Nutanix Inc (NTNX) 2018 Q3 法說會逐字稿

Good afternoon. My name is Mike, and I will be your conference operator today. At this time, I would like to welcome everyone to the Nutanix Q3 FY 2018 Earnings Conference Call. (Operator Instructions) I will now turn the call over to Tonya Chin, VP of Investor Relations and Corporate Communications. You may begin your conference.

Thanks. Good afternoon, and welcome to today's conference call to discuss the results of our third quarter of fiscal 2018. This call is also being broadcast live over the web and can be accessed in the Investor Relations section of the Nutanix website.

Joining me today are Dheeraj Pandey, Nutanix's CEO; and Duston Williams, Nutanix's CFO.

After the market closed today, Nutanix issued a press release announcing the financial results for its third quarter of fiscal 2018. If you'd like a copy of the release, you can find it in the Press Releases section of the company's website. We would like to remind you that during today's call, management will make forward-looking statements within the meaning of the safe harbor provision of federal securities laws regarding the company's anticipated future revenue, billings, gross margin, operating expenses, net loss, loss per share, free cash flow, business plans and objectives, product sales, plans and timing for and the impact of our transition to focus more on software-only sales; expectations regarding products, services, product features and technology that are under development; competitive and industry dynamics; new strategic partnerships and acquisitions; our intent to acquire new technology; changes in sales productivity; expectations regarding increasing software sales; potential market opportunities and other financial and business-related information. These forward-looking statements involve a number of 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. These forward-looking statements apply as of today, and you should not rely on them as representing our views in the future. We undertake no obligation to update these statements after this call. For a more detailed description of these risks and uncertainties, please refer to our quarterly report on Form 10-Q for the second quarter of fiscal 2018 filed with the SEC on March 15, 2018, as well as our earnings release posted a few minutes ago on our website. Copies of these documents may be obtained from the SEC or by visiting the Investor Relations section of our website.

Also, please note that unless otherwise specifically referenced, all financial measures we use on this call today are expressed on a non-GAAP basis and have been adjusted to exclude certain charges. We've provided reconciliations of these non-GAAP financial measures to GAAP financial measures in the Investor Relations section of our website and in our earnings press release. As a reminder, all results included in today's call and press release are using the newly adopted revenue standard, ASC 606.

Lastly, Nutanix will be at the Stifel conference in Boston on June 12 and also at the Nasdaq European conference in London, also on June 12, and we hope to see many of you there.

Now I'll turn the call over to Dheeraj. Dheeraj?

Thank you, Tonya. Good afternoon, everyone. Q3 was a quarter in which we accelerated growth while operating at above a $1 billion run rate. It was strong across the board, with billings, revenue and gross margins all ahead of consensus. This is the third quarter in our transition towards a software-defined business model, and I can say that we have managed change immensely well. Our results were bolstered by strong HP adoption as well as continued penetration of the Global 2000. We're just back from our fourth .NEXT user conference in New Orleans, and I couldn't be more excited about the way this event has evolved. Just 3 short years since our first .NEXT conference, attendance has grown from approximately 900 to nearly 5,000 attendees, raising the total to more than 20,000 who have joined us at our global .NEXT events over the past year.

What I'm proud of is that .NEXT continues to be hyper-focused on our customers. From our first hackathon, where customers like the County of San Mateo, Cardinal Innovations Healthcare and Tucson Medical Center, developed exciting new projects for our ecosystem, to the .NEXT Awards, where we recognized Home Depot, JetBlue, the Multi Commodity Exchange of India and CSRA for their contributions. Our customers are at the center of our success, and we're delighted to be able to celebrate them each year. We're also pleased by the positive reaction to the 3 announcements we've made at .NEXT. These 3 new additions to our product portfolio, Nutanix Beam, Era and Flow, are important not only for their strategic value to our business, but also for the new opportunities they open for customers on our platform. These offerings will help our customers better manage their hybrid cloud environments while helping to solidify our unique thesis of building an enterprise operating system from the ground up for a multi-cloud world.

I'd like to take a few minutes to tell you about each of these offerings so you can better understand the opportunity they present. When we introduced HCI in 2011, we disrupted the status quo by collapsing storage and computing to an easy-to-use platform. Legacy infrastructure companies weren't listening to their customers who complained about the cost and complexity of their legacy 3-tier architectures. The same was true when we collapsed a hypervisor into our stack, with the introduction of AHV in 2015. Security and the network were the next logical areas to focus our attention. These are legacy strongholds of complexity in the on-prem data center, and that's what we're tackling with Flow.

Flow will combine our native homegrown micro segmentation technology with application-centric network visibility technology from our recent Netsil acquisition, which we announced and closed in Q3. The innovators in the Netsil team, the Gill brothers and their college mates from the University of Pennsylvania, who developed the state-of-the-art observability technology for multi-cloud environments, are now a key part of Nutanix R&D. With Flow, our customers can secure network traffic and soon ensure their applications run optimally, with deep insight and visibility into how these applications are performing on the network.

As many of you know, our mission is to make infrastructure invisible. With Nutanix Era, we are redefining infrastructure to include databases. Enterprise platform-as-a-service, or PaaS, needs to straddle private and public clouds, edge and core computing and managed open source and bring-your-own licenses. Era dramatically streamlines the provisioning of new databases and automates the life cycle management of those business-critical databases. Similar to the functionality AWS offers in the public cloud with its relational database service, or RDS, we'll initially support Oracle and PostgreSQL databases, with future support for MySQL and Microsoft SQL Server. We expect Era to be available in the second half of calendar 2018.

