Good day, and thank you for standing by. Welcome to the Nutanix Q2 2026 Earnings Conference Call. [Operator Instructions] Please be advised that today's conference is being recorded. I would now like to hand the conference over to your first speaker today, Rich Valera, Vice President of Investor Relations. Please go ahead.

Good afternoon, and welcome to today's conference call to discuss second quarter fiscal year 2026 financial results. Joining me today are Rajiv Ramaswami, Nutanix' CEO; and Rukmini Sivaraman, Nutanix' CFO. After the market closed today, Nutanix issued a press release announcing second quarter fiscal year 2026 financial results. If you'd like to read the release, please visit the Press Releases section of our IR website. During today's call, management will make forward-looking statements, including financial guidance. These forward-looking statements involve risks and uncertainties, some of which are beyond our control, which could cause actual results to differ materially and adversely from those anticipated by these statements. For a more detailed description of these and other risks and uncertainties, please refer to our SEC filings, including our most recent annual report on Form 10-K and our subsequent quarterly reports on Form 10-Q as well as our earnings press release issued today. These forward-looking statements apply as of today, and we undertake no obligation to revise these statements after this call. As a result, you should not rely on them as predictions of future events. Please note, unless otherwise specifically referenced, all financial measures we use on today's call, except for revenue, are expressed on a non-GAAP basis and have been adjusted to exclude certain charges. We have provided, to the extent available, reconciliations of these non-GAAP financial measures to GAAP financial measures on our IR website and in our earnings press release. Nutanix will be participating in the Morgan Stanley TMT Conference on Monday, March 2, in San Francisco, and we will be holding our Investor Day in conjunction with our .NEXT user conference on Tuesday, April 7, in Chicago. We hope to see you at these events. Finally, our third quarter fiscal 2026 quiet period will begin on Friday, April 17. And with that, I'll turn the call over to Rajiv. Rajiv?

Thank you, Rich, and good afternoon, everyone. In our second quarter, we continue to see healthy demand for our solutions as reflected in our strong bookings and outperformance versus our guided metrics. We see this demand driven by businesses looking to modernize their IT footprints, adopt hybrid cloud operating models and deploy cloud-native applications, including AI. I'm excited to announce our strategic partnership with AMD, which focuses on the growth opportunity in agentic AI. This multiyear collaboration is focused on development and marketing of a Nutanix-powered agentic AI platform for enterprises and service providers built on AMD accelerated compute infrastructure. As part of the agreement, AMD will make a strategic investment of $150 million in Nutanix common stock and fund up to $100 million for R&D and go-to-market for the combined solutions. We look forward to delivering the first jointly developed platform from this partnership to our customers in late 2026. More broadly, we see AI as driving a whole new set of new enterprise inferencing and agentic applications for which we are in the early innings. Nutanix provides the ideal platform to run them efficiently and securely on the hardware of their choice across any location, enterprise data centers, edges or cloud service providers. We see this as a significant long-term growth opportunity for us. Moving to our results. In our second quarter, we delivered quarterly revenue of $723 million, above our guidance range, grew our ARR 16% year-over-year to $2.36 billion and saw solid free cash flow generation. We also added over 1,000 new customers, representing our strongest quarterly new logo additions in 8 years. Looking ahead, we continue to operate in a dynamic environment and supply chain challenges, which were not a meaningful factor in our first fiscal quarter, became much more acute subsequent to our last…

Thank you, Rajiv, and thank you, everyone, for joining us today. I will first discuss our Q2 '26 results, followed by an update on Q3 '26 and full fiscal year 2026. In Q2, we reported results that were above the high end of the range for all guided metrics. In Q2, we reported quarterly revenue of $723 million, higher than the guided range of $705 million to $715 ARR at the end of Q2 was $2.356 billion, representing year-over-year growth of 16%. NRR, or net dollar-based retention rate at the end of Q2 was 107%. Land and expand bookings in Q2 were higher than our expectations, which we believe is partially due to customers anticipating supply-related shortages and price increases for server hardware from our partners. As we move through the latter half of Q2, we began to see that the challenging supply environment for CPUs, memory, storage and other components is delaying our customers' ability to procure servers from our hardware partners in order to run our software. We did not see this as a meaningful driver of our results in our fiscal Q1 as mentioned in our last earnings call. But later in Q2, we did start to see it become more pronounced to a greater extent than anticipated, and expect it to continue through the rest of the fiscal year. Thus far, longer lead times largely driven by lack of CPU availability have been a significantly bigger challenge for us than pricing. Revenue in Q2 saw a headwind from these factors, but was more than offset by higher-than-expected TCV bookings that grew in the mid-teens percent and lower-than-expected percentage of land and expand bookings with future start dates. In Q2, average contract duration was 3.1 years, largely consistent with our expectations. Non-GAAP gross margin in Q2 was…

[Operator Instructions] Our first question will be coming from Radi Sultan of UBS.

