Digital Realty Trust Inc (DLR) 2017 Q3 法說會逐字稿

Good afternoon, and welcome to the Digital Realty Third Quarter 2017 Earnings Conference Call.

(Operator Instructions) Please note, this event is being recorded.

At this time, I would like to turn the conference over to John Stewart, Senior Vice President of Investor Relations.

Please go ahead, sir.

Thank you, Denise.

The speakers on today's call will be CEO, Bill Stein; and CFO, Andy Power; Chief Investment Officer, Scott Peterson; and SVP of Sales and Marketing, Dan Papes, are also on the call and will be available for Q&A.

Management may make forward-looking statements related to future results, including guidance and the underlying assumptions.

Forward-looking statements are based on expectations that involve risks and uncertainties that could cause actual results to differ materially.

For a further discussion of risks related to our business, see our 2016 10-K and subsequent filings with the SEC.

This call will contain non-GAAP financial information.

Reconciliations to net income are included in the supplemental package furnished to the SEC and available on our website.

And now I'd like to turn the call over to Bill Stein.

Thanks, John.

Good afternoon, and thank you all for joining us.

Let's begin on Page 2 of our presentation.

I'd like to focus today on capital allocation.

As you know, we closed on the acquisition of DuPont Fabros during the third quarter, a high-quality, highly complementary portfolio concentrated in top-tier U.S. metros in an accretive, prudently financed transaction that expands our relationships with a blue-chip customer base.

Integration is well underway, and Andy will cover this in detail in his comments.

Earlier this week, we announced a 50-50 joint venture with Mitsubishi Corporation to enhance our ability to provide data center solutions in Japan, and we've provided a summary here on Page 3. In a nutshell, we are contributing our recently completed project in Osaka, and Mitsubishi is contributing 2 existing data centers in Tokyo.

Although the venture is nonexclusive, the expectation is that this will be both partners' primary data center investment vehicle in Japan.

Japan is a highly strategic market, and we see tremendous opportunity for growth over the next several years.

This joint venture establishes our presence in Tokyo, which has been a longtime target market for us.

In addition, we expect this joint venture will significantly enhance our ability to serve our customers' data center needs in Japan.

In particular, we expect that Mitsubishi's global brand recognition and local enterprise expertise will meaningfully improve our ability to penetrate local demand.

Back in the U.S., we also entered into an agreement this week to acquire a data center in Chicago from a private REIT for $315 million.

This value-add play offers a healthy going-in yield along with shell capacity that gives us an opportunity to boost the unlevered return up into the high single digits.

This investment represents an expansion on our core markets and is occupied by existing Digital Realty customers, with whom we have been independently working to meet their expansion requirements in the market.

During the third quarter, we took steps to support our customers growth in key metro areas.

We acquired land parcels adjacent to our existing holdings in Osaka, Japan and Garland, Texas during the third quarter.

These parcels will ultimately support development of more than 50 megawatts of incremental capacity once our existing campuses in these markets are fully sold out.

During the third quarter, we also announced that we are breaking ground on a new 14-megawatt data center in Sydney, Australia adjacent to our existing facility in Erskine Park.

We also announced the expansion of our Silicon Valley Connected Campus with a 6-megawatt facility at 3205 Alfred Street in Santa Clara, California, which is scheduled for delivery in the first quarter of 2018.

We are pursuing LEED Gold certification for this project, and we also received LEED Silver certification for the most recently completed building at our Franklin Park campus in Chicago during the third quarter.

I'm also very pleased to report that we were ranked sixth on the EPA's top 30 tech and telecom list of green power users, and we ranked 12th on the EPA's national top 100 list of green power users across all industries.

As disclosed in our press release, we took a $29 million impairment charge during the third quarter on 3 underperforming properties to write them down to their estimated fair market value.

These assets are generally redevelopment projects in markets where data center demand has been proven to be comparatively thin, namely Boston and Sacramento.

These assets no longer fit our strategy or investment criteria and are classified as held for sale.

We have a strong track record of success on both development and redevelopment projects, but these 3 assets were clearly not our best performers.

And we are reallocating resources and capital to assets and markets that represent a better opportunity to create value for our shareholders.

We sold one asset during the third quarter, a fully leased power-based building data center in Austin, Texas.

We sold the asset to the user for $20 million at a 5% cap rate and recognized a $10 million gain on the sale during the third quarter.

