Greetings and welcome to CoreSite's Second Quarter 2019 Earnings Call. At this time, all participants are in a listen-only mode. A question-and-answer session will follow the formal presentation. [Operator Instructions] As a reminder, this conference call is being recorded.I would now like to turn the conference over to your host, Carole Jorgensen, Vice President of Investor Relations and Corporate Communications. Please go ahead, Carole.

Thank you. Good morning, and welcome to CoreSite's Second Quarter 2019 Earnings Conference Call. I'm joined here today by Paul Szurek, President and CEO; Maile Kaiser, Senior Vice President of Sales; and Jeff Finnin, Chief Financial Officer. Before we begin, I would like to remind everyone that our remarks on today's call may include forward-looking statements as defined by the federal securities law, including statements addressing projections, plans or future expectations. These statements are subject to a number of risks and uncertainties that could cause actual results or facts to differ materially from such statements for a variety of reasons.We assume no obligation to update these forward-looking statements and give no assurance that the expectations will be obtained. Detailed information about these risk is included in our filings with the SEC. Also on the call today, we will refer to certain non-GAAP financial measures, such as funds from operations. Reconciliation of these non-GAAP financial measures are available in the supplemental information this is part of the full earnings release, which can be accessed on the Investor Relations page on our website at coresite.com.And now I will turn the call over to Paul.

Good morning, and thank you for joining us. Today, I will review our second quarter financial highlights, recap several significant second quarter events and provide an update on our development pipeline. First, I'd like to welcome to today's call Maile Kaiser who is substituting for Steve to cover our sales results and our sales strategy and execution. Steve is enjoying some well deserved family vacation time. Jeff will then take you through our financial results and financing activities and provide an updated guidance overview.Turning to our financial highlights for the quarter, we grew operating revenues 4.7% year-over-year delivered a $1.27 of FFO per share, grew adjusted EBITDA 2.4% year-over-year and achieved a record quarter for sales with $27.3 million of annualized rent including solid retail, scale, and new logo sales, and an unusually strong quarter for hyperscale sales, all of which Maile will discuss in more detail.Next I would like to highlight several significant events, some of which we shared on our call in April. First, our pre-lease at SV8 in Santa Clara. For two of the three phases at that data center with expected in-service timing of late Q3 for Phase 1 and late Q4 for Phase 2. Second closing on our SV9 property purchase, for a future data center addition to our Santa Clara campus. Third, closing the financing of senior notes totaling $400 million, and fourth receiving permits for LA3 data center leading to construction starting in early July.In addition to the SV8, SV9, and LA3 milestones, we also made substantial progress on the rest of our development pipeline. We completed construction at LA1 a data center expansion of 17,000 square feet, and also leased nearly 30% of the computer room to a large network and content provider. We completed construction of our last phase of LA2…

Thank you, Paul. Today, I'll start off with a summary of our quarterly sales and leasing results and then talk more about our sales strategy and execution.Moving to our sales. For the quarter, we had new and expansion sales of $27.3 million of annualized GAAP rent. These sales reflected a record quarter for CoreSite where we delivered solid core retail colocation sales and a large pre-lease at SV8 in addition to other scale leasing.Turning to a few highlights on our sales. Our $27.3 million of annualized GAAP rent for new and expansion sales included $5.3 million of core retail colocation sales and $22 million of hyperscale and scale leasing including our pre-lease at SV8. These new sales included a 143,000 net rentable square feet reflecting an average annualized GAAP rate of $191 per square foot and strong sales of new logos. Our $5.3 million of annualized GAAP rent for core retail colocation sales included solid pricing on a per kilowatt basis compared to our trailing 12-month average. Looking at new logos we won 43 new logos, compared to 30 last quarter. These new logo additions reflect the highest number of quarterly additions in five quarters include quality new brand names that enrich our ecosystem and represent total annual rent nearly double our trailing 12-month average. We believe these new logos should help drive future growth and remain a focus and priority for us.Renewals were another key aspect of our leasing. During the second quarter, our customer renewals included annualized GAAP rent of $24.1 million reflecting our growing base of business and strong customer relationships.Our renewals represented rent growth of 2.6% on a cash basis and 7.4% on a GAAP basis, but included, as anticipated, higher than historical churn of 2.4%.Next, I'll talk more about our sales strategy and execution. I've…

Thanks, Maile and hello, everyone. Today, I will review our financial results for the quarter, provide an overview of our April and July financing activities, and discuss our financial guidance.Turning to our detailed results for the quarter. Our total operating revenues were $142.9 million for the quarter, which increased 4.7% year-over-year and 2.9% sequentially. Operating revenues consisted of $121.1 million of rental, power, and related revenue, $18.8 million of interconnection revenue, and $3 million of office, light industrial, and other revenue.Interconnection revenue increased 7.8% year-over-year and 2% sequentially. FFO was $1.27 per diluted share, which decreased $0.01 per share year-over-year and increased $0.02 per share sequentially. Adjusted EBITDA of $76.7 million, grew 2.4% year-over-year and 2.9% sequentially. Adjusted EBITDA margin was 53.6% for the quarter, consistent with our trailing 12-month average of 53.8%. Sales and marketing expense totaled $5.8 million for the quarter or 4% of total operating revenues. General and administrative expense totaled $12.3 million for the quarter or 8.6% of total operating revenues, including elevated legal expenses, adding approximately $0.04 per share in costs this quarter that I will address further related to the impact on full-year guidance. We commenced 65,000 net rentable square feet of new and expansion leases for the quarter at an annualized GAAP rent of $176 per square foot, which represented $10.2 million of annualized GAAP rent.Moving to data center sales backlog. As of June 30th, the annualized GAAP rent from signed, but not yet commenced leases was $26.1 million or $29.5 million on a cash basis. We expect most of the GAAP backlog to commence in the next two quarters.Turning to our property operations and development. Same-store monthly recurring revenue per cabinet equivalent for the quarter was $1575, reflecting an increase of 6.1% year-over-year and an increase of 1.2% sequentially. Q2 same-store turnkey…