Starting with Rajiv, just on the VMware replacements, like can you just take a step back for us and like Broadcom has been making a lot of changes like reworking the hyperscaler agreements for VMware on cloud, ramping up audits, things like that. Like just taking a step back, like how do you think about the magnitude of the VMware replacement opportunity versus 12 months ago? And maybe what have been the biggest drivers of that change?

Radi, Rajiv here. I don't think anything has really changed from an opportunity perspective. You can see that we are still continuing to add new customers at a healthy clip. We added 1,000-plus customers this last quarter. Our AHV percentage, which is our hypervisor adoption has hit an all-time high. To your point around the cloud partners, you've seen that we've extended our partnership with AWS post this Broadcom acquisition, and we're seeing some good traction. We talked about some examples on NC2 and how that's being used in the public cloud. And then we continue to add more external storage support. And you see the EverPure solution is now in the market. PowerFlex has been in the market and more public cloud options now with Google coming on. So all of these make migrations easier, right, from VMware. So we don't think anything has changed from an opportunity perspective. I've always said that this is a multiyear journey, and it will continue to be a multiyear journey for us.

Got it. Crystal clear. And then just a follow-up for Rukmini. Just as you think about like in light of the supply chain constraints, like the implied Q4 guide does imply a pretty steep upsell. So just like what gives you confidence in that? And if you could just speak to your level of visibility, like what leading indicators should we be looking at to kind of give us confidence in underwriting that Q4 upsell?

Yes, Radi. So I would say a few things. One is that Q4 is traditionally our -- a very strong quarter for us. It's the end of the fiscal year, of course. So our sellers are all incentivized to maximize what they can do in Q4. And we also are guiding to Q3 and by implication to Q4 based on what we're seeing today in terms of supply in the market and when we can expect those to be reflected in our revenue, Radi. So those are sort of the two things I'd point to. And I'd also say, look, our RPO had a nice growth in Q2 at the end of Q2. RPO grew 24%. So that's one other metric you can look at in terms of how much is remaining, right, in terms of our performance obligations and cRPO as well. So those are sort of the 2 or 3 things I'll point to in terms of Q4 -- Q3 to Q4 seasonality.

And our next question will be coming from the line of James Fish of Piper Sandler.

I mean, Rukmini, you probably were teeing that up for me with the RPO comment there. But I did want to touch on that. Help me with this dynamic where you guys have mid-teens TCV bookings, but RPO bookings growing [ 4% ] against that really tough 34% comp last year. I mean that implies about 10 points coming from sort of off-balance sheet bookings that is a materially higher amount than what you guys have historically suggested? Or what am I missing here that you're starting to see more off-balance sheet bookings?

Jim, yes, so as I said, we saw strong year-over-year RPO bookings growth of 24%. And I believe what you're referring to, Jim, is the calculated bookings, which for everyone's reference is calculated as Q2 revenue plus quarter-over-quarter change in RPO. So similar to last quarter, Jim, the exclusion of cancelable backlog from our reported RPO has a meaningful impact on the RPO bookings calculation and year-over-year growth rate despite the fact that, that cancelable backlog remains a very small portion relative to our total RPO number. It's in the low single-digit percent as a percent of the total RPO. Now including that cancelable backlog number, which historically has very rarely been canceled, in the RPO bookings calculation, that results in a mid-teens percent year-over-year growth rate aligned with the overall TCV bookings growth rate that I alluded to in my prepared remarks.

Okay. And Rajiv, for you, I guess you're talking about a really good pipeline of conversion opportunity. But how are customers balancing the outlook of having to renew again with that incumbent and delaying their hardware purchasing versus the savings they get by deploying the hardware now and purchasing Nutanix?