Beyond databases, enterprise PaaS also includes message buses for event-driven qbased applications. At .NEXT, we showcased Project Sherlock, a multi-cloud message bus with which IoT applications can take advantage of machine learning and AI algorithms at the edge, where machines are producing data. Rather than bring immense amounts of data to the cloud, Project Sherlock brings the cloud PaaS to the edge. It was the most well-received keynote at the conference.

Beyond databases and message buses, enterprise PaaS also includes objects storage for next-generation applications. At the conference, we showcased how our high-performance, highly secure S3-compatible object storage service will become an integral part of our enterprise cloud offering.

Moving further up the stack, we also introduced our first-ever SaaS offering, Nutanix Beam, a multi-cloud cost governance and security compliance application that is based on technology from our acquisition of Minjar. Based on machine learning of metering, billing and network data, Beam makes continuous recommendations and provides one-click remediations to cloud deployments that have suboptimal economics and violate compliance checks. Over time, we expect Beam to become the policy engine that will drive mobility of applications between on-prem and off-prem.

At the conference, we also showcased the progress we have made with Nutanix Xi, spelled X-I, Nutanix Calm and Google. As we talked about at our Investor Day in March, we are providing early access to Xi this summer, delivering a true hybrid cloud with a seamless experience that harmonizes the architectures of owned and rented clouds. It hasn't been done before. Like hyperconvergence of hardware on-prem, the attention to detail in engineering and design to hyperconverged private and public clouds will create a very large opportunity for us in the enterprise. Our bet is that mobility will be the killer app in multi-cloud. Our engineers, designers and product managers have been burning the midnight oil to make sure that replication, multi-tenant networks and runbook automation are indeed one-click in this hybrid experience. Simple yet secure, and most importantly, that our own engineering processes are one-click in cloud deployment and change management. Our relationship with Google continues to make very good progress. Google Cloud's CTO, Brian Stevens, joined us onstage at .NEXT to provide an update and live demonstrations of native Kubernetes support for containers and Calm support for Google GCP. And this entire demo was working with no hardware on-prem. The entire Nutanix stack was running in GCP on Google's hypervisor. That is the power of our software-centric architecture. Like Dell, Lenovo and IBM ports, we showcased how our cloud operating system will be ubiquitous, running inside hyperscale environments like Google's. In addition to support for GCP, Calm now includes support for VMware environments. This continued expansion across public and private cloud environments brings Calm closer to reaching its full potential as an application orchestration and life cycle management offering for the multi-cloud world. Interest in Calm is accelerating with some of our most innovative customers. And in this past quarter, it was adopted by a diverse set of customers and industries including health care, insurance and banking.

Now let's move on to a few of our noteworthy highlights from Q3. During the quarter, 44% of our bookings came from large deals. We now have 67 customers with over $5 million in lifetime bookings, 22 of which have lifetime bookings of more than $10 million. The majority of those customers are in the Global 2000. In this quarter, our top 4 deals were in -- all in excess of $5 million each, and 3 of those deals were software-only. Of those deals, 3 were noteworthy for the customer's decision to go all-in with our hypervisor AHV. And our third largest deal of the quarter, with a Global 2000 retail customer, with whom we have more than $35 million in lifetime bookings, the customer will leverage our platform to run call center operations on AHV. This deployment will also enable the company to roll out a new tech support offering and is our largest deal with a customer to date. In addition to its success within the Global 2000, AHV is also a major decision factor for government entities worldwide. AHV was a significant factor in both our largest deal of the quarter, a new customer win with a large U.K. central government department, and our fourth-largest with an entity of the U.S. Department of Defense, DoD, which has lifetime bookings of more than $5 million. Public sector entities continue to adopt AHV for both the cost savings it provides and the increased functionality of our fully integrated stack. With these large deals and continued adoption across our customer base, AHV adoption rose to 33% on a rolling 4-quarter basis, up from 30% in Q2. We expect this adoption will continue to rise over time. In addition to AHV, our robust security continues to be a major decision factor for our public sector customers. In 2 of our top deals this quarter, with the previously mentioned DoD entity and another DoD entity with lifetime bookings of more than $16 million, that is composed of military, federal civilians and contractors, our security-first approach, scalability and ease of use were noted as major factors in our selection. In the second DoD deal I mentioned, the military, federal civilian and contractor-composed entity, plans to use Prism Pro for planning and analytics. We continue to see adoption of Prism Pro and AFS throughout our customer base. In Q3, a new customer, an American investment management firm with a focus on asset management, serving customers worldwide, signed a deal for more than $1 million that included AHV and AFS across 13 locations, all managed by Prism Pro. Another new customer from our APAC region, a data processing unit within an agricultural co-op in a large prefecture, will leverage AHV and AFS to integrate all the infrastructure within the prefecture onto our software platform.

As mentioned, Q3 saw continued adoption from our G2K customers, which accounted of 42 -- 43 of our top 110 deals in the quarter. One of our 9 deals worth more than $3 million in Q3 was a software deal with a U.S. brokerage firm that has lifetime bookings of more than $10 million. This G2K customer continues to expand its production VDI environment on our platform and has made purchases in every quarter since its initial buy in Q3 of 2017.

Another software deal with a G2K customer in Q3 was with a French multinational banking and financial services company. This customer has lifetime bookings of nearly $10 million in a Nutanix environment running a diverse set of enterprise application workloads across more than 22,000 virtual machines. This deal was one of our 47 deals worth more than $1 million in the quarter.