Thank you. We will now be conducting a question-and-answer session. [Operator Instructions] Our first question comes from the line of Jordan Sadler with KeyBanc. Please proceed with your question.

Thank you, and good morning out there. So first I wanted to just address the changes in the guidance. If I could dig in a little bit more, the churn elevated again for the second consecutive quarter. Can you sort of identify with the specific drivers were that caused you to increase it again, sort of 90 days later, Jeff? I mean, I think you probably had a decent handle 90 days ago which caused you to increase the guidance there? And then here we are 90 days later with an incremental sizable bump at the midpoint, at least, to the overall churn rate. And I'm just wondering what's driving that, especially seven months into the year?

Yes, good morning, Jordan. Let me see if I can give you some color related to your question. And hopefully it helps explain the increases and what we're seeing inside the portfolio, but just a little bit more broadly speaking, if you look at our churn over the last several years, our average has been right around 7.3%. If you look at the bankruptcy that we talked about last quarter plus another small one we've got, that increases that average churn by about 120 basis points. So, all in there, you're at about 8.5% and what we've seen since that as we continue to look forward is we've got some additional customers who are looking to migrate some portion of their IT architecture into the cloud and that element of what is really resulting in the increase for this particular guidance is the incremental amount we're seeing and looking at for the rest of this year. And so that explains the timing associated with it. I think we are optimistic that we will retain a larger share than what we are of their deployments, but knowing what we know today, we felt more comfortable to increasing that guidance to the levels of which we have guidance.

So I guess I would characterize that reason as sort of a secular driver, I mean something that you guys have been experiencing and seeing broadly, you and your peers, for a couple of years. Is there, are you baking in some incremental room for additional potential customers to migrate to the cloud? Or how do you, what's with the sort of out-sized migration to the cloud to sort of happening in the back half, is just one large customer?

No, it's a handful of customers that are driving that. And most of that as Paul alluded to, and we're experiencing in three markets, two of those markets are on the West Coast, and so it is a-- I'd say, four or five customers that are really elevating that churn. And again, we just, we're estimating, we'd be able to retain more of their architecture than ultimately we think we're going to.I think to your comment is to it being a more broad secular driver, I think inside our portfolio, it hasn't been as big of a driver as what we're seeing this year. And so where we're spending more and more time is getting our arms around what that looks like in the next one to three years, so we could better understand that dynamic and make sure we're obviously communicating that as best as we can. We hope this is elevated for a single year and not a sign of a trend to come, but that's something we got to spend some more time on.

Jordan this is Paul. I would just add two things. There is a little bit of a good news element to this churn we're experiencing because from what we can see pretty much all of it is going into the edge cloud cases that we're hosting. And so, it further cements the demand in our data centers for that activity. Secondly, as you know there is always different used cases that increase and then subside. And I think we're sort of at the tail end of the cycle of some of those historical use cases that are subsiding. So, I would expect, and Jeff and his team will dig into it much more extensively over the next couple of months, but I would expect that this will be a crest of churn for us and we'll see it subsiding in the coming couple of years.

Okay. That would be good. In terms of the timing of the churn, is it, I was surprised that it's a couple of hundred basis point increase since we're already past June 30, but has it churned out, have these guys churned out of the portfolio already or is this sort of a third quarter or fourth quarter event?

No, I mean the churn we experienced in the second quarter was right in line with what we had originally communicated, Jordan. So we had expected the Q3 churn to be elevated again at that 2.5%. I would say based upon what we know today, I would expect that churn in the third quarter to be somewhere between 3% and 3.5% and then back to around 2.5% in the fourth quarter.

Okay. Last question is just on legal costs. You're elevated, what is the source of these? And then could you give us the breakout again of, I just missed it in your prepared remarks, how much the impact was in this quarter?

So, Jordan. This is ongoing litigation. So there's not a whole lot we could talk about. I can tell you it involves a dispute with a landlord in one of our smaller markets that started with the construction issue, there is, the normal disclosure in our 10-Q, which we'll be filling soon. And as with all ongoing litigation, there's just not a whole lot I should comment on beyond that. To clarify the numbers, though it affected the second quarter, $0.04 a share, and we expect the full-year impact and obviously it's not perfectly predictable, but to be about $0.09 a share, full year including the $0.04.

And just to add one more level, just for your modeling purposes if it helps, the incremental $0.05 for the year will be weighted toward the third quarter.

