot of greenfield opportunity, just a lot of... Nikesh Arora There's lots of opportunity. You just have to go a grind it out. There are not enough perfect go-to-market motions being built. There's not enough solutions architects out there. There's not enough cloud consumption capability out there. So, there's a lot of work to be done, but the TAM is huge. Saket Kalia Right. I'd love to maybe shift to Cortex a little bit. And there's a lot to talk about there, I mean, with Xpanse, with XDR, with XSIAM, to your point. And maybe Xpanse is a good place to start, particularly with the large U.S. federal deal that we announced last quarter. Congrats on that, by the way. I mean, maybe the question is, what's the breadth of that opportunity with Xpanse? And are there other potential deals out there like that federal deal that -- you remind me of the size, right, but those are... Nikesh Arora It's $125 million deal, of which you recognized $67 million last quarter. Look, again, I'm sorry to keep twisting your question slightly differently. Saket Kalia No, please. Nikesh Arora This past weekend, I spent more hours than I needed to on GPT-3. I don't know how many of you have been enthralled by this conversation on GPT-3, it seems to be the talk of town. I was in a Board meeting in a private company before this, and the whole conversation was how is GPT-3 going to upend your business model. And now GPT-4 is about to show up. And the reason I bring that up is 4.5 years ago, when I started at Palo Alto Networks, I showed up with two words, because I didn't know security. I had to show up some value-added. My value added was cloud and AI. So, cloud, we know the impact. We built cloud security business. We reoriented our network security business to build a SASE product, which we're one of two players in the market. And we haven't seen our AI capability yet. Our entire AI capability is centered around everything we're doing in Cortex. And we took a point of view that saying that security is structured. To do post-breach analysis, security needs to become real time. You can't do anything real time unless you have good data. So, we spent the last 4.5 years figuring out how to get good data, which is why we launched the products 12 weeks ago called Cortex XSIAM. The whole notion there is you collect all the data you can, you run AI models against it, and try and remediate security. Xpanse is one aspect of it. It is the aspect which looks at data which you can see from the outside in. And Cortex XSIAM looks at it from inside out. Question is, can you marry outside in, inside out? Outside in would say, a hacker looks at it and say, "My God, I see seven windows open. I can go in." And look from inside out saying, "I've locked all doors and windows. So, wait a bit, what about the seven that you can see from the outside?" So, it's a way to marry the data from the outside in, inside out. So, I think, yes, AI is going to have a huge impact on how real-time security is deployed. I think everybody's got awareness as of last weekend that, yes, it's going to be real. It's going to happen. And I think [indiscernible] we think the next big market that's going to be created. So, remember, in the last 4.5 years, we've created three businesses net new which are going to be hitting the $1 billion mark in the next 12 to 18 months; there's Cortex, there's Prisma Cloud, there's SASE. We think this whole notion of XSIAM will be our fourth business that gets there in a similar timeframe that we were able to get the first two businesses there. Saket Kalia So, great segue into XSIAM. And frankly, the way that you positioned it just an inside looking out, too, makes a ton of sense. Maybe said another way, it almost feels like a little bit of a cloud-based SIEM replacement to me. Maybe the question for you... Nikesh Arora I think SIEMs are old tech. They're sitting ducks waiting to be replaced. Saket Kalia So, I mean, maybe we can dig into that a little bit. Like how can Palo disrupt that? What I would argue is a relatively competitive SIEM market and also... Nikesh Arora It's not competitive. It's like you're selling chariots and I'm selling cars. We do the same thing, right? Horses shit, and cars don't. Sorry. I don't know why I said that. Saket Kalia One of the interesting things, though, that you mentioned about XSIAM last -- on the last quarterly call was that you kind of -- it felt like you wanted to balance demand with capacity there, right? Like -- or just want to be careful with how that scales. Can you just talk to us about the idea behind that? Nikesh Arora Traditionally, the way security resolution happens is you deploy a lot of security vendors and infrastructure, you set a bunch of policies. Any time a policy is violated, you get an alert and there's people who stir at all these alerts and go and figure out what the alert was. If the other was caused by a bad actor, then you go out and remediate, you block it, if you haven't blocked it by automated needs. Remember, if firewall didn't block a bad DNS or bad URL, you have something else happen, it flips an alert. The alert looks at it and says, "Oh my God, there's an