We sold another asset in Northern Virginia after the end of the quarter, likewise to a user, one of our publicly traded data center REIT peers for $34 million at a 7% cap.

This asset was held in a consolidated joint venture with Equinix.

And we expect to book a $15 million gain on the sale during the fourth quarter, which our prorated share will be approximately $12 million.

Let's turn now to market fundamentals on Page 4. Construction activity remains elevated across the primary data center metros, but leasing velocity remains robust.

And industry participants are mostly adhering to a just-in-time inventory management approach, helping to keep new supply largely in check.

Demand is outpacing supply in most major markets.

The near-term funnel remains healthy.

And demand seems to be picking up as we head into the end of the year.

In addition, vacancy rates remain tight across the board, prompting us to bring unmeasured amounts of capacity to meet demand in select metro areas like Sydney, Silicon Valley and Chicago.

We have seen a flurry of recent land deals in core markets, and a number of new competitors is on the rise, although we believe our global platform, scale and operational track record represent key competitive advantages.

Given the sector's recent history, any prospect of an uptick in speculative new supply bears watching.

However, we remain encouraged by the depth and breadth of demand for our scale, colocation and interconnection solutions.

We expect that demand will continue to outstrip supply while barriers to entry are beginning to grow in select metros, which we believe bodes well for long-term rent growth as well as the enduring value of infill portfolios such as ours.

Now let's turn to the macroenvironment on Page 5. The current monetary, fiscal and regulatory climate remains broadly supportive, and the prospect of tax reform remains intact.

Global economic growth forecasts have improved somewhat over the past 90 days, and most economic indicators have edged up as well.

As I've said before, data center demand is not directly linked to job growth or the price of oil.

However, data is the new oil driving the digital economy, and data wants to be close to the compute engine.

Compute engines reside within the scale data center footprints, and Digital Realty's scale strategy is the prerequisite to enabling these deployments.

Workloads need to be logically close and connected to data lakes, and we are well positioned to connect workloads to data on our global Connected Campus network and through our Service Exchange offering.

Enterprise architectures are going through a transformation, and workloads are transitioning from on-premise to a hybrid, multi-cloud environment.

Our comprehensive product offering is critical to capturing this shift.

Cloud demand continues to grow at a rapid clip, but future growth in the data center sector will come from artificial intelligence.

The power, cooling and interconnection requirements for AI applications are drastically different than traditional workloads, and Digital Realty is well positioned to support the unique requirements and tremendous growth potential of this next-generation technology suite.

With that, I'd like to turn the call over to Andy Power to take you through our financial results.

Thank you, Bill.

Let's begin with our leasing activity here on Page 7. We signed total bookings for the third quarter of $58 million, including an $8 million contribution from interconnection.

We signed new leases for space and power totaling $50 million during the third quarter, including a $6 million colocation contribution.

The weighted average lease term on space and power leases signed during the third quarter was 9 years.

Our third quarter wins showcase the strengths of our combined organization, as the bulk of our activity was concentrated on our collective campuses in Ashburn, which is not only the largest and fastest-growing data center market in the world, but also the combined companies' largest metro area in terms of existing capacity and ability to support our customers' growth.

Although the lion's share of our bookings landed in North America, 2 of our top 5 wins during the third quarter were with leading Chinese cloud and Internet service providers sourced by our Asia Pacific sales team.

We closed 6 fairly sizable transactions with customers from China during the third quarter, including 3 direct deals with existing customers, 2 direct deals with new logos and 1 channel transaction with an existing customer.

We were quite pleased to be able to support the growth of these magnet customers' respective footprints here in the U.S.

Our enterprise team notched several strategic wins, including a new logo win that brought a global provider of food, agricultural and industrial products to our Chicago campus.

We also continue to support the growth of an existing customer, a provider of data protection solutions for businesses and IT professionals, with their expansion into a third location within our portfolio.

Our network team continued to support the growth of a wired and wireless communications infrastructure customer on our Dallas campus, in addition to ecosystem-rich wins with global cloud and network service providers in our New York Internet Gateways.

We added a total of 36 new logos during the third quarter, and we signed over 200 leases for space and power.

Including interconnection, we completed more than 900 transactions.

We continue to build and refine our partners and alliances capabilities to extend our reach to serve enterprise customer demand, both directly and through various leading IT services and cloud service providers.

Through our alliance program, we delivered an integrated solution for a second site deployment for a leading multinational networking and telecommunications equipment manufacturer during the third quarter.