Okay. I'll yield the floor. Thanks guys.

You bet. Thanks, Jordan.

Our next question comes from the line of Erik Rasmussen with Stifel. Please proceed with your question.

Yes. Thanks. Maybe just circling back on that. I know you had given the outlook for FFO per share and it was a $0.16 sort of impact and $0.09 was coming from legal. You gave sort of three other reasons for the other $0.07, if you had to sort of weight those three Northern Virginia or the churn or timing of commencements, how would you say that those would be weighted in terms of those, the $0.07 for that remainder?

Yes, good morning, Erik. Yes, let me just go to some of the, obviously the public comments I said earlier in my prepared script, but you're correct, I did say that the $0.09 that was attributable-- the $0.09 of the $0.16 was attributable to the legal costs and the incremental $0.07 is largely attributable to the churn, the market conditions in Northern Virginia, and the timing commencements. Those three are relatively weighted equally to be quite honest. And just so that you also have again for just full transparency, given everything we've disclosed around our guidance, there are some offsetting, expense savings, some of it in SG&A, and in operating expenses, but it's largely attributable to those two items that we highlighted, and those three items weighted relatively equal.

Great. Okay. Well then, in terms of, it maybe just asking in another way, the top line, there was a $10 million impact, what would you say sort of the primary driver there? Is it more of the churn that happened or is it the impact of slowness that you're seeing in Northern Virginia? Is there something else? And then I'll have just one more question.

Yes. I know that-- if you look at it from a topline perspective, those three have contributed to that $10 million reduction fairly equally at the top line. There is some additional power cost revenue also but the larger ones are the three that I mentioned weighted relatively equally.

Okay, great. And then, I guess then, thinking of the set up for next year and expectations for sort of accelerated growth. And I'd say return to low double-digits, at this point where do you foresee sort of the biggest risk to hitting this target as we stand today?

Well, it's a great question, Erik. I think obviously based upon what we've just talked about in some of those items influencing 2019, is providing some headwinds, and those headwinds will impact 2020, and I think we've obviously got a lot of work ahead of us related to the back half of this year. In terms of the highest risk, I think ultimately, sales and churn are probably the two, I would elevate to the highest risk. Obviously, Steve, Maile and the team worked hard on the sales perspective, and we're doing what we can to minimize the churn, but I would say, those would be the two that I would say to elevate the highest.

Yes, Eric. We have a lot of opportunity between now and year-end, we have a lot of capacity, so we definitely have the raw materials to have a very strong. We'll have a good growth year next year regardless, but the strength of it will depend on how well we can capitalize on the raw materials we have in place.

Our next question comes from the line of Colby Synesael with Cowen and Company. Please proceed with your question.

Great, thank you for taking my questions. I guess just to be specific, do you still expect to achieve low double-digit growth in 2020 or is that now less likely? And I guess as part of that, are you still wanting to do a scale lease for SV8 Phase 3 or that something that you're looking to do retail with? And then, my other question, how much are you recognizing in termination fees or expecting to recognize in termination fees, considering all the churn in 2019, including specifically what you just occurred maybe in the second quarter to the extent that it's immaterial. Thanks.

So, obviously, it's a little bit more challenging with the headwinds in Northern Virginia and not expecting as much scale and hyperscale sales out of Northern Virginia as we did six months to nine months ago, to hit double-digit growth, but it is still possible to do that for 2020, and then you carry that out with the capacity we have, in the markets that we have, and the much increased number of markets in which we have capacity, and it should be more sustainable thereafter. It's still possible. It's going to be a harder lift to get there.What was the second half of his question?

SV8.

Oh, on SV8, yes SV8, we will build that out as our traditional colo floor. Colby as you know we are flexible on all of our floors. They're designed where we can accommodate both scale, small hyperscale deployments, as well as retail deployments. And as we always do, we will pursue it on a broad range and make decisions based on customer ecosystem fit and pricing.

Colby, to address your third question as it relates to termination fee. We don't have anything material in the second quarter regarding termination fees. However, I think what you're referencing is probably the bankruptcy, the customer that on the previous bankruptcy that we talked about, and I think it's important to just give you a little bit of color around that. When we went through, we basically negotiated a lease modification with that particular customer. And as part of that process, we drove overall economics that would include some level of a termination fee, but it doesn't get called that just because of the way it gets accounted for. But just so you're aware, in the second quarter we had $0.01 a share from that particular customer. And we expect another $0.01 a share in the third quarter from that particular customer before they vacate that particular computer room.

Great. Thank you.

You bet.

Our next question comes from the line of Robert Gutman with Guggenheim Securities. Please proceed with your question.

Thanks for taking the questions. First, in the scale leasing of $22 million, does that include any other hyperscale or hyper block or small deals besides SV8 1 and 2? Secondly, you mentioned $3 million energy efficiency rebate, and I think in the quarter a footnote it said you realized $1.7, so does that mean the balance gets recognized over the next two quarters and more broadly, how should we expect recurring at capex to play out in the second half of the year in light of that and ordinary spending?

Hi, Robert. This is Maile Kaiser, just to answer your first question on the scale leasing and hyperscale, there was other scale leasing in addition to the hyperscale that made up the $22 million and I'll let Jeff answer the other questions for you.