anomalous activity." Now, we used to get 67,000 alerts a week. It's humanly impossible to interpret them, analyze them and remediate them. So, you prioritize. That's what a SIEM does. SIEM says, "I got 67,000 alerts. Let's focus on these 500. These are very important. They look like bad alerts." Well, there's 66,500 more where the first 500 came from. That 67,000 is becoming 150,000 now, because you are deploying more tech and more vendors. So, this is not a human problem. This is an automation and the AI problem. So, what I meant by -- that the SIEM industry is sitting duck is most people are trying to solve this by giving you prioritization tools and UEBA tools to say, "Here's a SIEM, collect all the data into a large data lake, there's a query language that lets you query that database and you can figure out what happened." Well, if you take through 27 days to figure out what happened, the bad actors in and out and stolen your data and put it in the dark web. So, you've got to get it to as real-time as possible. It's taken us 4.5 years to get our own internal SOC from 27 days to under one minute. Now, the question is, can I deploy that for all of my customers? So, instead of going on in a big rush and saying, "Here's the product. Go deploy it," I'm saying, "I'm going to work the customers myself." We've done the first nine. We signed design partnerships. All of them have converted into customers for us in the last three months. We're going to take them down the journey of reducing their meantime remediate from 27 days to a lot shorter. And I think we can prove that model again and again, and we're going to go partner with a bunch of SIs out there who will then take our product capability and data execution capability and go take the market on. That's our plan. Saket Kalia Makes sense. Maybe we can go to Prisma SASE. I mean, clearly, a big change, I think, this year is moving this out of specialist sales and into your bigger general sales teams, maybe I'll call it. It's still early, but maybe what are some of the early signals that you're seeing? And any risks that you want to make sure you manage just around having too much to sell, maybe focusing on one tool versus the other? How are you thinking about that? Nikesh Arora Look, as I said, I spent the first two years doing product. Now, the next two, three years, I'm going to spend on making a go-to-market more efficient and deploying to every customer. At the end of the day, we used to sell firewalls, right? It's our primary business 4.5 years ago. Our core sales team out there is the primary -- we used to be a firewall sales team. We supplemented that with cloud security and Cortex and SASE. But as we've converged our SASE capability into a larger network security capability, which is Zero Trust focused, the functionality is 80% the same; what a firewall does and what a SASE product does. And we've combined the management pain in such a way that you can deploy SASE from us or you can deploy entire Zero Trust framework from Palo Alto, which is pretty consistent. So, our sales reps are actually not conversion. They're becoming network security, Zero Trust salespeople, and they say, what is the problem you're trying to solve? Are you trying to protect your data inspected? Great. The way to do it is through a hardware firewall or software or SASE. What hybrid architecture would you like? But don't forget we have 62,000 firewall customers out there who all want to go to SASE. So, there is a part which is a transition path from their existing Palo Alto firewalls to assess the outcome, which is a combination of what they already have and what we can give them on top. So, our salespeople need to know how all of that fits together and work. So, we're not merging two different sales teams, we're making sure they have a higher order pitch that allows them to present a Zero Trust solution and we feel comfortable enough that we've made enough product integration and marketing capability integration and training that these people can do it in a comprehensive way to all of our customer base. Now, that gives you more efficiencies even better. Saket Kalia Yes, absolutely. I'm going to shift to the Strata firewall part of the business, but maybe before I go there, any questions here from the audience? I guess, Nikesh, if we think about the Strata piece or maybe I'll just call it the product business, Palo, just like a lot of the other industry participants are seeing easing supply, but also more scrutiny, sweating appliances longer [indiscernible]. Now just to be clear, appliances are a much smaller part of Palo's business. But maybe the question is, how do you think about demand for appliances in this coming year, and maybe beyond that, to the extent, you've got a view? Nikesh Arora Look, as I said, the firewall-need functionality is not going to go away. You'll continue to have to inspect traffic. The question is, what is your use case and what is the best form factor for that inspection. If your use case is a data center, you better off putting a box, right? If you use case is a distributed set of 2,000 stores in the country, you're better using