Good afternoon, and welcome to the Equinix conference call. [Operator Instructions] Also, today's conference is being recorded. If anyone has objections, please disconnect at this time. I would like to turn the call over to Katrina Rymill, Vice President of Investor Relations. You may begin.

Good afternoon, and welcome to today's conference call. Before we get started, I would like to remind everyone that some of the statements that we'll be making today are forward looking in nature and involve risks and uncertainties. Actual results may vary significantly from those statements and may be affected by the risks we identified in today's press release and those identified in our filings with the SEC, including our most recent Form 10-K filed on February 28, 2014, and Form 10-Q filed on August 8, 2014. Equinix assumes no obligation and does not intend to update or comment on forward-looking statements made on this call. In addition, in light of Regulation Fair Disclosure, it's Equinix's policy not to comment on its financial guidance during the quarter unless it is done through an explicit public disclosure. In addition, we will provide non-GAAP measures on today's conference call. We provide a reconciliation of those measures to the most directly comparable GAAP measures and a list of the reasons why the company uses these measures in today's press release on the Equinix Investor Relations page at www.equinix.com. We'd also like to remind you that we post important information about Equinix on the Investor Relations page of our website. We encourage you to check our website regularly for the most current available information. With us today are Steve Smith, Equinix's CEO and President; Keith Taylor, Chief Financial Officer; and Charles Meyers, Chief Operating Officer. Following our prepared remarks, we'll be taking questions from sell-side analysts. [Operator Instructions] At this time, I'll turn the call over to Steve.

Okay. Thank you, Katrina, and good afternoon, and welcome to our third quarter earnings call. In Q3, we delivered both revenue and adjusted EBITDA above the top end of our guidance ranges despite significant currency headwinds as the global demand for Platform Equinix continues to drive our business. As depicted on Slide 3, revenues were $620.4 million, up 3% quarter-over-quarter and up 14% over the same quarter last year. Adjusted EBITDA was $283.9 million for the quarter, up 3% over the prior quarter and up 14% year-over-year. We delivered near-record bookings fueled by solid performance in our core verticals, global expansion with key customers and accelerated momentum in the development of our cloud ecosystem. The strength of our ecosystem strategy continues to manifest itself in very healthy fundamentals, including firm MRR per cabinet, lower churn, rapid interconnection growth and healthy operating margins. Interconnection continues to be a critical differentiator and source of sustaining value for Equinix. Interconnection revenue outpaced overall revenue, growing 17% year-over-year. And the robust growth of both our established and emerging ecosystems generated a record 5,700 cross-connect adds, 36% higher than our previous record. Importantly, secure private access between cloud consumers and cloud service providers is now the fastest-growing category of interconnection at Equinix. Secular trends driving this cross-connect growth are also driving similar momentum in our leading digital exchanges. This quarter, we added 143 ports, 3x the net adds we saw on the same period last year. Strong port growth on our Internet exchanges has translated into 27% year-over-year growth in traffic to 2.8 terabits across 2,500 ports, with network content and cloud customers driving this increase. Our global platform also remains a unique differentiator and is a key driver of our bookings momentum and revenue growth. Today, 67% of recurring revenues come from customers deployed…

Thanks, Steve, and good afternoon to everyone. In the third quarter, we saw strong performance across all regions with near-record growth in net bookings. Our bookings activity produced record billable cabinet adds of approximately 3,400, 70% above the average 4-quarter trend. We added a phenomenal 5,700 cross connects and 143 exchange ports in the quarter. This clearly highlights the benefits of our current strategy. Also, as demonstrated by our financial results and the strength of many of our key operating metrics, our ecosystem effect not only increased our revenues but preserved the attractive yields we enjoy on a per cabinet basis. As Steve outlined, capturing the cloud and enterprise opportunity is the next phase of growth for Equinix, and we're in a unique position to lead and benefit from these market changes. To accomplish this, we need to make some investments. We need to target new and existing customers. We have to determine how to deploy our capital for new product initiatives such as Cloud Exchange and how do we augment our go-to-market efforts. We'll be making these investments alongside our key system initiatives being Equinix Customer One, which was rolled out in Asia Pacific earlier this month, in the Americas and EMEA earlier next year and our financial systems conversion to support the REIT compliance effort. So now let me move to Slide 4 from the presentation posted today. Global Q3 revenues increased to $620.4 million, a 3% increase over the prior quarter and up 14% over the same quarter last year. Our overperformance was due to higher gross bookings, continued custom sales order activity and lower-than-planned churn. For Q3, revenue performance reflects a $2.3 million negative currency impact when compared to the average rates used last quarter and a $3.7 million negative currency impact when compared to our…