We're also continuing to expand our capabilities and footprint with cloud service providers.

Through our Connected Campus and Service Exchange interconnection platform, we are enabling customers to establish direct private connections to the leading infrastructure as a service, PaaS and higher-value SaaS offerings globally from our data centers on a connected network with a single-user interface.

These capabilities help our customers accelerate deployments of hybrid cloud solutions and open new opportunities to serve enterprise customers, either directly or through our partners and alliances programs.

As we discussed on previous calls, we see substantial opportunity for our partners and alliances programs to create meaningful upside for our customers and our business over time through a set of referral sell to and sell with programs.

Our current partners and customers see the value of our focus on working with quality solution and service providers to create a comprehensive IT solution offering.

In terms of integration, activities related to the DuPont Fabros acquisition are well underway.

We have established a comprehensive communication program for the team members who joined Digital Realty at closing, and we have executed a customer outreach plan to embrace all legacy DFT clients.

At the property level, we are completing security system upgrades to ensure that each of the acquired properties meets Digital Realty standards.

We're also performing system audits and updating operations, policies and procedures.

The systems integration is progressing well and the migration of accounting, finance and HR-related applications is expected to be finished by year-end.

We will continue to update all customer-facing material.

And we'll sunset the DFT logo and marketing materials during the fourth quarter.

We continue to anticipate that the vast majority of integration activities will be wrapped up by the end of the second quarter 2018.

Turning to our backlog on Page 8. The current backlog of leases signed but not yet commenced stands at $106 million.

The step-up from $64 million last quarter reflects the $50 million of space and power leases signed, along with the $59 million backlog inherited from the DuPont Fabros acquisition, offset by $67 million of commencements.

The weighted average lag between third quarter signings and commencements improved to 4 months.

Moving on to renewal leasing activity on Page 9. We retained 86% of third quarter lease expirations.

And we signed $66 million of renewals during the third quarter, in addition to new leases signed.

The weighted average lease term on renewals was over 6 years.

And cash rents on renewal leases rolled down 3.8%, primarily due to 2 sizable above-market leases that were renewed during the third quarter, one on the East Coast and one in Phoenix.

These are the 2 markets we have previously called out as the primary culprits of above-market rents in our portfolio.

These are not -- these are fortunately not the last 2 above-market leases in the portfolio [does both mix] of renewal leases signed in any given quarter could push our cash mark-to-market into the red.

On balance, however, we expect cash re-leasing spreads will be positive for the fourth quarter as well as for this full year 2017.

We continue to see gradual improvement in the mark-to-market across our portfolio, driven by modest market rent growth and steady progress of cycling through peak finished lease expirations.

In terms of our third quarter operating performance, overall portfolio occupancy improved 170 basis points sequentially to 90.8%, partially due to the inclusion of the stabilized DFT assets.

However, we also registered positive net absorption within the stand-alone Digital Realty portfolio, notably in Dallas, Chicago and the Bay Area.

And same capital portfolio occupancy also improved by 30 basis points sequentially to 90.2%.

The U.S. dollar continued to soften during the third quarter and finally rolled over relative to the prior year, as you can see depicted on the chart at the bottom of Page 10.

As a result, FX represented roughly a 30 basis point tailwind to the year-over-year growth in our third quarter reported results from the top to the bottom line, as shown on Page 11.

Same-capital cash NOI growth was 4.1% on a reported basis for the third quarter and 3.9% on a constant currency basis.

Core FFO per share grew by 5% both on a reported and a constant currency basis.

Core FFO per share came in $0.03 ahead of consensus as well as our own internal forecast.

And as you may have seen from the press release, we are raising guidance by $0.03 at the midpoint, reflecting the flow-through from the beat.

In terms of the quarterly distribution, we still expect the first half will represent roughly 51% of the full year results, while the second half is still expected to contribute roughly 49%.

This does imply a step down from the third quarter to the fourth quarter at the midpoint of our guidance range.

As you can see from the bridge chart on Page 12, the primary driver is a full quarter with a higher share count outstanding, following the close of the DFT acquisition late in the third quarter.

We still expect to realize approximately $18 million of annualized overhead synergies.

And we still expect the transaction will be roughly 2% accretive to core FFO per share in 2018 and roughly 4% accretive to 2018 AFFO per share.

However, these synergies will not be fully be realized until 2018, and the quarterly run rate is expected to springload in the fourth quarter before bouncing back in 2018.