Yes, it's a good question, Jim. So -- and we see kind of a combination of different things here, right, for all these customers. So -- there are customers who have gone out there and ordered hardware and they realize that prices are going to go up and they've ordered that ahead of time, and they are going ahead with the migrations. There are also customers who are looking and saying, okay, well, I've got my external storage now, and you guys can actually help me on that front without having to have me replace hardware. And that's a good -- that works very well for us actually. And so the migration, if you look at a customer migrating from VMware, it's obviously already tied to hardware refresh cycles. And if a hardware refresh is needed, we do very well from a TCO benefit compared to our competition, which is why people are moving over. Now if a customer decides to sweat their hardware in this environment, given your pricing of new hardware, we can still enable a software migration, right, because of the fact that now we're supporting a larger number of third-party storage and also being able to work on existing HCI hardware that runs VMware today. So we are offering our customers all of these options. And on top of that, they can also migrate to the public cloud, by the way, if they choose to. And there are some customers who are doing that. In fact, we talked about one of them here on the call. So we have all these choices, and that helps offset some of these hardware migration concerns -- hardware cost constraints, I should say.

And our next question will be coming from Sanjit Singh of Morgan Stanley.

Just some follow-ups on some of the headwinds in terms of revenue. Rukmini, last quarter, we talked about deferred start dates and that higher mix of deferred start dates. That doesn't seem to be the sort of near-term. It seems like that maybe came in line with your expectations and it's more of the supply constraints around CPU. I wanted to make sure that I understand the dynamics happening right now. And then it seems like the longer term or the midterm strategy here is to keep the bookings momentum alive by -- or keeping the bookings momentum strong by giving customers more flexibility when it comes to licensing and start dates. Is that how I should be thinking about it in terms of how you plan to navigate kind of the near-term turbulence around on long run on revenue of Nutanix?

Yes. Sanjit, let me perhaps first lay out sort of how we take our software to market, and then I'll address your question on future start dates and flexibility and so on. So we have two types of orders for our software, right? So one go-to-market motion, which typically accounts for the majority of our ACV bookings, involves software not linked to any hardware delivery, which we often refer to as software only. Now for the software-only deals, we typically will build a customer and recognize revenue -- so sorry, we'll build customer at the time the purchase order is made and revenue is recognized at the time of the license start date, as we've talked about before, and that's what you're alluding to. Now the other type of order involves software that is tied to the delivery of the associated server. Now this includes business that's sold through our OEM partners as well as integrated offerings such as we offer with Supermicro. And in that case is where the software is tied to hardware delivery, revenue recognition is also tied to that delivery of the hardware. So now to your question on future start dates, you are correct that the percent of orders and future start dates in Q2 was lower than we had expected and was lower than what we saw in Q1. And that percent or that mix depends on a variety of factors, right? It includes customer migration time lines. It includes that time line to procure hardware as we're talking about now, and our approval processes around that. So -- and that number can move around from quarter-to-quarter. And given the recent supply chain dynamics and ongoing migrations, we have factored in a higher percent of orders than we had previously to come in with future start dates going into the second half of the forecast.

That's very clear, Rukmini. I appreciate the color there. Let me talk a little bit about the partnership with AMD. So Rajiv, maybe in terms of the time line of having a solution out to the market, maybe get your viewpoint on that. It sounds like it's going to be later this year. And then in terms of which customer segments within the Nutanix or AMD customer base will be most interested in the joint solution? Is it sort of particular industries, higher segments like large enterprise. I just want to get a profile of who's going to be most viable as potential joint customers of the partnership.

Yes. First of all, the target, just to be very clear, is we are building a platform together, that supports essentially inferencing and agentic applications, right? So this would be a full stack platform on top of which people could use run models and build these multi-agent applications or even simpler inferencing applications. The target customer base, to your question, really are enterprise customers, okay? Enterprise customers. And if you were to say within that enterprise customer base, it would be typically customers in more regulated industries, customers who care about sovereignty wanting to run their AI close to where their data is located. So it could also mean people who are running AI applications at the edge. So if you were to look at this set of customer base, right, what it translates into is people who care about protecting their data, running it locally, running it inside their data centers for the most part, running it in edges. But also some subset of these customers are going to consume them from service providers. We have a lot of service providers these days who are also offering GPU as a service. So that would be the second way of serving the same end customer being the enterprise. So in short, targeting enterprise use cases for these agentic AI applications, mostly regulated industries to start with, but also people are interested in being service providers delivering to these enterprises. And as you said, the first solution out of this will be delivered by the end of this calendar year, and we are excited about the long-term potential for this partnership.