Okay. Thanks, Keith. Let me now cover our outlook for 2014 on Slide 13. For the fourth quarter of 2014, we expect revenues to be in the range of $627 million to $631 million, which absorbs $11 million of negative foreign currency impact compared to our prior guidance rates. Cash gross margins are expected to approximate 68% to 69%. Cash SG&A expenses are expected to approximate $139 million. Adjusted EBITDA is expected to be between $291 million and $295 million, which includes $6 million in costs related to the REIT conversion and absorbs a negative foreign currency impact of $5 million compared to prior guidance rates. Capital expenditures are expected to be $210 million to $230 million, which includes approximately $35 million of recurring capital expenditures. For the full year of 2014, we are raising revenues to now range between $2.433 billion and $2.437 billion or a 13% year-over-year growth rate. This absorbs $15 million of negative foreign currency impact compared to prior guidance rates. Excluding the negative impact of foreign currency, the revised revenues range reflects a $20 million increase compared to our prior midpoint guidance. Full year cash gross margins are expected to range between 68% and 69%. Cash SG&A expenses are expected to approximate $553 million. We are raising adjusted EBITDA to now range between $1.110 billion and $1.114 billion. This revised guidance includes $32 million in costs related to our REIT conversion efforts and absorbs $8 million of negative foreign currency impact compared to prior guidance rates. Excluding the negative impact of foreign currency, the revised adjusted EBITDA range reflects a $10 million increase compared to our prior midpoint guidance. We expect 2014 capital expenditures to range between $630 million and $650 million, which includes approximately $110 million of recurring capital expenditures. So in closing, we are executing our strategy of driving differentiated growth, reflected in the strong performance of our core ecosystems and record interconnection activity. We believe Equinix is the best location to access the variety of cloud services critical to managing enterprise workloads. And we continue to close strategic deals driven by the strength of new offers like Performance Hub and Cloud Exchange. Going forward, we remain focused on leveraging the advantages of our global footprint, driving both network and cloud services density and distancing ourselves from the competition through innovative products and investments in Platform Equinix. So let me stop here and open it up for questions. I'll turn it back over to you, Rachel.

Our first question comes from the line of Mr. David Barden of Bank of America.

I guess a couple, if I could. Just first, Keith, I think you said you were going to maybe kind of look opportunistically at CapEx -- or capital structure and balance sheet opportunities in the fourth quarter. I believe the original intention for the stock buyback program was to complete it in 2014. I was wondering if you could kind of comment on whether that's still the intention. The second thing was just getting some questions on your assumptions that the euro and pound foreign exchange rates relative to your -- relative to spot rates are going to improve in the fourth quarter. I'm assuming that has something to do with your hedging game plan, but that'll be good to know. And then the last thing, again maybe for Steve or Keith. I think in Keith's comments, he referenced something about investments that need to be made to attack the cloud opportunity. And I was wondering if that was trying to foreshadow some kind of CapEx or operational type of expense that might be -- we need to start thinking about for 2015.