With respect to AFFO, I would like to highlight again this quarter the long-term trend in straight-line rent.

As you can see on Page 13, noncash straight-line rental revenue has come down from a run rate of $23 million in the fourth quarter of 2013 all the way down to less than $2 million this quarter.

Over that same time, quarterly revenue has grown by 60% from $380 million to more than $600 million.

This trend reflects several years of consistent improvement in data center market fundamentals as well as the impact of tighter underwriting discipline, which has driven steady growth in our cash flows and sustained improvement in the quality of our earnings.

We had another particularly active quarter on the financing front, with numerous steps taken to further strengthen our balance sheet over the past 90 days, which we have itemized here on Page 14.

In late July, we issued 2 tranches of sterling-denominated bonds with a weighted average maturity of 10 years and a blended coupon of just over 3%, raising gross proceeds of approximately $780 million.

In early August, we prefunded a portion of the DuPont Fabros acquisition with the issuance of $1.35 billion of U.S. dollar bonds with a weighted average maturity of 9 years and a blended coupon of 3.45%.

This was only the sixth time an investment-grade U.S.-listed REIT has issued $1 billion or more in a single tranche of bonds.

The transaction was well oversubscribed and priced 10 basis points inside of where our existing bonds were trading on the secondary market prior to the transaction.

Needless to say, we were pleased with the solid execution and the strong investor interest in the Digital Realty credit story.

We also raised $200 million of perpetual preferred equity at 5.25%, an all-time low coupon for Digital Realty and the lowest rate ever achieved on a REIT preferred offering with a crossover rating.

In mid-September, we closed on the DuPont Fabros acquisition and exchanged all the outstanding DFT common shares and units for approximately 43 million shares of DLR common stock and 6 million OP units.

Also, in conjunction with the DuPont Fabros acquisition, we exchanged the DFT 6 5/8% series C preferred for a new Digital Realty series C preferred with a liquidation value of $201 million.

We also tendered for the DFT 5 7/8% high-yield notes due 2021, settled nearly 80% of the $600 million outstanding at closing in mid-September and redeemed the remainder within a few day post closing.

Subsequent to quarter end, we redeemed all $250 million of the DFT 5 5/8% high-yield notes due 2023 at a blended 106.3% of par, or a total cost of $270.5 million, including accrued interest and the make-whole premium.

When the dust settled at the end of the quarter, debt to EBITDA stood at 6x and fixed charge coverage was just under 4x, as you can see from the left side of the chart on Page 15.

However, that's a bit misleading since all the debt related to the DuPont Fabros acquisition was on the balance sheet as of September 30, but the third quarter only included a 17-day contribution from the DFT assets.

Adjusting for a full quarter contribution, the balance sheet actually improved as a result of the DuPont Fabros acquisition, and debt to EBITDA dips down below 5x and fixed charge coverage remains above 4x, as you can see on the right-hand side of the chart.

Finally, as you can see from the left side of Page 16, we have a clear runway with nominal debt maturities before 2020 and no bar too tall in the out years.

Our balance sheet remains well positioned for growth, consistent with our long-term financing strategy.

This concludes our prepared remarks and now, we will be pleased to take your questions.

Denise, would you please begin the Q&A session?

(Operator Instructions) And your first question will come from Jonathan Atkin of RBC.

I wondered on the international front, if you could talk a little bit about Europe and any sort of update on Frankfurt as well as the momentum you're seeing with retail and interconnect in the acquired Telecity assets, as well as over in Japan with Mitsubishi joint venture, and how you intend to staff and brand that?

Would that be kind of its own operating entity?

Or would that be DLR branded in some fashion?

Jonathan, this is Andy.

Maybe I'll take the first part, touch on Europe, and then I'll hand it over to Scott to give you an update on the Mitsubishi JV.

So over -- I think the first part of your question was with regard to Frankfurt.

I had the opportunity to spend some time with the team in Frankfurt a handful weeks ago.

I was able to tour our new scale campus that is quickly coming online, which I think is going to be a great flagship-scale product offering in that market.

I know we are in advanced discussions with one of the top 5 cloud service providers to serve as an anchor customer for that site.

So that will be a great win.

And then also, I spent some time with the team operating the colocation site, which was one of the 8 properties we acquired a summer previously.

And we've had great wins there on the colocation, interconnection front with a fairly strong lease up to that specific asset's occupancy.

Broadly speaking, I think Europe in general had a fairly strong first half.