Now coming to talent, we hired Sankalp Saxena as Managing Director to lead our operations in India. Sankalp will oversee our India technology center and further the great contributions that the team has made to our product and business. Built by our India's sales leader, Sunil Mahale, the Indian customer base also continues to pleasantly surprise us. I call India the AHV country. Their extremely higher adoption of AHV convinces us that virtualization is still an under-penetrated market. And by making hypervisor invisible, we're growing the surface area of Enterprise Cloud. We have very high hopes for India, as it grows to invest in software automation, a Silicon Valley-like start-up culture and cloud. From 5 years ago, when we had nothing in India, to where we are today, we've come a long way in proving that a high-quality business and a corporate development strategy can be built by being geographically dispersed without compromising on culture. No wonder we've been recognized in 2018 by Glassdoor and Battery Ventures as one of the top public companies to work for in cloud computing.

In summary, Q3 was a great milestone in our journey towards our goal for $3 billion in software billings in fiscal 2021. There's much work to be done in the coming 12 to 18 months, both in products and our go-to-market motion. We hope you'll keep us honest.

With that, I'll hand things over to Duston, who will provide a more detailed breakdown of Q3. Duston?

Thank you, Dheeraj. Our Q3 mirrored the underlying trends of prior Q3s, and we were pleased with our results that reflected continued strong growth.

Revenue for the third quarter was $289 million, growing 41% from a year ago and up slightly from the previous quarter and significantly ahead of our guidance of $275 million to $280 million. In Q3, we targeted to eliminate $45 million of pass-through hardware revenue, and I'm pleased to report that we exceeded our plan and eliminated $52 million of pass-through hardware revenue during the quarter.

Software and support revenue was $227 million, up 55% from the year-ago quarter and up 9% from the prior quarter. On a revenue basis, the product mix was 78% software and support, and 22% pass-through hardware. We billed $351 million in the quarter, representing a 50% increase from the year-ago quarter and a 1% decrease from Q2. Once again, and consistent with last quarter, this billings performance far exceeded the current Street estimates of $332 million.

Software and support billings were $292 million, growing 67% from the year-ago quarter and up 6% from the prior quarter. On a billings basis, our product mix for Q3 was 83% software and support, and 17% pass-through hardware revenue. Notably, on a bookings basis, pass-through hardware decreased to less than 16% of total bookings. The bill-to-revenue ratio in the quarter was 1.21, in line with our previous estimate of approximately 1.2. We have -- we continue to have slightly higher revenue deferrals with our software-only deals, which has the impact of lowering our current quarter revenue, along with increasing our deferred revenue balance. As we execute more software-only deals, we expect the bill-to-revenue ratio to continue to elevate into Q4. New customer bookings represented 27% of total bookings. On a regional basis, EMEA continued its resurgence and turned in another very nice performance in Q3. In Q3, our software-related bookings from our international regions were 45% of total software and support bookings versus 37% in Q3 '17. Our Q3 deferred revenue increased by $62 million, up 62% from a year ago and up 13% from the previous quarter. Non-GAAP gross profit for the quarter was $198 million, growing 57% from the year-ago quarter and up 9% from Q2. Our non-GAAP gross margin for the quarter was 68.4% compared to 61.2% in the year-ago quarter, 63.5% in the prior quarter and to our guidance of 67% to 68%. This continued gross margin expansion is being driven by a transition to a software-defined business model.

Last quarter, I mentioned we had fallen behind in our hiring and that we would, and I quote here, "Have a full-court press on hiring in the second half of the fiscal year to try to make up for this headcount shortfall." Well, we executed this full-court press flawlessly and ended up hiring more new employees in Q3 than in any previous quarter, by a wide margin. During the quarter, we added over 60 new sales teams, which is critical to our planned growth for future periods. We were very pleased with our hiring performance in the quarter. And although not yet at our planned headcount, we did significantly exceed what we thought was possible when guiding Q3, including the addition of almost 85 employees from our 2 recent acquisitions, Netsil and Minjar. This strong hiring performance drove expenses to $232 million in Q3, exceeding the high end of our guidance by $12 million. Expenses grew 37% from the year-ago quarter, but at a significantly lower rate of growth and total billing -- than in total billings of 50%; software and support billings growth of 67%; and gross profit growth of 57%. Our outperformance in hiring combined with higher-than-anticipated payroll taxes from RSU vesting, contributed to almost all the operating expense overage as compared to our guidance. The higher payroll taxes are directly correlated to the Nutanix stock price appreciation that has occurred over the last several months. On a non-GAAP net loss -- our non-GAAP net loss was $35 million or a loss of $0.21 per basic share.

A few balance sheet highlights. We closed the quarter with cash and cash equivalents of $923 million. That's up from $918 million in Q2. DSOs on a straight average were 60 days compared to 58 last quarter. The weighted average DSO was 26 days in Q3. We generated $13 million of cash flow from operations in Q3, which was negatively impacted by $10 million of ESPP outflow. And we generated negative $1 million of free cash flow during the quarter, and this was also negatively impacted by the same $10 million of ESPP outflow.

Now to the guidance for the fourth quarter. Again, on a non-GAAP basis, we expect: Revenue between $295 million and $300 million, assuming the elimination of approximately $95 million in pass-through hardware revenue; gross margin between 73% and 74%; operating expenses between $250 million and $260 million; and a per share loss of approximately $0.20 to $0.22, using weighted average shares outstanding of 171 million.

Although somewhat fluid from quarter-to-quarter, in Q4, we are assuming a bill-to-revenue ratio of up to 1.25 versus the current Street estimates of approximately 1.2. Based on our revenue guidance and our expected bill-to-revenue ratio, we expect billings of at least $25 million to $30 million higher than current Street models. This would also indicate, based on assumed software and support billings only, a 50% growth rate in billings from the year-ago period.