And our next question will be coming from the line of Wamsi Mohan of Bank of America.

It's Ruplu filling in for Wamsi. I've got two questions, I think, today, both for Rukmini. Just looking at the margin dynamics, I mean, the fiscal quarter you reported was very strong, stronger than expected. You're guiding fiscal 3Q operating margins to 16% to 17% and then the full year unchanged at 21% to 22%. Can you just help us understand the dynamics impacting fiscal 3Q margins? And what's the -- what are the factors that will help get to the full year 21% to 22%? And the same question on free cash flow. It looks like revenues are down $20 million for the full year, but you're taking down free cash flow down $60 million, but there's also potentially some -- the $100 million from AMD and they're also buying back shares or buying your shares. So can you just give us your thoughts on the cadence of free cash flow, the investments you're making and how we should model out free cash flow? And I have a follow-up.

Thank you, Ruplu. So on operating margin, I'll start there. So historically, our seasonality around operating margin is that it's higher in Q2 and Q4 and lower in Q1 and Q2 on a relative -- Q1 and Q3 on a relative basis. And that's because the revenue roughly follows that pattern as well. And of course, we're investing typically through the year. So that also flows in. But it's not unusual for us to have operating margin be higher in Q2 and Q4 relative to Q3, which I think is the dynamic that you're pointing out. And I will say that in terms of overall -- our overall view on investment, look, we still continue to believe there's a large growth opportunity ahead of us. And so we're investing for that growth, but doing so in a thoughtful way in a prudent way while also working on efficiencies and productivity within the business. So that's on the margin point, Ruplu. And then on free cash flow and your point about revenue versus free cash flow. So one thing I'll say is that when you think about the orders that I described earlier, where orders that are linked to the shipment of server hardware, and we said that those are a minority of our bookings, we can only recognize software revenue and collect cash alongside the shipment. And as a reminder, we typically recognize about half of an order's TCV bookings upfront as revenue, while we typically collect all of the cash upfront. So you would impact the cash impact on free cash flow to be higher than on revenue, and that is, in fact, what we're expecting. So the change to our free cash flow guidance was largely due to these longer server lead times, which is delaying the timing of when we can bill and collect from customers. And that's why the free cash flow guidance change is higher than the revenue change, Ruplu. So those are some of the dynamics around free cash flow and on operating margin.

So Ruplu, you also asked about the AMD piece, right? Do you want to cover the accounting?

Yes. So AMD is choosing to invest $150 million in our common stock, and that really has no impact on free cash flow because it's not going to be in the operating cash flow section.

Right. And -- but they're also investing $100 million. Is that all this year?

Correct. So that's -- I will say, look, I think we aren't getting into the specifics of that, Ruplu, but we did say it was a multiyear strategic partnership with AMD, and it's up to $100 million over that period over that period -- over a multiyear period.

Got it. If I can ask a quick follow-up. So you mentioned 2 delays. One is the delay because of availability of servers. And then you also, in the commentary said that there's a higher the orders with delayed start dates, I mean, that is also growing versus your thought 90 days ago. Which is -- can you help us quantify this? I think last quarter, you had said it was about $80 million of revenue that was being pushed out because of future start dates. Can you help us quantify how much is the impact from delayed availability of servers versus orders with later start dates? Like how are each of these impacting revenue recognition?

Yes. So for this quarter's updated guide, Ruplu, I would say the more meaningful change is in the supply chain environment. It is worse than what we had expected 3 months ago. And so that is a meaningful portion of the change in guidance. And as we said on the call, we're also assuming that the second half orders with future start dates, the mix of those orders is also expected to be higher than we had thought. So that is also factored in there. But we're not breaking out the quantification, Ruplu, but the more significant change was what we saw in the supply chain environment.

[Operator Instructions] Our next question will come from Matthew Martino of Goldman Sachs.

Maybe to start with Rukmini. You mentioned in the prepared remarks, extended lead times have been the bigger headwind and less so server prices. Can you provide us with some color on kind of what you've contemplated around guidance in the back half of the year, whether it's largely extended lead times or potentially lower initial deal sizes on higher server costs, just given the fluctuations out there in memory prices today?