SASE, because you don't want to go send a truck every year to try and upgrade your boxes in 2,000 different stores. So, generally, software has a lower total cost of ownership, higher security, because you can update it simultaneously from a remote location. So, you want to preference software form factor. There are specific use cases for hardware form factors, which are low-latency, high-bandwidth use cases. I think generally, the firewall requirement doesn't go away. I've heard the comments perhaps somebody else made recently today, I think the industry grows at 5% to 8%. I've always maintained that. I think you get to see 5% to 8% growth. In tough economic times, you go towards the lower end. In better times, you get to 8% or 10%. I think the last two years have been confounding. So, nobody actually has a good sense of what happened. Most of the industry players did price increases. One or two of them, I think, they've gotten a bit confused that two price increases of 10% each can get you to 20% growth on a for -- on a unit per unit basis. Now some of them, we're not lucky enough, we cannot pass prices to our customers. In the enterprise space, our customers negotiate our prices. I increase price by 8%, my yield is 1%. So, I have no intent to keep increasing price and get a little yield. If you're servicing the small to mid-size businesses, you get to pass price to small to mid-size businesses, so you get a better uplift on your business. Couple that with people ordering ahead, because of the supply chain issues, couple that with some of our industry competitors not being able to supply for 12 months, you get a distorted market with differentiated growth rates, which you're not clear what the annual growth rate is. I think you get the 5% to 8% growth rate. Hardware is the easiest thing to postpone, because you already have it deployed, you can say, "So I did for one more year, buy it another year." You can't postpone SASE. You can't postpone cloud consumption. You already paid for it. You can't postpone cloud security. You can't postpone SOC automation. So, hardware is this susceptible part of the market, 5% to 8% growth, we feel very comfortable. We can deliver that -- I think, that kind of sustainable rates for the next two to three years. Saket Kalia Make sense. I'd love to wrap up with just some more specific financial questions. And maybe we can start with billings and cash. I mean, clearly, you beat your guide last quarter on billings, but we beat NGS ARR by a lot more. And so maybe the question is, are there any timing or duration points that we should keep in mind when... Nikesh Arora Are you saying, is that number real? Saket Kalia Sorry? Nikesh Arora Is that your question? Is that a real growth? Saket Kalia No. Of course, it's not... Nikesh Arora So, if I miss, you don't like it. If I do well, you don't like it. Saket Kalia No, it's more of a question around when will we see that growth come through billings? Nikesh Arora That's not -- there's no lag in financial metrics. I can't add anything to ARR unlike private companies which is not already in our billings. Saket Kalia Okay. Nikesh Arora It can only be in my ARR, if it's in my billings. Saket Kalia Okay. So, it's not -- so there's no invoicing -- there's no lag in sort of invoicing versus... Nikesh Arora I can't recognize something as potential future revenue if it's not in my RPO or it's not in my billing. Saket Kalia Okay, got it. Nikesh Arora Right, mathematically true. Saket Kalia Well, RPO for sure, but with billings maybe differently, right, in terms of invoice versus the contract? Nikesh Arora Yes. But remember, I bill for the year. And ARR has this funny word call A, which is annual. I build it this year, it's annual. What I've not built for next year is not -- it's next year anyway, it's not this year's annual. Saket Kalia Okay, got it. Understood. Nikesh Arora So, there's no gap. Saket Kalia Okay. Nikesh Arora I think your question -- I shouldn't be answering questions and asking, but I'll answer it anyway. Saket Kalia Yes, please. Nikesh Arora I think the question was, why is our NGS ARR growing so well, right? And generally, the answer is if your billings stay constant, your NGS ARR does a lot better, that means your non-NGS is now shifting to NGS. Now you could say our non-NGS is not performing or you could say our transformation is working, but you wanted to say just truth. We said we're going to make our entire business cloud enabled, and we're aggressively converting our business to cloud. So, a lot of our subs are cloud, a lot of our new products are cloud, a lot of our old business we are replacing with cloud-delivered capability, which is part of NGS. Saket Kalia Got it? Maybe just in the last couple of minutes that we've got left, again, one of the things we said at the beginning that I was really happy to see it was just the profitability, right? I mean, to me, it was clear that Palo pivoted quickly given the macro backdrop. And I think we took operating margins up by 50 bps, free cash flow margins up by 100 bps. You mentioned some front-loading of hiring. Maybe the question is, are there any other