It is also very important to note that in Q4, the change in the estimated bill-to-revenue ratio of 1.25 from the current Street estimate of 1.2 ends up resulting in approximately $12 million more of incremental deferred revenue that would have otherwise been reflected equally in revenue and gross profit. We continue to fuel strong growth (inaudible) in Q4. We expect operating expenses to increase between -- to between $250 million and $260 million, finally getting to a level assumed in our operating plan that we put in place at the beginning of the fiscal year. Even with this growth, on a year-over-year basis, expenses will grow significantly less than the growth rate of our software and support billings. This growth in expenses has been thoughtfully crafted and is closely correlated to the superior growth rates we are experiencing in our software and support billings. We will continue to accelerate our spending around existing products, new products and on new technology derived from M&A activity if we believe it will lead to notably improving our already current market leadership position.

If not for the projected increase in the bill-to-revenue ratio for the quarter, which results in additional deferred revenue, this expense growth would have been funded by higher revenue and an equal amount of gross profit dollars. Cash flow remains unchanged, as the increase in billings offsets any expense growth. And lastly, as you may recall, during our Investor Day in March, we talked a lot about the Rule of 40, with the Rule of 40 score basically summing up the annual revenue growth rate percentage and free cash flow as a percent of revenue. We also showed that very few software companies can consistently maintain a Rule of 40 score of more than 40%, and that companies that execute to this are rewarded with superior valuations. We also stated that we would continuously trade off higher or lower revenue growth with corresponding lower or higher free cash flow and profitability, while always targeting a Rule of 40 score of 40% or greater. Based on a rolling 4-quarter basis and using our software and support revenue as a proxy for total revenue, our Q4 guidance would suggest a Rule of 40 score approaching 50%, which we believe, puts us in the top 10% to 15% of all public software companies.

So in summary, in the third quarter, it was a strong one. We had good growth in both billings and revenue, saw a rapid gross margin expansion. We ramped hiring to support our growth plans, and we delivered on our stated milestones to shift to a software-defined business model.

With this execution, we are well-positioned to reach our goal of $3 billion in billings for fiscal 2021. And with that, operator, if you could open the call up for questions, that would be great. Thank you.

(Operator Instructions) Your first question comes from Matt Hedberg with RBC Capital Markets.

Maybe starting, Dheeraj, with the pending launch of Xi Cloud and the related services, which I think still sound to be absolutely on track from Analyst Day, how do customers think about operating in a true multi-cloud world, where -- and how quickly might we be able to realize true application and data mobility?

Great question, Matt. We, at .NEXT, did some really good touchdown with customers on how they feel about this and some great sessions, more than 900 attendees and people flowing out of the conference room. We were almost violating the Fire Code of the building, with so many attendees. There was immense interest in Xi, because we really built a cloud for the IT operator. Up until now, most of the cloud is being catered to developers and builders of new applications, but there's a ton of legacy to really migrate from an owned on-prem world to a rented off-prem world. And we're really putting a lot of attention to detail to those mundane things in storage and compute and networking and high availability, business continuity, disaster recovery, a lot of those things are the things that we actually are doing to make sure that people get a one cloud -- a one-click experience. And we're fairly close. I mean, as I told you earlier, the thing that we're really building is a bullet train between on-prem and off-prem, and that is a lot of work in terms of compute and storage and networking and security and identity, that all have to become hybrid. But you'll start to see those results show off in the coming couple of quarters. I mean, we talked about summer is when we actually launch this for our customers. And from there, you'll start to see us report about how many migrations, because this is really adjacent to our existing customers. They really like the fact that they can get this in a one-click experience.

That's great. And then maybe just one operational question. With the record hiring this quarter, and I think 60 new sales teams, what process are in place to ensure these teams scale appropriately? I'm curious, are you splitting territories in certain regions? Or are more of these in net new pads that did not have prior coverage?

Well, segmentation is a journey, not a destination. So we'll continue to segment. We are continuing to look at how some of our reps actually graduate from enterprise accounts to global accounts. Within global accounts, we'll continue to segment, because some of them might have 10 accounts, and over time, I think it's a great rite of passage to have our reps only focus on 2 accounts. But there's a lot of farming and hunting that has to go in parallel, so we continue to look at this as a pyramid. And a lot of experienced reps will actually go from commercial reps to being enterprise reps. And a lot of really good enterprise reps will become global account reps, actually. But there's a -- there's also a focus on channel as well, and that's the other area that we're focused on. As we segment a lot of our own sales force, how does the channel actually go and deal with the SMB and the lower end of the mid-market.

Your next question comes from Alex Kurtz with KeyBanc Capital Markets.

Dheeraj, just on the OEM portfolio and your hardware partners. How's the ramp been year-to-date? And taking maybe a sneak peek into fiscal '19, which one of the big OEMs do you think offer the most leverage and maybe getting international growth or maybe into new verticals that you guys really aren't today?

Thanks, Alex. So one thing that has really gone on well for us is, during the transition, we have not pissed off our OEM partners. That's very important. When you talk about change, especially of this kind, the last 3 quarters have been immense change in the business model, we've done a really good job of saying it's a win-win. It's not a zero-sum game. And many of our large customers are actually working with our OEM partners to provide them the hardware that they actually prefer, which -- the freedom that they actually need between Dell or Lenovo or HP or Cisco. All in all, I think we've seen really good partnership across the board. We just announced the XC Core partnership with Dell as well, where people can now buy software licenses on top of Dell hardware. Lenovo has done a tremendous job. It's probably the best quarter to date, that Duston can also attest to. But also, I think we're seeing a few things in Japan and the rest of Asia, including Germany. Fujitsu has actually been one of our recent partners. And IBM in the power architecture, this is not x86. We've just announced our AIX solution, so the fact that people can virtualize AIX environments has been seen really positively by IBM customers.