Yes. Matt, so in terms of the forecast for the second half, we have factored in delays in availability of servers, right? So that is -- we factored in based on the visibility we have right now and what we see. But I will say, I think we've said this before, that we are indirectly impacted by this given that we -- it's our server partners that we rely on and our customers rely on to procure their hardware. But yes, we factored in some of the delays around server lead times and when we expect those to become available. Now the second point about potentially impact of pricing on our software. As we've said, we continue to see solid demand for our solutions in Q2. We've talked about the bookings growth. And so far, I would say the impact of this higher server price and longer lead times can be mixed, Matt. And what I mean by that, in some cases, it is accelerating projects, while in other cases, it's requiring more approvals and lengthening procurement time lines. And as we said, we are also working with our customers to offer them many options, including choice of server vendor, external storage support, public cloud with NC2 solution and swapping software on existing hardware. So, so far, we have not seen a meaningful impact to our software deal sizes or pricing in the market. Although customers are, in fact, experiencing higher server prices and looking for ways to mitigate that. So we will continue to emphasize our software value proposition and provide a range of options to customers as we have described. So that's the way we're approaching this, Matt.

Yes. Very clear. And then for you, Rajiv, you guys are offering multiple workarounds. Rukmini just flagged it with NC2, expanded server platform choices, external storage. I'm curious, Rajiv, from your perspective, what's proving most effective for those deals that are staying on track in terms of providing that wider array of options? And then in context of that, the new logo velocity was really strong this quarter. So wondering if Pure Storage was a large contributor to that dynamic?

Yes. I think Pure Storage or EverPure is still very early. We did close our first wins this quarter, Matt. But I wouldn't say that was a big contributor to logos yet. Most of the logos came from our typical VMware migrations on to the HCI platform. External storage is growing, but it's still a small portion. Now as we look at sort of which of these are moving the needle in a significant way, it's really a combination of all of these factors. The fact that we're able to run our HCI platform, I mean, now support a broader set of configurations and provide more choice to customers. There was a lot of arbitrage in the quarter about people looking at different server vendors to say, who can get me the best price and best lead time. And the fact that we could support all of them, I think that's clearly a value proposition and a factor that our customers took advantage of for sure. NC2 had a pretty good quarter, right? So people are starting to use the public cloud more and more as in terms of the -- getting around the fact that they can't get servers on-prem. They can also potentially use NC2 to do a migration to the cloud. And if they want to move it later on-prem and they have servers available, they can do that, too. So we are starting to see some of that come through as well. The external storage, like I said, is a third component, which is still early days. We're starting to see the I see the traction very clearly, though. And I do expect that as customers -- if they start looking at sweating their existing hardware more, given the supply constraints that they're seeing, then it's going to start to become even more important. So I think it's not that there's one thing that's driving -- helping us in a big way compared to the others. It's a combination of all of these things that's helping us mitigate the supply situation.

And our next question will be coming from the line of Samik Chatterjee of JPMorgan.

This is MP on behalf of Samik Chatterjee. So firstly, just wanted to double-click on your AMD deal. What is the best way to size up your revenue opportunity here and the go-to-market strategy? Will it be just the attach rate with AMD's inferencing platforms or you will be able to do independent software sales as well? And then what is the likelihood of similar future deals? And how does the market look like?

Good question, MP. Let me take that. So first, in terms of the revenue itself, right, obviously, so the solution is the Nutanix software stack running on servers that include AMD's accelerated compute platform, right? So we take -- by the way, we already have solutions in the market today with NVIDIA on this front, right? In fact, all our AI solutions today run on NVIDIA, and now we are adding AMD to the mix to provide more choice for customers. So the go-to-market model is that we sell our software, the full stack, which is our Nutanix Cloud Platform, our Kubernetes platform and then our NAI, the Nutanix AI piece that runs on top of all of that. All of that is now -- the R&D effort is to take all of that, get it to work very nicely and optimally on AMD hardware, right? That's inside the servers, AMD's GPU hardware and accelerated compute hardware. And so we sell our software just like we sell today. And more of that software will land on platforms that have AMD hardware inside. So that is the go-to-market collaboration, right? So AMD brings together their hardware solution, their ecosystem. And we then put our software stack on top, and we take that to market, take it into our enterprise customer base, and that is the go-to-market portion of this. Now revenue-wise, of course, the solution is going to be in the market at the end of this year. We expect to start seeing small amounts of revenue next calendar year or the second half of FY '27, our fiscal FY '27.