Okay, David. So let me take the first couple, and then Charles or Steve will take the last one. I think when we look at capital structure, it's clear to me as we continue to move towards conversion to a REIT and recognizing where we may spend our capital dollars, all else being equal, given where the market is, there's still an opportune time to raise capital relatively inexpensively. And so one of the things that we've said before, to the extent that we do raise capital, particularly given that we've taken out the 2014 conversion over hopefully not too long distance, we'll take out the 2016 converts, we'll have the capacity to put more debt on the balance sheet and use it effectively. So that would be one thing that we will continue to look at. There's no commitments yet other than we are actively looking at the opportunities that are present in the marketplace. Also, I'd tell you that given the size of our revolving line of credit, that's another area that you expect us to change as we exit the year and start operating as a REIT effective 1/1/15 so that would be one other area that we would actually look at. As it relates to our share repurchase program, we have roughly 150 -- sorry, $150 million left on the program. And clearly, we will use the cash opportunistically, taking into consideration some of the other decisions that we're going to make about how to spend our capital. So clearly, that is what's spent, roughly $350 million to date. We'll continue to look at that opportunistically. As it relates to FX, a couple of things. Number one, when we look at the FX rates, particularly on the forward guidance relative to spot rate, you're absolutely right.…

Yes, and David, this is Charles. I'll take the last piece relative to some of the investments. As the results demonstrate, we continue to see very significant momentum in certain building the cloud ecosystem, both on the service provider side as well as sort of significant early traction on the enterprise or the buy side of the ecosystem. And we intend to continue to invest behind that momentum given that we see the cloud overall as really quite a transformational opportunity and growth opportunity for the company. So there's probably several categories in this of investment that we would be looking to incorporate into our 2015 plan and through the remainder of this year. One, on the product side, really building on the momentum of our Performance Hub and our Cloud Exchange offers to continue to add feature functions to those and make sure that those are the leading platforms from a cloud perspective in the industry, which we're very confident is the case. Secondly, on go-to-market, we're going to continue to invest in our solution architect team, which has really been critical to ensuring customers can leverage those offers to their benefit as they implement their hybrid cloud architectures as well as our -- we'll probably selectively augment the direct force in response to continued strong demand and then also continue to invest in our channel program, in our professional services capability, both of which are really designed around: one, completing whole offers by sort of adding third-party value-add to our core offers and then expanding our reach from a distribution perspective as we -- particularly as we tackle the enterprise market. And then finally, sort of targeted demand generation, again probably focused on the enterprise and really building off of our lighthouse wins, several of which we talked about in the script, and in fact, have talked about a handful of them in every script for the last several quarters. And now what we're beginning to see is a broader -- people sort of looking to those lighthouse wins and saying, "Hey, if they're doing it, we ought to be considering it," and people who are implementing hybrid cloud and using Performance Hub and Cloud Exchange to do that. So we'll probably be putting money behind increased awareness and demand generation activity to build off of that momentum.

Our next question comes from the line of Mr. Jonathan Schildkraut from Evercore.

I have a question about Open IX. A lot of the folks that I speak to, including myself, have been a little bit concerned about the development of Open IX in the U.S. And one of the things that Open IX talks about is kind of connecting its locations with a fiber ring and sort of extending its density that way, which is very much, as I understand it, the way that the London Internet Exchange works. Now you guys have been very successful across Europe, in London in particular, in terms of winning key customers. And maybe if you could give us some insight into how you compete there and how you differentiate there while sitting on that ring to give us a little bit more confidence in how you guys will defend against any sort of perceived move here in the U.S.

Sure. Jonathan, this is Steve. I'll start and I'm sure Charles can probably supplement here. A couple of data points. First of all, after 16-plus years, we have a critical mass, as you all know, of cross connects and ports on our switch fabrics and a couple of switch fabrics now that we're deploying, one with the Internet switch fabric that's been around for a long time and more recently our cloud switch fabrics. So if you just purely look at the results that Keith mentioned in this quarter, just in the Americas, where the Open IX is -- I think the heart of your question is before we go to Europe. Just to remind you guys, we added 2,600 net cross connects just in Q3 and over 101 ports in the Americas alone. And as Keith mentioned, that -- over the last 4 quarters, we've added more ports in the U.S. than we have the prior 3 years cumulatively. So that tells you something that the value that we're delivering and the leadership position and the critical mass that we've accumulated is still advantaging us. So there's no discernible impact to Equinix really at this point. Like any competition, we are tracking it closely in the U.S. We'll continue to monitor these alternative exchanges. But our real focus is on the next generation of interconnection, advantaging our leadership position and continue to deliver a superior peering model to our customers. And I think the scale and advantage we have there is proving out in the metrics. I don't know, Charles, if you'll add anything. Maybe...