And then Duston, just a quick question about steady-state OpEx growth, as maybe we start to look to fiscal '19, maybe an early look on how we should be thinking about that.

Yes, I think Q1 -- again, we only guide to one quarter at a time, but we've got some other chunky things in Q1, so we'll have some reasonable expense growth there. But again, it's going to be highly correlated to our billings performance, specifically around software and support. And that's -- I think we've done a good job at balancing that. And the more comfortable we get with that growth rate and what's going on there, we'll continue to put up some reasonable expense growth to fund that growth. Q1 maybe a little bit more than the other quarters, but we'll have to see how that plays out, just from some planning things that we have going on in Q1 and our sales meetings and things like that.

Your next question comes from Aaron Rakers with Wells Fargo.

A couple of questions, if I can. So first of all, I wanted to understand exactly what you're saying about the billings relative to revenue trend. If I look at just the software plus services billings, and you're saying about 50% year-over-year growth, it would -- unless my math's wrong, it would seem that you're actually assuming that the billings relative to revenue for software and services alone, that actual ratio comes down a bit? So I'm just curious if -- why would that be, or is my math off?

Yes, I would probably have to take it off-line. I think your math is probably off a little bit there, maybe just picking up some numbers. I haven't dissected it into the individual pieces there. Clearly, it was 1.21 in Q3, and we see it on a total billings basis going up to roughly the 1.25, maybe, upwards of 1.25 there. So I'll have to go through the detail with you.

Okay, fair enough. And then when we look at your guidance around the gross margin line, you look at the delta between what you're guiding, 9%, it looks like, hardware pass-through, hardware as a percentage of billings, and there's been about a 4 percentage point delta between that relative to the percentage of revenue that you've driven from pass-through hardware. So if I do the math and assume about a 98% gross margin on software alone, and you assume a 60% gross margin on the services business, then it would appear that you're being rather conservative in the gross margin guidance for this quarter. So I guess I'm trying to understand what exactly are you assuming. Is it the services gross margin kind of goes back into that mid-50% range? And why would that be?

Yes -- no, we've always had that pool of other expenses there that doesn't easily get allocated to the software piece or the other...

We have professional services as well.

Yes, the professional services, which is in the support piece there. So that, we've talked about it in the past, is roughly $2.5 million a quarter plus. That's kind of a fixed amount there. To your point on the margins, we've been really happy with the progression on the margins, coming up this quarter. Obviously, we beat guidance a little bit, and we're excited about what we're doing, and we'll see how Q4 plays out here.

So to be clear, you're assuming a mid-50% services gross margin?

I think if you bundle those other fixed costs into there, that's what you'll probably come up with, yes. Because right now, it probably -- it has a home in the other bucket there. And ultimately, if there's nothing left, it has to get allocated somewhere.

Your next question comes from Katy Huberty with Morgan Stanley.

The acceleration in software and support growth to 67% in the quarter, is that entirely driven by faster node growth? Or is there also a dynamic of higher attach software that's playing into that trend?

Yes, I think -- thanks, Katy, for the question. I guess it's a combination of quite a few things, large deals being one. A year ago, we were in this business where we talked a lot about DRAM pricing and things of that nature, where hardware and movements in hardware pricing were actually affecting our overall ability to extract value of the software itself. So a big chunk of this is just the fact that we can now sell software at scale. And obviously, people are looking at Ultimate, which is one of the additions that we sell, which has quite a few features. We're selling Prism Pro now. At scale, we're selling Prism Pro. In fact, it's -- the other thing that you've seen is AFS, which is our fastest-growth product in the history of the company. So there's a few other things coming along, but the biggest factor, I would say, is the fact that we can decouple software from hardware and not worry so much about hardware pricing.

Yes, Katy, we've done, I think, as Dheeraj mentioned, a good job of maintaining value. We typically look at it on a software and support or what we call implied software value of software and support, ASP per node, and that held pretty strong in the quarter. So it's a pretty good determination there of us holding some pretty good value.

The other thing that we also did, Katy, in the last year, has been use of consumer-grade flash. That has actually really helped us in all-flash environments. We're been doing a lot of all-flash environments now. And that's happening because of a lot of the work that we did with consumer-grade hardware.

And can you remind us when some of the new products, like Xi, Calm, Beam, Flow, that will ship this year, when that can contribute meaningful revenue? Are we still a few years out from that?

So Flow is a GA product, and there's a lot of education. In fact, we're going down a path of saying you don't need VXLANs. That's actually a complete different thesis. Just like we talked about no SAN and no VTEPs, we talk about no VXLAN. And there's some education here, but we can definitely see some customers, many customers, large customers, who have been extremely affected by the complexity of VXLANs and the network virtualization stack of the last 3, 4, 5 years, we are seeing some really good traction in POCs. So I would say that meaningful revenue, we probably expect sometime in calendar 2019. Now exactly what that looks like will be a function of execution, not just product, because product is pretty much there in Flow, but a lot of work that we are doing, go-to-market itself, and saying there's a different way of doing network virtualization. Era, I think, again, is going to be GA by the end of this year. So it's early days. It's like 2011 for Nutanix, is what Era is right now. And Beam, the good thing is that it's a SaaS product and it's already GA-ed. So we'll learn our lessons about how do you really pay the channel, how do you pay the sales force on a subscription product itself. And I think there'll be some good lessons to come and talk about in the coming quarters.

And then just finally, where are you in integrating AI capabilities into the software? And do you think that those are features that you can charge for? Or are they just necessary to compete in the software market going forward?