We had a comparable anchor win to our campus on the scale side.

In Amsterdam, we had growth from a U.S. specialty cloud provider in Dublin, and we had some great wins on the colo interconnection side as well in some of our London assets.

I think we'll -- we had a little bit lighter quarter in the third quarter, largely due to some of the timing of inventory coming online, but I do expect to see a pretty strong fourth quarter on the Europe front.

And I'll hand it over to Scott to talk about the Mitsubishi JV.

Jonathan, on the Mitsubishi JV, that's going to be cobranded, but it will be very clear that it's a joint venture between Mitsubishi and Digital Realty.

We're very excited about this opportunity.

It's going to allow us to further serve our customers in this important Japanese market and specifically, access to the Tokyo market.

As you can imagine, it's a long process to go it alone there.

We have -- a lot of our international customers have immediate needs in Japan, which we would like to be able to serve.

And this really gets us up the curve a lot faster on that.

We get to bring our global customer base, our global platform and operations development capabilities.

A balance sheet obviously, they bring a strong balance sheet as well as local presence.

Land positions and relationships that will help us both address the market in a much more timely and effective manner for our customers.

And then finally just -- is there anything post the end of the quarter that you could provide any granularity on in terms of preleasing in the portfolio?

No new news to post you on the handful of days we're into October.

But I think you can expect an update from us somewhat close to mid-quarter on our Investor Day coming in early December.

And the next question will come from Jordan Sadler of KeyBanc Capital Markets.

Bill, in your prepared remarks, you touched on either the demand picking up heading into the end of the year.

And Andy, you just ran through some of the strength.

It sounds like your -- that Europe could be a nice pace setter for next quarter.

I'm trying to figure out on a pro forma basis with DFT, what the new bogey is in terms of leasing for DLR on a quarterly basis.

And I know you don't really give guidance around that, but should we expect, if demand is picking up, that the 4Q could be as good as the volume we were able to produce in 3Q?

Thanks, Jordan.

This is Dan Papes.

I'll start with an answer to that.

I just characterize it this way, that the acquisition and -- sorry, the merger with DFT has certainly stirred up interest from our -- their customer base and our customer base.

And we're optimistic that that's going to be helpful, but at the same time, I don't want to -- we don't really commit to a run rate or anything like that.

I would just say that we feel like the demand in the fourth quarter is healthy.

And from what visibility we have into the first quarter, it's healthy as well.

We're going to just have to see how we execute against the pipeline of opportunities that we have.

And you'll see that unfold as we're able to close transactions.

If anybody wants to...

Jordan, what I would add, too, is that the deals are definitely getting bigger.

There are bigger deals in the pipeline, but with that comes lumpiness.

So your -- we could have a great quarter in the fourth quarter, or it could be a great quarter in the first quarter.

But there is a lot of potential volume in Dan's pipe for sure.

Okay, that's helpful.

And then I was glad to see that the asset sale pipeline appears to be picking up a little bit.

It seems like you guys are cleaning out some of the slower growth or less -- or assets with less opportunity.

Can you maybe flesh out, Scott, for us what this looks like, in this held-for-sale pipeline in terms of overall volume?

And then maybe what else might be in the portfolio in terms of scale to dispose of?

Yes.

Sure thing, Jordan.

So these were -- we went through and looked at our portfolio and tried to identify some of the other assets that would be considered noncore or noncore markets.

And we came up with these additional assets, 4 assets that we're putting into the held-for-sale category.

We're also conducting regular reviews on our portfolio.

So it's a little premature to tell you what the volume might be of other ones.

So if you can give me a few months to work through that.

We're expecting that these assets will probably trade somewhere in the first half of 2018, the ones that we have right now, but I think -- I think it safe to assume that we will regularly review our portfolio to cull noncore or underperforming assets.

And the next question will come from Colby Synesael of Cowen.

Two questions, if I may.

Of the 33 megawatts of TKF that was leased in North America in the quarter, how much of that was in legacy DLR facilities versus legacy DuPont Fabros facilities?

And then secondly, the $59 million of backlog that you referenced for DuPont, is that the backlog that they had at the end of their second quarter?

Or is that the backlog that they had as of the last day before the transaction closed with you?

Thanks, Colby, Andy here.

So the -- going to your first question, I would say a large share of our North America signings were in the Ashburn market, and there was some fairly large signings in there.

But there -- it was not -- it was fairly spread as well.