Well, all I'd say is I think although the industry structure is slightly different in Europe than it is in the U.S. and Open IX is, I guess to some degree, attempting to shift some of that industry structure and introduce a dynamic that looks more like the European model, I would argue that the fundamental basis of competition on the exchange market -- or in the exchange market isn't meaningfully different across its various regions. At the end of the day, the customer is making a fundamental decision about how effectively they can peer their traffic. And they're looking at a range of options that goes from transit to private peering -- I'm sorry, to public peer to private peering. And the reality is the most effective way to do that, typically if you have -- particularly if you have large volumes of traffic, is to ensure that you can peer off traffic to the right counterparties most effectively. And when you look at the results that -- I mean, that Steve has articulated in terms of our port growth, it's very clear that customers are voting with their wallet relative to how they can do that most effectively and they're doing with Equinix. So they can move from: one, accessing transit partners; two, then peering their exchange -- their traffic over a public exchange inside the facility. And then quite importantly, when they get traffic between a counterparty to a certain level, they want to pull that off in private period over cross connect. And them being able to do that inside of an Equinix facility where we have a very high critical mass of participants is just much more effective for them. And so in the end, we feel very good about the product. We continue to invest in its future functionality. And I think if you look at our momentum compared to what limited growth and trajectory that we're seeing from Open IX participants, I think we feel very good about where we are.

The only thing I'd add, Jonathan, is as Charles said, the traction in the U.S. is negligible. But in Europe, if you listen to our team in Europe, those 3 firms are doing well. And actually, they're growing in our facilities and we're benefiting from that. So we're -- they're getting the port growth and we're getting the colocation growth. So if you talk to Eric and his team in Europe, he's still growing on the back of the relationship that we have with all 3 of those in Europe. So the dynamics are a little different between the 2 regions. We're competing in Americas. We're working together in Europe.

Our next question comes from the line of Mr. Amir Rozwadowski of Barclays.

I just wanted to touch on a bit of a bigger picture question here around the enterprise arena. I was wondering if you could give us a bit more color in terms of the type of traction you folks are seeing here overall in the enterprise. Clearly, it seems as though cloud-based opportunity seemed to be reaching a tipping point for you folks. But I was hoping that you could see how -- or at least get a little bit more insight on how we should think about when we should start to begin to see similar momentum on the enterprise front. Clearly, you guys have a number of partnerships, most recently with Cisco, as you mentioned on the prepared remarks. I would love to hear how these partnerships have been resonating with some of your potential customers.

Why don't I start, Charles, and let's double team this again. I think there's a lot going on here in the enterprise. It's a good question, Amir. First of all, if you talk to any CIO or the head of applications or head of infrastructure in any company -- any enterprise company today, they're going to tell you that they're faced with a variety of disruptive forces going on. One, they've got congested network issues. They have performance and security limitations that they're challenged with. And then more recently, their user experience is becoming more social and more mobile. So most of the clients we're dealing with are pushing to rearchitect their IT to be able to access the hybrid cloud. So they're all studying. Some are further ahead than others. And they're doing that in a distributed hub architecture. So that's why you hear us, and that's why we develop a network Performance Hub, which is to address this distributed hub architecture that they're going to want to get to, to get to the -- to ultimately get to the hybrid cloud. So they're reinventing their wide area networks, their local area networks. Their workplaces are becoming virtual. Their hybrid cloud is the choice of the future. That's been validated pretty much anywhere you look or talk to. So there's a lot going on there. Now we have to get use cases built and we're doing that. You hear about that every quarter. Our approach to this is to get the cloud access nodes, the cloud service nodes, get our data centers as populated as possible with these access nodes so that enterprises can look inside of these data centers and see the cloud density and network density they need to put private workload in there to enable the hybrid cloud. They can connect private to public cloud access points and create the hybrid cloud. So there's a lot going on in the marketing function of the company to build these use cases. And as you get to an industry vertical and you get a use case started, then the rest of vertical will follow suit. So we have a lot of that going on today.