Yes, it's a great question about AI. There's a lot of AI talk, just like when you think about the Gartner hyper cycle. There's a lot of AI software. There's a lot of AI in open source. There's a lot of AI coming from specialty models and algorithms coming from public cloud vendors. Our goal will be to make AI and ML accessible to customers, because there's this design principle called the MAYA design principle, and the acronym stands for most advanced, yet acceptable. And a lot of AI in the '90s failed the MAYA principle, just like RFIDs in 2000s and Docker a couple of years ago. So beyond the hyper cycle, there's a very good opportunity for us to really go deeper into making this consumer grade. One of the things that we talked about at .NEXT was Project Sherlock. And Sherlock's goal is to actually make AI more useful and accessible through one-click consumption. And how quickly can you build applications? Not just the fact that you can use GPU hardware or flash hardware, but really the fact that you can build applications? And that's one of the big things that we're focusing on with Project Sherlock.

Your next question comes from Rod Hall with Goldman Sachs.

I want to just start off, I go back to this comment you made, Dheeraj, about all-flash and sort of understand who you think you might be taking business from there. So could you just comment on who you think you're competing with there? Who may be losing business as a result of your progress there? And then I've got a follow-up.

Yes, thanks for the question. So on one side, there's definitely a lot of disruption going on with mid-range arrays, because a lot of them are up for refresh. Most of them were actually using spindles in the past. And on the other side, hyperconverged infrastructure itself, we've done a really good job of mixing and matching hybrid nodes with all-flash nodes. And that is a really hard problem to solve for because you have to really understand data locality and peering and metadata and a lot of the things that otherwise goes awry when you think about spindles and flash in the same cluster. So on one hand, we're becoming super competitive in heterogeneous clusters and hyperconverged infrastructure. On the other, we think that we're doing a pretty good job of selling flash to the application administrators, rather than going and selling to the storage guys, where it's actually a cents per gigabytes or dollars per gigabyte kind of conversation, like, what does it need to really make this application fast and be able to deploy this at scale.

Yes, it's just strange, because we've seen some flash vendors pivoting to like a DAS type of messaging. And we wondered if maybe you're putting some pressure on them from a competitive point of view.

I mean, definitely, look at developers and containers. I mean, at the end of the day, a lot of the next-generation DevOps and developers will want commodity servers and be able to really do things like the way they did with Hadoop and Spark and Apache, Kafka, and all those different kinds of DAS architectures, really need to come together and need to provide all the bells and whistles that you had in a 3-tier architecture in this collapsed commodity hardware architecture.

And then Duston, I just wanted to check, pass-through hardware expectations in fiscal Q4. Could you just maybe update us on what you're expecting, now that you've seen the -- got the quarter under your belt and a little bit of fiscal Q4 as well?

Yes, so nothing to do with where we are in Q4, but roughly, what we said we'll eliminate 95%. We had a target to get to roughly, I think, 9% for billings. That's always 3% to 4% higher for revenue. And as we see it now, we should be pretty close to that estimate that we put out 3 -- over 3, I guess, quarters ago. North America, as we mentioned, team did a great job. Basically, facilitizing all of North America. So that's done. Europe is well underway, with a lot of progress being made there. And now the focus is turned to APAC. So a lot of good stuff going on. And as we see it now, we should be pretty close to that number.

Your next question comes from Jayson Noland with Baird.

Dheeraj, I wanted to ask on this shared accelerated storage term kind of an umbrella phrase for NVMe. You're agnostic to the hardware underlay, but is this an opportunity for Nutanix? And does it change anything within the industry overall?

Absolutely. We talked about how close can storage get to the application. Because now with NVMe over fabric, we actually need to bring it even closer to the application, because at the end of the day, NVMe's going to be that much faster. And while we're talking about 40-gigabit networks and 100-gigabit networks, still the commodity is 10-gigabit. And only in the last 2 years have people moved away from 1-gigabit on-board network cards to 10-gigabit now. So I think it's a great opportunity. In fact, we can also virtualize the NVMe over fabric protocol to be hyperconverged, which is again, a very unique opportunity. The fact that you can have apps not going through NFS, which was the tax that we had to pay for hypervisors. In the AHV LAN, I think when you look at our hypervisor, there's some immense opportunities to go and optimize the NVMe over fabric kind of protocols, without really thinking about legacy hypervisors.

Okay. And then a follow-up on AHV. You said it's a significant part of some of your larger deals. Is that driven primarily by cost savings? Or is there more to it?

Well, I think we don't lead with cheap. And Nutanix is not that cheap a product. If anything, we get a lot of flak for being a premium product, but we lead with ease of use and the stuff works. And the fact that application folks can now use virtualization. That's the thing. I talked about, for example, India is a country that was under-virtualized. Now they can just use it without having to really become virtualization experts. Look at Splunk administrators and Oracle administrators and Citrix administrators. A lot of these folks, including Mode 2 applications, I mean, a lot of Mode 2 apps need the soft layer underneath to be able to make things highly available and replicable and backup and disaster recovery-proof and so on. So we're seeing a lot of adoption because of the fact that we're making things invisible rather than just making it cheap. Now budget is definitely a reason. I mean, people are like, "Wow, I don't have to pay for it anymore." But the real education on the fact that we should not have to pay for it comes from public cloud. People are like, "Well, I use the public cloud, and I don't really pay for a hypervisor. I don't have a team of people managing virtualization software," is the education that people are getting on-prem as well.

Your next question comes from Wamsi Mohan with Bank of America Merrill Lynch.