Yes, Robert. Paul alluded to his prepared remarks that we actually received $3 million, which is ultimately the cash that we received, and ultimately we allocated the $3 million pro rata to both recurring capex and expansion capex based upon the way we spent those dollars on the original project. And so, when we were spending the money. A portion of what we spent was included in recurring capex and that related to that portion that was already built. And then another portion was expansion capex, it related to the expansion of that particular data center, and so the $1.7 million just reflects the pro rata portion of that allocated to recurring capex initially and that's what's given you that, that decrease in recurring capex for the second quarter. Does that explain what you were asking?

Sure, but I think it was also, recurring capex in the first quarter was elevated, now it's actually a positive number with the rebate, what should we expect it to be in the next several quarters, what's the normal rate?

Yes, that's a great question. So, I think if you just ignore the rebate in the second quarter, we would have had about $1 million of recurring capex in the second quarter, and obviously that's down just a little bit from the first. We've typically been forecasting somewhere between $1 million to $2 million per quarter of recurring capex for 2019, and that's what I would expect in the third and fourth quarter of this year.

Thanks. One more if I may? In the commencements that were delayed, was that a function of the opening times, like construction or was it a function of customer accommodation moving in later?

I think it really reflects the more new customers in this hybrid cloud type of architecture, and they tend to take longer to make some of the decisions they need to make to set up their space and even sometimes finalize the sales order for what they need. They have more decisions taking place than customers, typical customers two-three years ago would have had.

Got it. Thank you very much. I'm sorry. Go ahead.

I was just going to say, this is Maile Kaiser. And I would just add that, that it really is a lot of times customer-dependent on their timing and putting together all of that architecture in the design and if it's a migration versus a new project. All of those commencements can change.So, we're really at the mercy of their timing and requirements .

Thank you very much.

You bet. Thanks, Robert.

Our next question comes from the line of Nick Del Deo with MoffettNathanson. Please proceed with your question.

Hi, thanks for taking my questions. First, cloud has been putting some pressure on the MSP business model or at least forced them to evolve some. In a general sense, do you think MSPs and system integrators are as attractive customers for CoreSite as they used to be, or would you describe them as higher risk now?

You know what I alluded earlier to churn we're having in traditional business models. I think we've seen over the last four-five years kind of steady bleed-off of MSP, hosting, some private cloud options that are not specifically geared to a hybrid cloud architecture. So I think a lot of our churn this year is related to some of those companies come into their final reckoning. So, yes, and we don't see those, Maile do we see those types of customers budget in our funnel these days?

Well, I think that we see them in a new form. So we definitely do quite a bit of business with SI and MSP but there are a smaller deployments and they're more aggregating hybrid cloud architecture and then helping to manage that hybrid cloud architecture. So, traditionally, they used to purchase larger deployments and try to build out their own private cloud. Now, what we're seeing those customers do, is provide additional services on top of a hybrid cloud architecture and deploying those services in our data centers.

Got it. Can you share what portion of revenue they represent?

Yes. I think if you do a deep dive, just depending on how you categorize it for us, Nick it's about 2% of our embedded here today.

Okay. So it's pretty small.

Yes.

Okay, that's helpful. And then one last one, would you consider the Northern Virginia market soft enough that any scale leasing renewals you're going to see in coming years are likely to face pricing pressure that you haven't historically seen?

It depends on how quickly the market recovers. I don't think we have any coming up for a couple of years, but if market conditions persist today we probably would have some negative market-to-market on some scale leasing.

Our next question comes from the line of John Atkin with RBC. Please proceed with your question.

Thanks. Got a couple, I'll ask them upfront and you can answer them in whatever order. But when you fail to close on a scale deal opportunity, is it because of churn? is it price? is it product? just curious to get a little color on deals that maybe came close, but didn't quite finish it? Secondly, on the new logos, I'm interested in kind of any qualitative color around the cross-connect intensity of the new logos you're bringing in? And also, is that primarily coming from channel or is that coming from the efforts of your direct sales people in terms of the new logos?And then on the churn, I was wondering if you could help us localize where on the West Coast? Is it some of your older inventory outside of Santa Clara and San Jose in Milpitas in Northern California? Or is it on the center of the campus that you were seeing some of the churn? And then in LA is it Alameda or One Wilshire? Thanks very much.

So let me try to address everything except the new logos, and I'll let Maile address those. Obviously I get involved in every large scale deal and every one of them that we have missed on, I believe, were just 100% pricing, where the pricing was below what we thought was the right trade-off. And understand that if our retail sales in North Virginia-- if our retail sales in Northern Virginia were not going as well as they were, we'd probably be more tempted to take some of this discounted pricing, but we're doing pretty well on the retail bucket and that continues to seem to gather momentum. So right now I think we can continue to pursue that strategy for leasing up VA3 until better opportunities come along on the scale side.In terms of the churn again, most of this is, most of the churn within the Stender campus in Santa Clara and in Alameda at LA2 and obviously the Boston situation. So Maile you want to talk about the new logos?