Yes. I mean, just maybe a couple of things. I think that again, we're seeing very strong momentum on the service provider side of the cloud. We are getting these sort of key lighthouse wins on the enterprise side that leverage our Performance Hub offering. And interestingly actually, the momentum on the enterprise is across a range of enterprise types. We talk about enterprise distinctly from cloud, content, digital media, financial services, et cetera. But in reality, all of those companies are enterprises and they have very distinct enterprise needs. And what we've been able to do is actually to leverage our relationships with them and leverage the fact that they already use us for many of their revenue-facing applications to meet their enterprise requirements. And that is their ability to deliver distributed applications with high performance, get anything anywhere to all of their users in the enterprise, moving to a more distributed architecture that allows for the high-performance application delivery. And so those lighthouse wins are really starting to open the eyes of the broader enterprise community. And I think when we're -- so we see that momentum. And I think when we're really going to start to see it at a new level is as we ramp up our go-to-market capabilities. We have relatively small direct enterprise selling force today. We're selling across all -- our enterprise offers across all our verticals. But as we ramp the professional services in the channel, that's when I really think we're going to begin to see it contribute more meaningfully. And we've been prudent about what pace that's going to be in our planning, but we think we're excited about the investments we're making in 2015 and believe that we'll begin to see those very positively impact our numbers.

Our next question comes from the line of Mr. Jonathan Atkin of RBC Capital Markets.

Yes. So I have a question around the interconnect. The networks are very strong, and Keith sounded very optimistic about the outlook going forward. And I just wondered how to think about that in terms of translating to gross margins starting to ramp and even margins starting to ramp. They've been fairly flat in most regions except for Asia Pac. So that's kind of my financial question. And then I wondered, we've been kind of talking around this in terms of sales headcount and solutions architect, but can you kind of quantify where you are now and what your targets are for sales force and solutions architecture?

Good question, Jonathan. So a couple of things. Number one, yes, we are quite optimistic about what we're doing with interconnection and how we see that continue to scale, not only within the Americas but across the broader platform. So I think the benefit that you're going to see is interconnection revenues, there was roughly a 50 basis point improvement on a total company basis in interconnection revenue as a percent of our total recurring revenue this quarter. All 3 regions were up. The company therefore as a whole was up, and that's, obviously, a net positive. Equally, from a positive perspective, we are continuing to increase, if you will, the product density in some of our deployments. And both of those are driving, if you will, more profit into the system. The offsetting part of that, of course, is our product mix and also some of our IBX. And as we alluded to, we're also being very thoughtful about some of these larger cloud-based strategic footprints that we're taking. And so what offsets some of that -- some of the successes, of course, is some of those decisions we make. As Charles alluded to, I think we're being thoughtful and very prudent about the decisions we make, recognizing overall, we believe that cash gross margins are still -- we're still driving for a 70% or better cash gross margin line. And we're still driving towards a 50% or greater EBITDA line. All that said, we also want to make investments. Again, when you look at the size of the opportunity that we see out there, and again, Steve and Charles alluded to it, in a way that when you think about trying to capture that enterprise and continue to win the cloud business that exists in the marketplace, you got to scale the business. And one of the ways that we're going to scale is, of course, adapting our go-to-market strategy, looking at the channels, looking at the solution architects, looking at how we create more demand gen, investing our technology and our technology platforms behind Ihab and Brian Lillie. So for all those reasons, I would tell you that we're making great headroom -- headway and continuing to drive leverage in the business and scale and profitability. But we're taking some of that value and we're putting it into the future of our business because we think that's going to be really important as we look towards us being a much larger company than we are today.

Sales headcount, Charles?