One for Duston, one for Dheeraj. Duston, the new customer acquisition pace has slowed down to about 4% year-on-year and that's been decelerating. But you also noted, like, increasing traction over the last several quarters on larger deals and even including software-only. So has there been a change in focus towards larger deals versus new customer footprint? And your ramp in sales teams that you alluded to in this quarter, how should we expect that to trend going into next quarter as well? And I have a follow-up for Dheeraj.

Sure. Now if anything, we're focusing on a lot of things, and we've actually had a renewed focus with the channel on new customer logos, actually. We've got a rebate program in place now, and...

For the first time in the history of the...

For the first time in the history of the company, and we're really excited, actually -- now this has to obviously get into actual bookings -- but really excited about what's happening in the channel with the pipeline for new logos and things like that. So there's been, if anything, an acceleration of focus there, from a new customer. And the Q3 new logos usually come down. We had a really good Q2 in a lot of different ways, and...

Unlike in 2016 -- 2017, when Q2 was...

Yes, we had an opposite issue in -- a year ago, but we had a really good quarter in Q2, which brings up a good point, guys. We've internally now, and maybe you want to start doing this also, is that we're starting to analyze the business, really, on a rolling 2-quarters basis, because it's a lot more insightful, because I think potentially in this phase, maybe we have Q2s and Q4s, our January and July quarter, which also happens to be the end of a commission -- 6-month commission period for us, maybe be a little stronger. So you can't really look at a strong Q2 and a strong Q4 without looking at a potentially slower Q1 or slower Q3. Or conversely, you can't do the opposite. So we've started actually analyzing a lot of things on a rolling 6-month or 2-quarter basis, which, as I say, is pretty insightful for the business.

Also, we're comfortable with our long-term average that we want every quarter, because the Rule -- I mean, the Investor Day presentation, we talked about the average customers that we want every quarter.

Okay, that's helpful. And then Dheeraj, how are you thinking about the evolution of the product portfolio? You've been adding assets like Netsil, Minjar to make your offering even more comprehensive. And you alluded to, sort of, very, I guess, quite extreme focus internally on areas like replication and one-click cloud deployment, all the things you brought up. How far are you from what you view as sort of a comprehensive enough complete portfolio towards your cloud OS destination? Like how far are we from the asset base that would get you to your end state?

It's a great question, Wamsi. And definitely a lot of wondering that I need to do for this answer. So I was talking to somebody yesterday who joined Microsoft in 1991, and I was trying to hire him. And he said, "You know what, Microsoft was as big in '91 as you guys are today." And this is a timeless journey. What we're really doing is like the Microsoft or Oracle of 1991. To build an operating system is never going to end. I mean, even today, Windows is hustling with Windows Azure. That's a new consumption model of a hypervisor and operating system. And this, Windows was [seeded] back in 1990, 1991. So I think we're on a timeless path here. We'll go and do many, many things in layers above us, like PaaS. I mean, I talked about enterprise PaaS. We'll double down on enterprise PaaS. We are great open source leveragers. We leverage a lot of open source to build things that are commercial-grade that are enterprise-grade. So expect us to do many things where open source becomes a big part of our leverage.

Your next question comes from Jason Ader with William Blair.

Two quick ones, maybe first for Dheeraj. As you've shifted to an all-software model, what has surprised you the most, I guess, both positively and negatively?

Great question, Jason. Thanks for the question. I think the biggest positive thing that I see is that our front office actually likes it, and our customers like it as well. Because anytime you do a transition of a business model, and you have to do it really methodically, it's like at 35,000 feet, how do we change the wings of a plane, is kind of the thing that we went through in the last 9 months. But it was done without a blip. And that's what I feel really good about, that our front office embraces it. There's not a whole lot of, like, what did we just do? And we're seeing some large deals, and I mean, in terms of what we see on the horizon. And that probably would not have been possible had we continued to sell appliances. And the big reason is because people would not want to buy all the hardware upfront. Now that's not to say that we don't see a lot of customers saying, "Please do not discontinue NX," because they love the single throat to choke, in a single hardware plus software kind of integration. So I think what pleases me is that we're able to actually have our cake and eat it, too. Your question about what is the con, like what do we not like about it? I think it's the design piece of it. We have to really go in -- I mean, and this is not about the last 9 months, but the next 12 to 18 months, we really have to think about a hybrid license model. And it's not that easy, because for the first time, we'll have to expose what it means to deal with on-prem and off-prem, that nobody in the world has ever done before. What do the accounting standards and rev rec issues look like? There's a lot of pain, and we'll probably be a great company if we come out of it in a way that is not reminding people of the vRAM licensing that VMware did. Every time we do something in licensing, a lot of our ex-VMware employees are like, "Remember what VMware did vRAM? Don't ever do that." So I think that's the thing that we have to be watchful for.

And do you think the software flexibility has actually helped the business, just from the standpoint of now customers don't necessarily need to rip and replace, they could use existing hardware?

There's a lot of requests coming for what does it mean to do this on NEC and Fujitsu and this and that. So definitely there's more requests coming. And we are seeing the TAM grow, because now -- kind of you're meeting in the middle with the customer. They're like, "We love your software, but can you do this with my favorite hardware vendor?" and so on. So I think it's definitely helped from that point of view.

Okay. And then just one quick one for Duston. Duston, are you willing to give us sort of a ballpark breakeven level from a revenue standpoint?

No, no. I mean, again, that's kind of the whole purpose of when we talk about the Rule of 40 and growth and trading off growth with free cash flow and profits and the likes there. So we're fueling a lot of growth, and we see a lot of growth ahead, but, as I say, we've done a pretty good job balancing in the past and still growing pretty rapidly. And I think you can assume we'll continue to do that balance going forward, but I'm not prepared to give you a revenue breakeven number.