And my second one would be around supply constraints, between CPUs and memory, which are bigger supply constraints in your opinion, currently for server deliveries? And then how -- when do you expect the supply constraints to ease? Like do you expect it to ease within this fiscal year or go beyond this fiscal year as well?

Both good questions, again, MP. So currently, I mean, look, I think the real answer is actually both. Right now, the critical item in the path is the CPUs. But very quickly, if we start -- if the CPU supply situation starts getting better, in fact, Intel has talked about this, right? During their last earnings call, they talked about how they were way behind in terms of meeting their demand. And certainly, we're seeing some of that, right? So that's in the critical path right now. We also are bringing in more and more AMD SKUs to market, right, so that we can actually offer both AMD and Intel to ameliorate the situation a bit. But then very soon right after that, you run into memory. So the answer is actually it's both. And with respect to timing on when this could get better, frankly, I think it's unclear at this point. I don't think it is very short term. I think it's going to be there for a period of time because obviously, this is being driven by the massive uptick in AI spending across some of these bigger hyperscaler providers that's causing this big supply issue overall in the industry. And so I do expect it's going to take some time to play out for the supply situation to normalize. And it's hard for me to tell you exactly when that's going to happen. And keep in mind that we are a step removed from all of this, right? We are indirectly impacted. So it's very hard to predict exactly how quickly. But I do expect that -- I mean, obviously, the suppliers there, both the CPU folks as well as the memory folks are working hard to go get more supply online, and this is an industry-wide phenomenon. So again, it's hard for me to give you an exact prediction on when this is going to get normalized, but it will get normalized over some period of time here over the next couple of years.

And our next question will be coming from Ben Bollin of Cleveland Research Company.

Rajiv, I'm curious your thoughts how you think about the mix of revenue or ARR orders that you're getting from AI inference, agentic opportunities today. Essentially, how big is it? And what type of growth rate do you think you're seeing from that opportunity? And then I had a follow-on for Rukmini.

Yes. So Ben, first of all, it was 0, 1.5 years ago. And so clearly, you're growing from a very small base, and it is still fairly small for us in the scheme of things because we are in the very early innings of enterprise AI adoption, really early, right? Our customers today, we talked about one of these customers in the EMEA region, who's now created the shared GPU platform that they are providing to various departments. So we are -- the use cases today tend to be fairly simple applications, simple inferencing applications, not multi-agent use cases yet. So we are in the very early stages of enterprises building these more sophisticated AI applications that they will consume. So for us, I mean, I look at this and say, over the next 5, 10 years, there's going to be a whole bunch of new applications that are going to be built in the enterprise. Existing applications will all start incorporating AI, workflows are going to change. So there's going to be a whole new set of workloads from our perspective that are -- including gen AI as part of these applications. And so for us, the opportunity is to become the underlying platform to run those applications, manage all the data associated with it. So we are pretty early in the cycle at this point. And enterprise adoption, like I said, is just starting really. I mean, because most enterprises at this point are still grappling with AI. They're starting to use it more and more, but they haven't really built a lot of their own custom applications and custom workflows. but they're all starting to do it. And when that happens, I think this will grow. So short answer to your question, it is -- it was 0, 2 years ago. It is small right now, but growing and growing very nicely. And we have -- again, the fact that we now have a broad ecosystem that cuts across NVIDIA and AMD, along with more third-party people coming to the mix here, we are very focused on this. Please come to our April conferencing event where we will talk more about our AI strategy and road map. But we do expect this to be a good long-term driver for us.

That's helpful. Rukmini, I'm also interested in your thoughts. When we look at the RPO and booking strength, obviously, the appliance pricing is going higher from these vendors and it's changing by the week or the month or whatever. How much of what you're seeing right now would you attribute to customers getting in line because they'd rather put the PO in today, try and get current prices versus assuming the price may be higher at some point in the future? How do you think about that?