Sure. So I think as we look to target new logos and bring on a new customers, we're looking for customers that value our ecosystem and our community effect, and with that and those deployments that are looking to be closer to the edge and interconnect their workloads with our service providers in our cloud on-ramps. We see that that brings in more cross connect activity. So we feel pretty good about the new customers that we've brought on, and the mix of customers, they're actually, about 55% of the new logos were coming from enterprise, and the rest were coming from network and cloud, and those enterprises are looking to interconnect with several cloud providers and network providers and services. So we expect to see a growth in cross connects with those customers coming on board.

And then lastly, is there anyway to kind of characterize the close rate that you've seen, is it same close rate, but just kind of a smaller deal funnel late stage or what was kind of some of the dynamics that you've seen? And then going forward, would you characterize the late stage funnel as being at average levels, above average, below average as we sit here today?

So I think that the good news is, we are improving our win ratio that is one of our top priorities for 2019 in addition to improving sales productivity and increasing the funnel of new opportunities. So, both are critical. I think that we've seen an increase in the quality of new logos and opportunities in our funnel where we're trending in the right direction with our improved win rate. And again, our focus is on bringing in more and more customers that value our ecosystem and need access to cloud on ramps and the interconnection support that we provide in our data centers.

Thank you.

Our next question comes from the line of Nate Crossett with Berenberg. Please proceed with your question.

Hi, thanks for taking my question. I just wanted to get a sense of how you're thinking about yield for the development pipeline as you do more scale leasing? And do you think you'll be able to maintain that 12% to 16% that you quote or should we expect that range to kind of migrate lower the more scale that you do?

So, Nate, it really depends on the market being very candid about Virginia, if we did a significant amount of scale in VA3 we would at today's pricing not hit that 12% threshold. On the other side of the coast that with, as long as we have an appropriate mix of retail, we can still do a significant amount of scale and hit our underwriting hurdles.And I believe that will be true for LA and Chicago based on the interest we're seeing in the early stages of construction, time will tell on those. But in those markets right now the supply demand balance seems to be much more healthy than it is in Northern Virginia.

Okay, and then just longer term like what percent of your overall portfolio should we expect to be scale. I guess is there a certain threshold you don't want to go over, or is it just customer?

It's really hard to give you any guidelines on that because it's going to vary based on pricing and supply and demand in each market. So I think that is going to be a derivative variable as opposed to an outright target.

Okay and then on just the land bank, are there any other purchases like SV9 that are in the works? What's the amount of capital you're allocating toward kind of land purchases?

I think it will be very modest going forward. We do want to, we've got SV9 in Santa Clara and very glad to have that. We will, down the road, probably need to acquire additional pieces of land in Los Angeles and Chicago. But we are good for a long time in Northern Virginia. We're good for a long time in the Northeast. And obviously we're in good shape in Santa Clara.

Okay, so there is no new markets on the horizon any time soon.

We keep looking at them. I'll be honest, the more we look at new markets, the more grateful I am to be in the markets that we're in. I think relative to our size and based on the quality of our markets in terms of supply and demand in the other drivers of demand. I think share per share we have more opportunity for growth relative to our size as far as I could tell among our peers. So time will tell if that's an accurate prediction, but I feel pretty good about where we are right now.

Our next question comes from the line of Frank Louthan with Raymond James. Please proceed with your question.

Great, thank you. So, talk to us a little bit about the pace of some of the hyperscale deals that you've got, when can we expect to see those kind of layer in and start billing? And I apologize if you already addressed this, but how do you feel about funding for that? do you, do you need to raise additional capital over the next 12 months to 18 months to fund those?

Hi, good morning Frank. Well, I think we did reference it, but let me just to make sure you caught what we said is, when you look at SV8 Phase 1 and Phase 2, which is really what is a majority of our backlog today, we've got 6 megawatts of that 12 coming online late Q3, and the remainder coming on late Q4. And so that's the timing associated with that particular deployment.In terms of liquidity, we've got about $335 million of remaining spend based on what's under construction today, and our liquidity, we ended the quarter at somewhere right around $460 million. So we've got plenty of liquidity to take care of what's under construction today, so I don't see any needs for additional liquidity before year-end. But, it is something we look at, and we always want to be watching the markets to ultimately see if we want to do something based upon pricing and where markets are headed. But as I look at it today, and I don't think there is a need to do anything. And we'll obviously probably have through the first half of next year without any real issues on the liquidity perspective.

Okay. Thank you. And then just a follow up to the question, so SV8 Phase 1 and 2 in Q3 and Q4, what sort of the revenue ramp we should expect? Is this, with the hyperscale deal, is this going to be a longer revenue ramp for the full billing run rate or will they be sort of at full rate in those, by the end of each of those quarters?

I got to be a little careful here just due to the confidentiality we have with a customer. But just to help you out there, there is not a significant ramp there. I'd look at it fairly reasonable in terms of when it's coming online late Q3 and late Q4.

Our next question comes from the line of Aryeh Klein with BMO Capital Markets. Please proceed with your question.

Thanks. So, of the 7% or so on historical churn, how much of that has been associated with cloud migration? And then in your prepared comments you mentioned that you're hopeful that pricing would improve in Northern Virginia over the next year or so, what could drive that since it seems very challenging right now?