actions that you've taken or plan to take that give you confidence in driving the higher profitability despite the more uncertain macro backdrop? Nikesh Arora Well, remember, there's revenue and there's cost. Now, with revenue, you have -- the more you become a [reliable] (ph) business, the more visibility you have on revenue. So, we have ample visibility in revenue. Our percent of revenue that we know for next quarter continues to go up as you get more and more business converted to NGS ARR. So, I know what my revenues going to be, because they've booked in the past, right? So, I have a good sense of my top-line, barring my product sales, which I do in the quarter. Most of my billings or most of my software sales in the quarter don't have a huge impact on revenue in the quarter, it actually benefits next quarter. So, hardware benefits this quarter, most of everything else benefits next quarter, because most of the businesses are done in the last three, four weeks, right? So, I have a reasonably high visibility on my revenue on a quarterly basis, barring product sales. On cost, those are in our control, right? And I always maintain that the bigger our revenue base becomes, the more leverage you have on profitability, because then you have a lot more revenue to amortize costs against. So, we're at a point where we have a reasonably good sense of our cost bases. The macroeconomic environment gives us the cover to be able to make sure we can do cost-managed things. When everybody is talking about is hiring people up to -- and there's a -- two quarters ago, the question you were asking me was, what's going to happen to wage inflation? There are so many jobs out there. There's not enough people. There's a talent war. What about attrition? Now, you're asking me how are you going to maintain cost? Guess what? My attrition has gone down. My cost of hiring has gone down, right, which is good for me, because I have less productive people leaving. Don't underestimate if my attrition is running at 18%. There's a huge cost of onboarding, a fifth of your workforce gets renewed every year. That's a lot. Plus, the added people. I'd be at 14,000 people. I joined Palo Alto with 5,000 people. You do the math. 12,000 people are new to Palo Alto out of 14,000, just based on attrition. Attrition has gone away. Wonderful. My cost of onboarding, my costs are training, my cost of productivities, I mean -- so there is leverage in the system. So, that's the leverage we're trying to bank on to try to make sure that we can focus on better operating margins. Saket Kalia Absolutely. Maybe the last question here, just in the short amount of time that we've got left. Capital allocation. I mean just with the increase in free cash flow margins and just the cash that the business is generating, how do you think about sort of rank ordering the priorities that you and the Board have in terms of uses of cash? Nikesh Arora You should be asking that question to the FAANG companies. We don't have a cash -- we don't have excess cash problem. They have excess cash problem. Look, we have convertibles to pay, $3.6 billion of converts we have to pay out. We will generate north of $1.5 billion of cash this year. If we pay all the converts, we'll probably end up with $3 billion, $4 billion in cash. The one silver lining of that big hammer is I can get interest on my cash now, which I never got in my first 3.5 years at Palo Alto. Saket Kalia That's right. Nikesh Arora So, we spent a lot of time with our Treasurer. We're trying to maximize our interest income. I asked this question, and I will leave that question with you guys and feel free to e-mail Saket with the answer, because not me. I went to business school as well, like many of you, and we're taught capital allocation, and what is that Miller- Modigliani, or whatever? Saket Kalia Yes, Miller and Modigliani. Nikesh Arora That guy, right? Those two guys, right? And the challenge we have is that right now, if I use cash to buy back stock, the EPS impact I have is lower than if I put the money in the bank and get interest. So, then the question for MBA students is, what would you do? Would you get more EPS because the market is short-sighted? Or do you rather buy stock, because long term, your stock is worth a lot more? And I've asked this question to legendary fund managers, I won't name them, and they've sided on the, put it in the bank and get more interest income. What do you think, Saket? He writes every day about everything I do. What do you think? Saket Kalia Yes. No, listen, I mean... Nikesh Arora What should I do? Saket Kalia Well, I actually really like the chart from the Analyst Day that had a really balanced... Nikesh Arora I'm asking -- answer my question, dude. This is not fair asking one question, he hedges. What's the answer? Do you want me to do interest income? Or do you want me to buy back? Saket Kalia I would prefer whatever maximizes EPS. Nikesh Arora You heard it from here, Saket. Saket Kalia Well, with that, guys, I think that's about all the time that we got with Palo Alto Networks. Nikesh, thank you so much. Nikesh Arora Thanks, Saket. Thanks for having me.