Yes. Look, we do think that some of our bookings performance -- outperformance in Q2, Ben, was tied to that. It's hard to be very precise and try to quantify that. But we do believe that our bookings in Q2 benefited from some orders being placed, as you say, earlier than originally expected as customers look to get ahead of server price increases or try to ensure access to servers. But it is difficult to quantify.

And our next question will be coming from Brandon Nispel of KeyBanc Capital Markets.

I think just one for me. NRR decelerated this quarter, I think, Rukmini, you called out 107%. Could you give us some color in terms of why NRR decelerated? That would be helpful.

Brandon. So the first thing I'd say it's important to note that the factors that we outlined last quarter and I'm talking about this quarter as shifting out the recognition of revenue also impacts ARR and NRR, because the dynamics are similar in terms of how we recognize revenue versus when something would show up in ARR and therefore, will be also reflected in NRR. So just important to keep that in mind, Brandon, right, because those dynamics are very similar. Now in addition, I will say that our NRR in Q2 specifically was also impacted by timing delays we saw in some renewals in our -- with regarding our U.S. Fed business specifically that was created to some backlog related to the recent government shutdown, and we expect to receive those in Q3. So that was the timing delay from Q2 into Q3. And so those are sort of the things I'll call, other than things we've talked about before, right, in terms of, look, our ASP of new logos have become -- have steadily increased over time. And we've also talked about as ARR grows every quarter, the ACV dollars required to offset a point of churn increases even at the same churn percentage. And so that can make it challenging to achieve the same NRR over time. So the last two I mentioned are ones we've talked about before, but I will call out the first two as being -- just things to keep in mind as you look at NRR.

And our next question will be coming from Nehal Chokshi of Northland Capital Markets.

Yes. Congratulations on the good bookings quarter there. Talking about bookings, based on the color you provided so far, Rukmini, it sounds like cancelable bookings, while it's still in the single digit of the overall RPO, it is materially more than the total in terms of the year-over-year growth rate. So can you discuss what is behind canceled bookings year-over-year growth being much higher than non-cancelable bookings?

Nehal, so cancelable bookings, like you said, I think this is again referring just for everybody's benefit, a bookings number that I think folks can calculate based on looking at our revenue in Q2 and adding that to the quarter-over-quarter change in RPO. And the way RPO is defined for us like it is for most companies is -- only includes orders that are non-cancelable. Now as I said earlier, we have a small portion. It's in the low single-digit percent of our total RPO that is cancelable. It remains in the low single-digit percent. But it can move around somewhat, Nehal, right? And which is why it sometimes distorts these year-over-year growth rates because you're looking at a quarter-over-quarter change in RPO and then we're trying to do a year-over-year calculation in terms of growth rates for bookings. So I would say it moves around a little bit. And why do we even have these cancelable orders? Like I said, internally, we look at them all as bookings because historically, these cancelable orders, even though they're officially cancelable, have very rarely been cancelable. And so yes, there's really nothing else there other than we've had some old partners that have some of those clauses, and we're actually working to make sure that those are no longer the case, frankly, going forward, right? So over time, we would expect that cancelable number to go down, but it can move around from quarter-to-quarter because we have some of these from before.

Okay. So just to be clear, it's basically just quarter-to-quarter variability in these cancelable bookings, but it is 2 quarters in a row where we are seeing cancelable bookings have a greater growth rate than non-cancelable bookings.

Yes. As I said, when you calculate the bookings number, the using RPO that is taking the entire bookings and that's taking all of the RPO into account, the TCV bookings growth rate is in the mid-teens, which is also what we said in the prepared remarks. And there's some variation because what we report out externally to you all, we only include the non-cancelable orders in the RPO.

Okay. My other question, which I think is probably more for Rajiv. Can you explain -- I mean, clearly, Nutanix does not need the cash from AMD. Why accept an investment at $36 a share when arguably that's a very attractive price for AMD and not so attractive for existing Nutanix shareholders?

Yes. No, Nehal, that's a good question for sure, right? So our belief is that the equity investment from AMD aligns their interest with Nutanix' success. And they are a very important ecosystem partner for us. If you look at the AI ecosystem out there, it's built around NVIDIA to start with and AMD, right? So for us, it's really important to make sure that we are well aligned and both sides want us to succeed. And this equity investment from AMD makes it very much their interest to see Nutanix succeed. The dilution from this, clearly, we understand there is dilution. It is quite small overall. If you look at it, it's somewhere slightly over 1%, I believe. But the fact is that this partnership enables Nutanix to become a leader, right, in providing an agentic AI platform for enterprises and service providers. And together with NVIDIA, I think now we have a complete solution that we can take to market across both of the major players in AI.