So we had a top 5 cloud service provider sign a multiple megawatt deal, which really was the last major signing to cap off our previous Digital Realty stand-alone campus.

We had the first meg-plus deal from an international customer, sign on to the new Building L. And we also had sizable signings into the legacy DFT ACC complex.

So it was...

Is it fair to assume next year then 20 megawatts in -- of that 33 came from the DuPont Fabros facilities?

I would say the -- I can tell you all the signings -- new signings in the legacy DFT assets happened really just post the shareholder vote by both companies.

The largest signing I think was about 15 megs on a single lease, and -- but adding up the total customers, it's probably approaching that 20 number from one customer.

Great.

On the backlog?

Your second question, that was the 2Q backlog.

So did they sign anything between the second quarter close and when the deal with you closed that's worth calling out?

There were no new signings between 6/30/2017 and the shareholder vote.

The next question will come from Simon Flannery of Morgan Stanley.

I wonder, Bill, if you could expound on your prepared remarks about AI, and in particular, just help us understand what you see as the kind of the market opportunity there, and how the timing -- when does that really start to impact your volumes in terms of orders and revenues?

Well, we've already seen some demand for it.

So it's -- what I'd say, we're in the first half-inning of a 9-inning game there.

And our take is that it's going to be -- as I said, the demand will be substantial, and I think the power requirements will be significant as well going into this.

And I think we'll see a lot of this from our existing customer base.

Dan, you might want to expand on that.

I'm sorry, was there another question from you, Mr. Flannery?

Yes.

I don't know -- was Dan going to follow up on that?

No.

Bill, I think, sufficiently answered the question, Simon, unless you have a specific follow-up.

Yes, on the power requirement, is that something that would might require a kind of a denser build in future phases?

Or do you think you can accomplish that within your existing architecture?

I think we can accomplish more power density within our existing build.

Simon, this is Andy.

Just to add on to that, I mean, we're seeing a range of customer density demands, and it's not always dictated by specific industry verticals for that matter.

But you may have enterprise customers that wants a lower density in terms of power per square foot.

And you want -- may have other customers that are driving that well into the 200 watts per square foot.

And I believe we cleared our build to be able to bring incremental power to the floor as needed and respond to the customer demand.

And I think this AI trend is going to further those boundaries.

And the next question will come from Frank Louthan of Raymond James.

Talk to us a little bit about the DFT model for over the products set.

Are you looking to expand that model possibly through some of your other campuses?

How do you look at that?

And then as a sort of follow-up, there are a couple of markets DFT was in, particularly I think in Canada, that are a little bit new for you.

How has preleasing trended in those markets during the quarter?

Frank, this is Andy.

So on a model, so pretty much anything really in flight in terms of near-term deliveries or last legs of expansion are pretty much dictated by the existing DFT design.

So you can expect that for our most recent deliveries, just wrapping up in Santa Clara, the next builds of our ACC 10 phase 2 and similarly in Chicago.

When it comes to new swaths of available capacity, be it just down the road from the ACC complex and our latest land acreage in Ashburn, we are creating Building L, which is the first phase of 36 megawatts.

But that entire site could have capacity to call it 250 megawatts.

And we have greater ability on those land parcels to provide various build types depending on the customer demand.

So we think being able to have the flexibility to offer various designs to capture a broader amount of customer demand is a competitive strength.

And having that with the acreage and power capacity in core growing markets is also to our advantage.

And your second question was really, I think, speaking to cross-sells and some of the newer expansions.

Toronto in particular, we've seen great interest from several customers, legacy DFT customers that had looked to expand with a megawatt or so.

We've also seen interest from legacy Digital Realty customers that really had not done any business with DFT, but have been looking for capacity in Toronto where we were essentially out of inventory and fully leased and now looking to expand.

So we think there are substantial cross-sell opportunities here, and Toronto is just a one-off example of something that will, down the road, pay dividends.

Yes.

I think you can add -- this is Scott.

The -- Toronto is kind of a land-constrained market, so the supply is somewhat limited.

So I think that will help us in our leasing there.

Frank, it's Dan, too, just as the leader of the sales organization, I'm very happy to have inventory in Toronto for our customers because that's something that our customers have been expressing interest in since I've been here, similar to my being happy to have inventory down in Frankfurt for customers who have been asking for it for some time.

So it's good.

The next question will come from Dave Rodgers of Baird.