Your next question comes from Simon Leopold with Raymond James.

First, I wanted to ask a -- more of a modeling-type question and then more of a competitive environment question. On the modeling, you offered us the metric of the ratio of 1.25. And I'm also looking at trends in DSO. And really, my objective here is to try to think more precisely about the cash generation. So over time, how do you think that ratio of 1.25 plays out over the next 1 to 2 years? And how should we think about DSOs in general for the company?

Okay, 1 to 2 years is a long time. We're learning a few things along the way here. I would, quite honestly, we'll have to see how it plays out. I would be surprised if we saw something higher than 1.3, because sooner or later, you have enough of other stuff coming the other way off the balance sheet, too, that offsets deferrals. We've got a couple of other things we're working on there. So I think it's too early to tell, but I think anything above 1.3 would be a little surprising. But again, 1 to 2 years is quite a ways out there. And then DSOs, I don't think you should see much movement in the DSOs. We always look at it on the weighted average, and you can't do much better from that perspective. I don't see really anything getting materially worse there. We've done a pretty good job, maybe a little customer mix, maybe a little geography mix. But I don't think you should see a material change in DSOs. It was 2 days from last quarter's -- 58 last quarter and 60 in Q3, on a raw calculation perspective. So I think that stays relatively consistent quarter-to-quarter.

Also, I mean...

Yes, I know that's (inaudible) model. My other question was if we can get updates on the competitive environment, particularly with the shift to the software model and more partnerships. And we've got companies like NetApp talking about developing their position in hyperconvergence, although we haven't seen much evidence of that in our checks. I'd love to hear your perspective on where the biggest competitors and new entrants are playing against you.

Yes, thanks for the question. I also want to say, as a sequel to the last answer that Duston gave, that we're in transition, so until the software thing is completely flushed out and becomes the 9% thing, until you see that stability, it's hard to also predict some of these things with [billings].

Yes, very hard. Just on that point, too, just to make sure, the bill-to-revenue ratio has nothing to do with cash flow. So wherever that goes, if you bill, you bill. So nothing there on cash.

And I think once it's stabilized, if Xi has done a good job, we'll start talking about subscription.

Right, yes. So we'll talk about something different in the future here.

Especially if we've done a good job with the product. On NetApp, obviously, they've built a very good business. We respect them for their channel loyalty, for example, learning a lot from a company like that. They're fully focused on storage. And in fact, they talk about their cloud positioning and so on, and it's basically about storage. We are a compute, storage, networking, security and application mobility company. It's an entire operating system. And some of the things that some of our hardware competitors are assuming, they're assuming -- like, NetApp is assuming that the legacy enterprise apps are already there in the public cloud hyperscale setup. That was the second-order problem, to go and solve for filer, for a legacy app that has already moved. We're working on the first-order problem: How do you get the app there in the first place? And by apps, we mean in their entirety. It's a much bigger problem with way bigger rewards down the road, actually. In terms of real competition, I think there's a lot of legacy to go and displace. The growing brand helps. We expect a lot from the channel in the coming months and years. And we also know that the higher up we go in the Global 2000, the more the brand will trickle down to the mid-market as well. There's a -- across the board, there's a lot of hustle that we're actually doing, and that's just one part of the go-to-market. As we talked about different hardware platforms, different geographies, I think there's a lot of that work that we continue to do. And we don't underestimate the competition. But at the same time, we don't overly worry about them. That's one thing that's been really good about the company, is that we do not lose focus on the competition -- by focusing on the competition actually. The thing that we always see in every .NEXT user conference is just the customer love and it's one of those things that you would not see reflected in the P&L. If you haven't visited a user conference, I would love for you to actually come and see from directly from the customers what they see about us.

Our last question at this time comes from Andrew Nowinski with Piper Jaffray.

So I just have maybe a follow-up on your new customer growth. So I understand the quarterly seasonality with Q1 and Q3 being the weakest. But if we go back to your Analyst Day, I think $1.8 billion of that $3 billion target was expected to come from your non-Global 2000 customers, implying you had to add about 1,000 per quarter. So I guess while the transition to software-only certainly has led to larger deal sizes, are you concerned that the shift may have negatively impacted your smaller customer growth? And are you still comfortable with that $3 billion target?

Yes, I mean, Duston can answer this, but [so can I.]

Yes, I mean, that's a 3.5 -- let's give it some time here, Andy, right? This is one quarter or a couple -- a month or so after the Analyst Day.

And it's an average...

And it's an average over a 3.5-year period. We're doing some really, as I said, cool stuff in the channel with new logos. I don't think anything has suffered because of success in larger deals and things like that, no. So we're comfortable. That's going to bounce around. Let's see what happens in Q4, and then we'll have a discussion.

Well, I guess I was more focused on the software transition element. Has that had any impact on large customer versus small customer growth?

We don't think so.

No, I don't think it's impacted. Again, the customer can have anything they want. If they want an appliance, they're going to get an appliance. So there's no reason why that should impact smaller customers or smaller deals.

All right. And then the percentage of nodes that are managed by Acropolis continues to increase. I was just -- which is great, but that could be a factor of just existing customer footprint continuing to expand. So I was wondering if you're seeing any sort of change in the adoption of Acropolis by your new customers?

That's a data that I don't have at the back of my pocket, but we can actually....

Yes, we can follow up with you, Andy. Yes, we don't track it quite like that, but we can see if we can dig in and find it for you.

That was our last question. Thank you for joining today's conference call. This concludes today's call. You may now disconnect.

Thanks, everybody.

Thank you.