Good morning Aryeh. If you look back historically, as it relates to our churn, and you know, you referenced that 7.3%, in terms of what's migrating to the cloud historically, it's been a relatively small percentage. We've had some of it over the years, but nothing to the levels of what we're seeing this year. So I would say, if I had to quantify it, it's somewhere around, call it 25 basis points to 40 basis points, maybe 25 basis points to 50 basis points on an annual basis. And I'm sorry Aryeh but what was the second question?

Aryeh, let me, I think you're asking about the potential for change in the Northern Virginia market and what would drive that. Is that right?

Yes.

I mean, I think we're already seeing a little bit in that some of the buildings that we expect to come out of the ground by now, appear to have been postponed, what is behind that in every case, we don't know. We do, we are aware of a couple of prospective builders who have publicly said they're going to go on pause and just look for build-to-suit opportunities. So, supply will be affected by those types of actions, demand will pick up probably when the cloud hyperscalers get into their next, get out of their absorption phase of what they leased in 2017 and 2018 and move into the phase of having to lease or produce additional capacity to accommodate the growth. Historically that tends to be about 12 months to 18 months cycle between the bottom of those cycles and coming back up more to the top. So maybe we're, I don't know 12 months away from that, but it's hard to predict with certainty. Is that, is that responsive?

Our next question comes from the line of Michael Rollins with Citi. Please proceed with your question.

Thanks. So a couple of follow-ups, if I could. So, first on the question of Virginia, from looking at the rent schedule correctly there is about $53 million of annualized rent. Based on current pricing, what would you estimate the revenue risk would be over time in the mark-to-market?

We'd have to going to go back and calculate that relative to scale and retail, you know, Virginia still is primarily a retail data center campus for us and retail pricing has been pretty solid there. So, I don't know what the magnitude of that would be, but we'll try to present that on our next quarterly call.

Yes, Mike just to further that, as you look at the numbers we've disclosed, keep in mind, what's in DC1, DC2 is really the retail side of the business, and then the composition of the Reston campus. If I had to give you an estimate, it will give you something more solid, I'd probably say at least two thirds of that is going to be built up with our retail business. And so just keep that in mind as you think about that. And you guys are asking very good questions related to pricing and let us just dig into it more we'll give additional color and commentary on that as we move forward.

And secondly as, as we think about the comments that you made on churn and the risk of customers migrating some infrastructure into the cloud. If we look at the segmentation schedule on Page 15 of the leases by size of NRSF, is there a story to tell of where their susceptibility in terms of the size of the leases? And is there some way that you've internally estimated like the long-term risk ,long-term exposure, where a certain class of the revenue is more susceptible to some of the migration that you're seeing from the handful of customers that you mentioned for this year?

Well, I think you're trying to look at it with, obviously, the information we've given as best as you can. And when I think about what we're seeing this year, I would say most of that churn is probably bucketed in two portions, one would be the 5000 or less square foot bucket and the other one would be between 5000 square foot bucket to 10,000 square foot bucket, and that's why you've seen some of the numbers being elevated just due to the size, but in terms of what, what that looks like going forward, where is our biggest susceptibility, that's a, it's a great question. I think that the most important thing is ultimately as Maile and Steve have alluded to in the past, we're ultimately trying to make sure we can address those hybrid cloud solutions. And the question is going to be, what portion of their architecture that we, that might reside with us today, either is retained with us or that moves to the cloud, and I can't, I don't think I have enough data today. It says because just see how much of that we've been experiencing, to be quite honest. It's just not that we're going have to dig into more and see if we can get it quantified for you. But just to give you an idea, those two buckets is where most of that churn resides today that I referenced earlier.

And Michael, so far what we're seeing are really specialized use cases that are going into specialized edge cloud used cases. Most of what we have in our data centers, the network, the cloud obviously, that stuff is pretty impervious to migration of the cloud, and most of our enterprise occupancy is come in the last two, three years from customers that have already moved primarily into our data centers to do the hybrid cloud type of activity. I should point out, we also, one of our new logos this last quarter was somebody coming out of the cloud with a significant amount of capacity. So we're going to continue to see these dynamics and Jeff and his team have been working diligently this year to develop better predictive analytics around these things and it's still a work in progress, but I think we have enough visibility to not feel terrible about where the churn situation is with respect to cloud migration.

And this is Maile. I think just to highlight what Jeff and Paul have shared, as we see our customers start to take advantage of some of the workloads that they can put into the cloud, we are going to see that right size of the deployment. But we're also finding that it's the exact same reason that we are winning new hybrid cloud deployments due to that exact need to have their workloads connected with the cloud providers and with low latency in the edge market. So where on the one case, we do see some churn on the other case, we do see new logos and then we also see it feeding into our edge campus cloud customers. So kind of a three pieces to that puzzle.

And so one final question. And so in the guidance change, the guidance for revenues changed, but not the interconnection revenue guidance. So, does that provide some observations about how your customers are valuing connectivity in that portion of what they get out of the data centers maybe to your point about the size of the deployment and if there's maybe some additional color there that would be great.