So DFT, I think, negotiated a lot of their super-wholesale deals at the various highest level of the executive pool.

I think you mentioned it was a pretty simple lease structure.

Do you guys have a plan in place to kind of go attack those same customers, to Andy's point of maybe changing your build structure to some extent?

So is there a sales plan?

Is there kind of a lead structure that you're contemplating to go after that group of tenants?

And as a follow-up to that, can you update us maybe on the Facebook leases that you inherited?

So in terms of executive interaction with the customers, we have something that we call an executive sponsor program here at Digital.

So we have accounts that are assigned to members of our executive leadership team for coverage in addition to the day-to-day sales team.

So we'll have 2 or 3 accounts assigned to individuals on our leadership team.

I'm sorry, what was your other question?

I can help out.

The other one on the design side.

I think building a larger footprint buildings with larger data halls is something we've done on many occasions.

I think it's important to remember the most significant departure between us and DuPont from a design standpoint was really the way the UPS is designed.

We have typically built a battery UPS system, and they've done a rotary pillar unit UPS system, which, by the way, we have many of those deployments in our portfolio, so it's a matter of preference on that and, of course, we're always reviewing to see which is the best design.

And the second part of that...

I think you -- sorry, go ahead.

Just on Facebook leases was just the second part of that.

Sorry if I interrupted.

Yes, so I think you asked about an update on a major customer renewal.

Obviously, we can't provide any specific details that are confidential to our customer.

We did have a very large volume of renewals.

I believe it was our highest quarterly renewal quantum in the history of digital.

But we did not -- I can tell you, there were not any major renewals executed during the quarter.

And can I just add to Bill's answer, Dave, in regards to -- I think your question was something around senior executives being involved in transactions and bringing them to closure.

We have become much more active in that way as a senior leadership team in the field.

Our -- myself and my peers and Bill spend, I would say, daresay, much more time with clients now than we did a year ago.

And we also did -- one of the things that we saw from DuPont Fabros was they were very effective in that.

And we're -- as we look at best practices from DuPont Fabros, we saw that as a success factor that we feel like it is important for us to leverage.

So we are doing a lot more of that.

I think it's a great question, and something that we're focused on.

The next question will be from Michael Rollins of Citi Research.

Just first, I was just wondering, if you went back to Slide 7, if it was possible to break out the dollar impact of the leasing between the heritage Digital business versus DuPont in the quarter?

And then just second to that, you mentioned a lot of the activity was concentrated in the Ashburn market during the quarter.

And do you think that was just an issue of where customer decisions were getting made?

Or is there some share shift that may be happening in some of the other markets?

Or is it even possible to just -- what's in the available-for-sale category?

Just some insights on the concentration maybe versus historical diversity of bookings would be helpful.

Sure, Michael.

This is Andy.

Let me see if I can tackle -- I'll do it in reverse order.

So last quarter was a very hot Dallas quarter.

We had multiple customers with multiple megawatt signings.

This quarter was a very hot Ashburn quarter.

And I can tell you, Ashburn is a very competitive market -- player, most competitive.

But we had many wins in Ashburn during the quarter from various types of buyers, cloud buyers, other international companies and other technology companies.

I'm not sure -- there is obviously very well-known merits to why customers choose Ashburn, why this quarter versus last quarter or the quarter preceded made Ashburn the winner.

I'm not sure there was anything that pointed to that specifically.

I think it is in a market that has demand from the most -- the most amount of industry verticals.

You have top cloud companies, network companies, enterprise customers all seeking to deploy there.

And this was obviously a big winning quarter for us on all of our 3 major locations in Ashburn.

I think your second question was really breaking out how much of the third quarter signings landed in the legacy ACC construct.

I think I touched on this with -- in response to Colby's question.

I believe the largest signing was roughly 15-ish megs and the second-largest was 7 megs, both of those into legacy DFT, ACC complexes.

And then I think after that, there was a 4.5-megawatt signing into the last building on the legacy Loudoun Digital campus, and then over 1.5 megs landed in the brand-new building now on our latest Digital Realty campus and the largest capacity for growth.

I would say it's not all Ashburn.

We did have deals ranging from 300, 200 kilowatts to close to 1.5 megs also in San Francisco, down in Santa Clara, in Chicago and Dallas.

It's just that the predominance of winnings was in Ashburn this quarter.

The next question will be from Vincent Chao of Deutsche Bank.

I was just curious.

It seemed like you guys were making a special point about the demand or the signings that you had from some of the larger Chinese cloud Internet players.