Yes. So sales headcount, we are right now in the sort of 225-ish range globally for quota-bearing headcount. We had talked about 250 as a range that we could easily see getting to. I think we're doing that quite selectively in terms of where we see the demand there in markets or verticals that we -- where we want to make additions. So I think I can see us getting over a period of a few quarters up to that 250 type range. And then on the solution architect side, we're looking at probably adding another dozen or 2 of those over the course of the next several quarters just given the success we've seen in terms of them being able to really help our customers identify how Platform Equinix can fit in to most typically their sort of hybrid cloud architecture strategy. And so those are sort of the level of investments we're making. But importantly, I think also of note is that with the way we're going to get leverage in the long term is by also spooling up partners that can provide some of the professional service solution architect type capabilities as well as some of the sales reach from a distribution perspective. So we're looking at probably investing in the indirect side of our business from a channel program standpoint to the tune of 50 or more people in 2015. And obviously, we'll ramp that up in a disciplined fashion. But we can -- that's a big investment, a big focus area for '15.

Okay. And then just real quick on the channel program, as that does start to ramp, how do we think about the margin impact on that? Does it affect either way if the channel partners do account for a bigger portion of your incremental sales?

Not really. The cost of sale won't be meaningfully different, I don't think, for us in terms of the indirect. It might be, particularly in the early days as we probably look to invest in the level of channel harmony and making sure that we have a good collaboration between our direct force and the channel out there. You may have a slightly higher cost of sale. But I also think that we're going to be able to, particularly with some of those channels targeting certain enterprise segments with our Performance Hub and Cloud Exchange offers, which we believe have a very strong value proposition. I think we'll be able to sustain kind of the price points margins that we're accustomed to. So I don't see that as meaningfully different, and I certainly don't see it in terms of having any kind of meaningful negative impact on our overall margin structure.

Our next question comes from the line of Mr. Colby Synesael of Cowen and Company.

I have 2. First off, as it relates to CapEx, when I look at your buildout in terms of cabinets the last few years, it's been fairly consistent. And also, the cost on a per cabinet basis is somewhere around $50,000. Is there any reason to think that as we go into 2015, the number of cabinets you tend to build out and also the cost of those cabinets would change very much? And then also, as it relates to the cloud business, you talked about adding some larger cloud deployments, magnet customers as you referred to them. What do you anticipate the overall impact of those deals being on MRR per cabinet, particularly in the United States or the Americas as we go into 2015?

Yes, Colby, Keith and I can tag team this a bit. I think in terms of capital expansion, I think we're going to -- we're looking with some granularity as to where we need to invest given the demand. I don't think there would be a dramatic difference in terms of the pace of buildout. I do think we're looking at whether or not some of the large footprint opportunity that we are exploring and seeing in the cloud space will require us to make some incremental investment there. But that's something I think that we'll probably give you little more clarity around come our guidance in February. As to the cost per cabinet, we are constantly striving to sort of drive our cost per cabinet down while at the same time ensuring that we're delivering the premium quality service delivery that our customers require given the premium nature of their applications. So again, we're trying to drive it down to some degree from a continuous improvement standpoint. But I think the number that you articulate there, which is on the order of $50,000 a cab, is probably a reasonable planning number. And again, we'll probably continue to chip away at that, but that's probably not going to -- I wouldn't say change dramatically. And then lastly, as it relates to these big cloud footprints, honestly, we have been pursuing those as part of this sort of disciplined cloud ecosystem strategy for several quarters. And despite that, as you can see in our results, we've continued to maintain very attractive levels of yield. And that is because there are a number of levers available to us in terms of: one, driving interconnection growth, which as you can see, was very healthy this quarter; and continuing to manage power densities in the facilities with some rigor and some care that allows us to really sort of offset some of -- if we are taking on a cloud footprint that may not ramp from interconnection standpoint immediately or it will take some time to for an ecosystem effect to build to up around that, honestly, we continue to have levers that we believe will allow us to sort of maintain stability in the yield. And so that's where we are, and the results have showed that over the last several quarters.

Colby, I would just add one thing. The other thing I think that's important to add to what Charles said, we also -- when we think about CapEx for 2015 beyond -- and beyond, we want to continue to be very disciplined about how we deploy that capital. But as you're aware, we also have 13 projects of size and scale today that are underway such as a Toronto or Singapore 3 or a London 6 large deployment. And so we're going to be very measured about how we deploy our capital on a go-forward basis. And I think Charles is absolutely right about the comment that we're going to -- it's going to be roughly in and around the same order of magnitude. But we're also going to reserve the right to take the opportunity when it exists. But as you can appreciate, it takes a long time to build out some of these assets. So as we continually think about how to deploy capital, our teams are already thinking about '15, '16 and '17. And so I feel we're in a very good position to moderate the consumption of our capital on a go-forward basis and be thoughtful and prudent about how that gets deployed.