Yes, I think just to give you some quantity numbers, I think obviously year-to-date our interconnection revenues up 9.4% cumulatively year-to-date. I think our midpoint of our guidance is 8.3. So we're comfortable where our numbers are for this year, and obviously, we didn't change the guidance. I think we've seen, in the first two quarters, we've seen good volume growth inside our business and that does play toward ultimately those customers that are needing to connect, whether it's to the cloud, whether it's to the networks, actually it's some of both. The growth in interconnection inside our portfolio continues to be very strong for those people needing to connect to the cloud.And I don't think that dynamic is going to change or subside anytime soon. And as long as we can continue driving those types of deployments Maile described, I think that will continue, at what levels, not certain, but we're comfortable with where we are for 2019.

Yes, Michael, on a micro level for an individual customer to the extent they're keeping some deployment in our data center, they will either maintain their current cloud connectivity or even increase it depending upon how many additional clouds they use and how they architecture to potentially connect to different cloud regions for additional resiliency.

And we've added a couple of new products into our interconnection portfolio as we've mentioned earlier the CoreSite interconnection Gateway that allows them to interconnect and manage the connectivity between workloads to different cloud providers, as well as the inter-site capabilities between markets. So connecting their existing deployment into a new market, so more interconnection product that we have, will continue to drive the interconnection revenue.

Our next question comes from the line of Jon Petersen with Jefferies. Please proceed with your question.

Great, thanks for sticking with us here. So if I could just a little bit more on the churn, if you could just maybe give us how you're thinking about backfilling this level of churn, or maybe just to think about more from a real estate perspective, obviously your occupancy is going to be going down through the rest of the year. How should we expect that to trend in 2020? And as you think of about getting back to 2020 or getting back to double-digit growth in 2020, it's kind of a quick turnaround of releasing that space required?

Yes, hi, Jon, this is Maile. So we are actively pursuing backfill opportunities for any of the areas where we expect to see customers churn out. The good news is in most markets, the the location in which there, we're going to see that come back, is in turnkey colocation space and computer rooms, so very quick to be able to turnaround and support new customers. We've already been able to backfill a few opportunities and consistently look for pipeline to support the ones that are coming up.

It term of.

Go ahead.

Jon, I'm sorry, I was just going to say in terms of occupancy, I mean if you look at our same-store portfolio, it's declined a little bit here over the last couple of quarters and I would expect that to continue as you point out, it's probably going to decline probably another 200 basis points to 300 basis points or before year-end based on the churn we're looking at today.

Yes, that's kind of what I expected. I guess my thought, my question was more about as we look into 2020 do we expect that to reaccelerate?

I think it would, Jon. As Maile was saying, this churn space is basically our generic bread and butter colocation product. So we just keep selling.

Okay. I think you probably just answered this question, but just to be clear, when we think about your development pipeline versus this now vacant churn space, is there any different in kind of the ability to sell that space? And do you now have to kind of pulled in those markets where you now have more vacant space or are you going to pull back on new developments for a quarter or two while you backfill that vacant space?

All the development that we have in process right now, we will continue to press forward on because we think we're going to need it. The new development quite honestly gives us the ability to handle much larger opportunities than what the churn space enables us to handle. The churn space just comes back into our generic retail colocation inventory.

And Jon the only change related to that is we did purposefully move back the completion of the Boston computer room due to the one computer room we're getting back. And so to your point, it helped, we're going to delay the completion of that room, but we still think longer term, we're going to need it. And as a result, we're continuing to move forward to just the timing is pushed back by quarter.

Our next question comes from the line of Sami Badri with Credit Suisse. Please proceed with your question.

Thank you. My question has to do with visibility of these discontinuation of applications, you referred to some of your customers doing that. Now when this happens in your business, how much of general visibility do you have before it happens, before they have this conversation where they want to disengage out of colocation and move to another type of workload or just ended application. Do they communicate this to you, like every month when it happens, three months, six months just so we have an idea for as this transition takes place. We can just have an idea on how exactly the dynamics are impacting your ability to guide?

So it really depends on the customer and the type in our traditional types of churn the SI, MSP, hosting companies that appear to be into their subsided cycle. We've generally been able to predict that based on the account representatives interactions with the customer plus some details around their activity in this space and historically we have been pretty accurate with that. Little bit harder to predict that with the things that might go to the cloud, because the customers themselves even though we are in close contact with them, are still evolving in some of these decisions, and so one month, it's yes we're not going to do anything for some period of time and then two months later, after more understanding of what cloud products are available, they make a change in decision. So that's why that segment is a little bit harder to predict than our traditional churn forecasting.

Got it. And then we talked a lot about Northern Virginia market softness and most of the people on this call probably know that Northern Virginia is a very major market. So would you say that this specific dynamic is isolated, or you know like the supply and demand dynamics. Would you say that these are isolated to Northern Virginia or do you think this is a overall sector deceleration that we're going to start hearing about in some of the other markets over the rest of the year.

So in our markets. The only place we're seeing it is Northern Virginia, I've heard similar rumblings about the Phoenix and Dallas markets. But we're not in those markets, and I should also qualify that we're not, we don't have the capacity to be going after 12 megawatt to 18 megawatt deals in Northern Virginia. So, there might be less visibility for us into that particular subset of the market than others are seeing, and like you guys, I'm looking forward to seeing what we learn as our peers with activities in Northern Virginia report, in the next week or so, but so far, supply and demand, as I said, it looks pretty good in all of our other markets it looks pretty balanced, and again, the long-term secular drivers for demand. The new data products, the increase in the amount of content being pushed out. You know the increasing interaction with consumers and demand for lower latency, those all seem to be very much in place. And that's why we're seeing kind of consistent sales results across our core products and are hopeful of increasing that, over coming quarters, and why we still believe there will be periodically opportunistic sales for scale and hyperscale.