And our next question will be coming from Matt Hedberg of RBC.

It's Dan Bergstrom for Matt Hedberg. To follow on with another bookings question, it sounds like Q3 bookings expectations are higher than previous, which is great. Now with the supply chain challenges, can you just help us think through your available to renew second half pipeline on a year-over-year basis? And then maybe how that's changed versus assumptions coming into the year?

Yes. So look, I think the available to renew pool is largely land and expand. It's driven by land and expand bookings that we've made in prior years. And so -- and then the other thing to note on renewals is that typically, those are not as tied to hardware requirements, as land and expand would be, right? So a lot of the dynamics we're talking about here with regard to supply chain are really more tied to land and expand rather than to renewals. And so yes, I think the available to renew pool again doesn't change based on these which are usually set largely based on land and expand bookings that we did in the past and are coming up for renewal in the second half. And so that's still the case for the second half in terms of just how that's factored into our forecast.

And our next question will be coming from Mike Cikos of Needham.

I just wanted to come back to this AMD partnership announcement real quickly. And I think it's probably for you, Rajiv. But if I think about how Nutanix has developed or evolved over the past couple of years, the entire product road map really ties to providing customer choice, right? And I know that you're highlighting the expanded AI opportunity here. But is that the #1 takeaway we should be thinking about with this Nutanix-powered agentic AI platform in concert with AMD? I guess what else should we be thinking about in light of this platform offering?

Yes. I think I covered some of this a bit earlier, Mike, but there's a couple of points that I want to make, right? One is there's an AI ecosystem being built, and it's really being built around 2 players right now, right? NVIDIA is the lead by far, right? And of course, therefore, for you to be successful, you got to work with the NVIDIA ecosystem on the one side, which we are already doing. The other big player is AMD, right? And so it's very important for us to be able to go work with that AMD ecosystem as well. And at the end of the day, customers, like you said, we are all about offering customers choice, right? A choice of at every layer in the stack. And AMD is going to be an important player here, and they already are an important player here in this accelerated compute market. And so it's very important from a Nutanix perspective to be able to offer both choices to customers in a consistent way, right? So we provide an AI platform for agentic applications that underlying that could be a customer using NVIDIA-based servers or AMD-based servers, right? And they get similar capabilities at the top level using our software. So that was the motivation for us to be able to provide the complete ecosystem. And again, it impacts sort of the whole picture, right? There's people who are certified -- if you look at third-party software, the ISV ecosystems, there's people who work with NVIDIA and there's people who work with AMD, right? And we want to be benefiting from both sides of that.

Excellent. And for the follow-up for Rukmini. I appreciate the shortages out there. Great to see the margin preservation. Can you just walk us through how -- I guess, the key investment initiatives this year and how you're actually managing against that, just given the reduced revenue profile? It would just be great to hear how you're operating against that.

Yes. Thanks, Mike. So in terms of investments, look, it's consistent with what we've laid out, I think, at the beginning of this fiscal year. We think of this as a huge opportunity ahead of us, how do we make sure we're investing in a way where we're directing investments to where we see a clear return. So on the go-to-market side, it's around making sure we have the appropriate coverage around the world where we see the opportunity. It might be in areas like portfolio specialists, for example, where we think there's an opportunity to sell more of our portfolio across our customer base. So it's areas like that, that we're investing in, Mike, in the go-to-market front. And then there's also some investments in R&D as we continue to innovate across all the things we've talked about here, right, in terms of Kubernetes platform, our AI platform, supporting external storage, et cetera. So a lot of important innovation initiatives as well in R&D that we want to make sure we are continuing to fund. So if you look at the overall operating margin guidance that we've maintained here, we're getting a little more contra expense from partners than we had expected at the beginning of the year and some lower commissions as well, given the revenues are getting shifted out of the year. So we haven't really made any changes to our investment plans for fiscal year '26, but some ins and outs, and that makes us feel comfortable with our overall full year op margin guide of 21% to 22%.

And this concludes today's program. Thank you for participating. You may now disconnect.