Our next question comes from the line of Mr. Michael Bowen of Pacific Crest.

A couple of housekeeping things. I'm not sure I heard on enterprise. I think you gave out the percentage of revs the last couple of quarters. If you could provide that, that would be terrific. And then with regard to the recurring CapEx definition as it pertains to AFFO, do you guys have an update on that? Has that been finalized? And then finally, with regard to Performance Hub, you announced some wins. How can we think about that as far as revenue impact and margin impact going forward? If you can help us think about that, that would be terrific.

Okay. So let me take the middle one, if I could. And just as we get -- as we sort of -- and then Charles and Steve will take the other 2. So as it relates to AFFO, as I said, this quarter, we're raising our guidance to greater than $745 million. It includes a $5 million headwind relative to currency. And when you add in the $32 million of what I call REIT-related operating costs, you get a sense of roughly greater than $780 million. That also assumes, as you know, $110 million of recurring CapEx. There are still 2 lines. So when you look at basically the results that we've provided on the earnings deck, the PowerPoint presentation, there's a couple of areas that we're still looking at to make sure we look at our peer group, we look at [indiscernible] and we look at what makes a good sense for us. And so we're continuing to refine that. And so I'd ask that you just give us patience for at least one more quarter as we continue to work through that. But suffice it to say, when you look at CapEx, the $110 million that we allocate to recurring, which is about 4.5% of revenues, it has the potential potentially to change a little bit depending on how we allocate that small bucket of -- that bucket of cost.

Could you remind us the initial question?

Yes, the other question was with regard to Performance Hub, with regard to how impactful, with regard to revenue and how should we think about it from an incremental margin standpoint. The other was the percentage of revenue on enterprise. I think last quarter, if I'm not mistaken, it was 11%, up from 9% the prior quarter. If you have that, that would be great.

Yes, I think the deck shows that again, you have to understand that we are -- we separate out an enterprise vertical that we distinguish from our other verticals, financial services, content, digital media, cloud, et cetera. And that one is also -- that 11% now is shown in the deck, I think on Page 15 of the deck that we posted today. But one thing that is very important to remember is that the momentum of our enterprise class offers, if you will, is not necessarily reflected just in that number because the number of the buyers of those are customers who live in other verticals at least in our classification. In fact, our largest wins to date have been with customers -- well, some of our largest wins today have been with customers that live in the other verticals in our categorization. So for example, we referenced a very large global Performance Hub with an online banking -- for an online banking application that we have this quarter. That would be -- that would fall in our financial services revenue line. And we also had some with content and digital media or cloud customers that would be in there. And so that's what we're seeing in terms of momentum for our enterprise class applications. Having said that though, we definitely are seeing momentum in what we categorize as the rest of the broader enterprise or what our CMO likes to refer to as the other -- the rest of the enterprise. And in fact, that category is our largest source of new logos, and we've been very effective in getting in there, getting admittedly smaller wins and then really being able to land and expand from a geography and a product set standpoint with them.

Okay, great. And I guess where I was going with the Performance Hub, Steve, was maybe to put a finer point on there. Principally, the question is one of the things that I've been getting questions on is with regard to Performance Hub and Cloud Exchange. Technically, I guess, isn't there an opportunity cost for incremental revenue that might be going directly to the cloud provider rather than incremental revenue within your data center? But I'm assuming you're going to answer that in a way that -- or in a manner in that the overall incremental impact is positive. But can you help us think about the puts and takes there as far as opportunity cost but then offset by an increase in revenue to the -- for the Direct Connect to the cloud provider?