Our next question comes from the line of Richard Choe with JP Morgan. Please proceed with your question.

Great, thank you. Just a follow-up on the churn aspect and migration to the cloud. Normally, the life cycle is you go from being in the cloud and move to dedicated, and so this is kind of a reversal of that, but I guess, given your comments so far, it seems like it's with certain applications and very specific, do you think this is going to be more broad based or is this very application-specific? And then in terms of maybe on the other side of it, it seems like that you're getting some benefit in that, you're not completely losing the customer in that they continue to want connectivity and infrastructure around cloud providers. So can you kind of maybe give us a bit of a more holistic view on kind of the net loss versus the net gain on that? Thank you.

So, we actually track that pretty well, are pretty diligently, and so far between what we've leased for cloud applications, including all the different components of the cloud companies and edge cloud use cases plus what we've leased to companies coming into our data centers so that they could do hybrid cloud in the most secure and high performance way possible, we are way ahead on the whole cloud migration element. And I'm guessing If you talk to your colleagues in the enterprise space, they'd probably tell you that the vast majority of all migrations to the cloud are happening out of proprietary enterprise owned data centers.Again as I mentioned earlier, most of what we're seeing are specific use cases. And we're also seeing a lot of hybrid cloud that's coming to us specifically because they have the cloud operations right there on our campus or the direct cloud on ramps or the type of connectivity options we give them. So, net-net cloud is still a huge positive for us.

And then in terms of the backfilling, is speed of filling it more important or getting the right price in looking at the space that you're getting back?

All of the above.

Got it. Thank you.

Our next question comes from the line of Lukas Hartwich with Green Street Advisors. Please proceed with your question.

Thanks. I just want to clarify, has Northern Virginia become incrementally worse or you just reiterating prior comments around that market?

So we've been saying for about two years now that pricing for scale and hyperscale is coming down in Northern Virginia, we didn't really get to acid test that until we were out responding to RFPs for the new space coming on in VA3 Phase 1B and based on that, I would say that it has over the last three months to six months has continued to deteriorate a bit. It may be bottoming out right now. Everyone certainly hope so, but that's, that's our perception of the market.

Yes, it's maybe a little bit worse, and then you touched on this a little bit earlier, but in Boston. I noticed that a big chunk of square footage moved from stabilized to your held for development. But the AVR was the same. So I was trying to figure out what exactly went on there.

That was just a space that was leased on a powered shell basis to a wholesale customer that's in one of those business models that has been declining for a while and to release it according to our normal colocation model we'll have to put some additional capital into that space to convert it from powered shell to colocation space.

Got it.

And until we do that, it's not available inventory.

Got it. And then the rent didn't really move. Is that just a timing? Will that go down next quarter?

Yes, no it's, those powered shell, obviously are not leased at the same levels in which typical turnkey is, so it did come down, but it was offset by the lease modification we executed with the other customer, that will come down in the third quarter.

Got it. That's it from me. Thank you.

You bet, Lukas.

Our next question is a follow-up question from the line of Jon Atkin with RBC. Please proceed with your question.

Thank you. Yes, just real quick on the new connectivity products that Maile was alluding to a couple of questions back, and I wondered, maybe if it be possible to learn a little bit about customer adoption. Is it new logos taking it on day one. Is it long-standing retail customers that are, that are beginning to kind of migrate to it. Are we still in kind of really early days here or is there a mature run rate to think about as people kind of adopts the Internet Gateway and other newer products. Thanks.

Absolutely. Thanks, Jon for the follow-up. So these are relatively new products. So I think it is still early stages, we are actively proposing them to existing customers that can take advantage of new sites and interconnecting the two sites together as well as reaching new cloud zone, availability zones, from their existing site with us, but it's also, I think also on the CoreSite interconnection gateway product, we're going to see that be something that new customers are going to be interested in as they look to migrate out of their enterprise data center and into a colocation site, but they're not yet ready to make that move. So this, we see, is kind of a, a way to bring them into the data centers. So right now we're seeing some good pipeline with existing customers and we think that that's going to grow as we sell it with our solution partners.

Thank you.

Thanks, Jon.

There are no further questions in the queue, I'd like to hand the call back to management for closing comments.

So thank you all for your time and interest. Lot of good questions. I come out of every quarter, really proud of the team, the things the building blocks, we're putting in place for long-term growth have really come together well, and I like where we're going with that. I do believe the cyclical headwinds will pass and will be much happier as we have this additional space to sell, and I know one thing that you can't get any sales if you don't have the space to sell and so right now we have much more opportunity to sell than we've had in quite some time. So, I look forward to the future and I'm grateful to work with such a good great group of people to accomplish what we've accomplished so far. Thank you all very much.

Ladies and gentlemen this does conclude today's teleconference, thank you for your participation, you may disconnect your lines at this time and have a wonderful day.