Yes, I mean, I think what you're getting at is probably a very common question that we get, which is if people are taking workloads and moving them directly into the public cloud, doesn't that represent a net reduction in the sort of colo opportunity? And the reality is, is that in aggregate, that may be true across the broad colocation market. But as it relates to Equinix, who tends to play in a very specific sort of premium application segment where people are moving workloads that are -- require network density, global reach, ecosystem reach and sort of mission-critical performance, we see it as a very sort of net positive opportunity for us. And so what we're able to do is we look at enterprises who may, in fact, be looking to move certain applications from their basement into a hybrid cloud architecture. And they may take stuff that was living in their data center and move some of that, particularly back office applications or other nonperformance sensitive applications, into a cloud environment. But the way they're going to do that is they're going to purchase a Performance Hub implementation from us across probably a global reach, a global opportunity for us, 3 or 4 or 5 locations, interconnect that to Cloud Exchange and then move their workloads through that Performance Hub. And so we see it very much as an upside opportunity for us.

Michael, the simplest way to think about this, these are service offerings, both of them, that help customers connect their workloads to public cloud access points. So the cloud exchange is trying to accelerate that specifically. The network Performance Hub, we've been utilizing for several quarters. And it's to help a CIO who's distributing their wide area network in a more optimized, cost-effective, higher-performance way to get -- to solve all the challenges that they're all seeing with mobile, social, cloud, et cetera. So these are service offerings that actually is helping customers connect to cloud access points -- or access nodes. And when you consider that 80-some percent of the IT spend in the world today is sitting in-house, on-premise and a lot of that stuff is being predicted to come out to the market, get colo-ed or outsourced or managed service, we're going to catch a lot of that workload that's coming out to the market in the future because they're going to look inside of Equinix and they're going to see all this network density. They're going to see all this cloud density and they're going to say, "That's how we can build a solution, the multi-cloud, multi-network drives that I need to go run my company." As we've mentioned to you guys multiple times, even Equinix as a $2.5 billion company uses 30 to 35 cloud providers just to run our IT. That's exactly what we see happening with enterprises today. We're trying to figure out how to get to the multi-cloud environment. And these too -- these are offerings to help them do that.

Our last question comes from the line of Mr. Mike McCormack of Jefferies.

Maybe just a comment, Keith, that you gave on the MMR (sic) [MRR] outlook in the 4Q. What are the key drivers around that? And then thinking about cloud installments in particular, any differences in what you're seeing in churn among those installments?

Yes, Mike, at least from an MRR perspective or we refer to as yield, there's really no change that we're -- no meaningful change that we're expecting in Q4. And again, as you can appreciate, given the size of our installed base now, any movement in either direction is going to be relatively dramatic over a short period of time for us to meaningfully adjust those metrics. So that all said, we feel confident in the firmness of our metric both on a global basis as we calculate globally. We're up roughly 1%, I think, quarter-over-quarter and 3% year-over-year. We feel very good about the firmness of that metric. And it gets supported by the fact that, as I said earlier, increasing infrastructure density and the deployment, increasing interconnection and then just the volume of opportunity we see. All of that would lead me to believe that we're going to continue to be playing in this range for the foreseeable future.

Yes, and Michael, let me take the second one with regard to the sort of cloud footprint and the churn dynamics around those. What I would say is that for the most part, cloud is a relatively newer phenomenon over the last several years in terms of driving significant deployments. And so we're probably inside of the sort of contract length on many of those. But what I would say is that we have developed over the last several years a very robust understanding of the churn dynamics in our business and what represents a deployment that is going to be sort of subject to churn or price compression versus those that are going to have high -- very high levels of retention and therefore higher customer lifetime value. And what I would tell you is that the dynamics of churn -- of cloud deployments tend to really favor them being high retention, high lifetime value deployments. And the reason is several-fold: one, they tend to be revenue facing; two, they tend to be very well interconnected; and three, they tend to be global and so they're deployed cross region, cross metro. And we actually have a pretty hard handle, analytical handle, on something we call our -- it's basically an index that we have where we can calculate the likely retention of deployments based on those dynamics. And I would tell you that generally, cloud deployments stack up very well along those dimensions.

Thank you. That concludes our Q3 call. Thank you for joining us.