And I was just curious, are you seeing significant pickup in interest from that region in those types of tenants, such that we might see a bit of an uplift over the next couple of quarters from that category?

Vincent, it's Dan.

Thanks for the question.

I'll tell you the scenario that we're excited about the opportunity.

We did capture 3 inbound opportunities from -- as you heard from Chinese cloud service providers.

And we don't believe they're done.

We believe that it could be the start of something very important for us.

So we're on it and pursuing it and hopeful that it will continue to provide us some meaningful opportunities for growth.

I would not say these are anomalous transactions.

I think that there will be several more in the future, and we want to go capture those.

Got it.

And Bill, maybe going back to something you said in your opening remarks about rising barriers to entry in certain markets.

I mean, it seems like land availability is becoming a problem for certain markets.

But I was just curious if you could provide some more color on what you meant by -- what barriers exactly are you referring to?

You're right.

I mean, land is significant.

It's a significant issue in Silicon Valley.

It will be an issue in Loudoun.

I think that's why you're seeing this sort of mad rush to acquire land because it's -- there's a -- the availability of suitable sites is definitely depleting.

And then in certain markets, power is an issue.

It's not a permanent barrier, but it certainly can be a multiyear barrier.

I want to go back to the -- for your first question.

The other thing that we've done that is a change is that we now have a Mandarin-speaking salesperson based in Hong Kong in our Hong Kong office.

And so we're covering the Mainland Chinese companies from that office with this person, and that has clearly worked out well.

That person had prior relationships with the Chinese Internet companies.

The next question will come from Richard Choe of JPMorgan.

Great.

Just a follow-up on that quickly.

Where are you seeing more of the activity from the Chinese companies?

Where are they looking?

And it seems like we're hearing more about the Chinese cloud players coming to the U.S. So just kind of trying to get a sense of where they're interested in booking space.

Richard, so it's actually fairly broad-based.

We had signings in Ashburn.

We had signing in Santa Clara, we had signings in Chicago and then some smaller signings in Atlanta and Portland.

So it's actually been pretty broad-based.

The -- well, obviously, I think the larger signings are more aligned with our campus locations.

And many of those signings usually land with 1 meg or so, with the customer quickly coming back for -- looking to expand.

Great.

And I guess you've been very consistent with the 57% to 59% adjusted EBITDA margin.

With DuPont, should we expect that to go to the higher end as things get integrated?

Or it -- are enough things happening that you're investing in that range is kind of the range we should be looking for?

My guess is that with the quarter, really a full quarter and 0.5 month of the DFT transaction under our belt, we will be guiding slightly to the higher end of that range of our guidance.

And obviously, it should be more helpful as we move into 2018, given the operating efficiencies from that portfolio.

The next question will come from Jon Petersen of Jefferies.

I just wanted to ask about -- you did a number of kind of one-off dispositions and acquisitions.

I'm kind of curious -- some more color on, I guess, the ease of selling stabilized data centers on a one-off basis.

I think historically, we've seen more of M&A focused on large portfolio transactions.

I guess, is it becoming easier as the industry kind of normalizes to sell one-off buildings?

And then specifically on the 2 dispositions you did, the one in Austin had a 5% cap rate and the one in Sterling, Virginia at 7%, just curious for some more details on what the discrepancy on those cap rates would be.

I would think a Northern Virginia data center would trade at a lower cap rate, just given the strength of that market.

Curious about some more details specifically on those buildings.

Yes.

Sure, Jon.

And I'll take the 2 assets first and I'll then go to the other one.

But Austin, we sold it to -- as you know, we sold it to the existing customer there, and that was a separately negotiated transaction.

And we think we got a good outcome on that from a cap rate perspective.

Virginia -- and by the way, as a general statement, I think cap rates are probably not the best way to always look at these.

But if you look at Virginia, that asset was a joint venture with Equinix.

They were interested in disposing of that asset as well.

If you look at it on a whole picture of all the metrics and the price per foot, the quality of the asset, the amount of office in the relative data center, if you look at all of those components on it, you will see that it's pretty good outcome from a valuation perspective.

So the cap rates can be a little misleading on those two.

The ease of selling one-off assets in markets, I would say if you had a good one-off asset in a core market, they're probably still pretty easy to sell.

You get outside of the core markets, and they can get a little bit more difficult to sell.

Because then the size of them will determine who the potential customers are that might -- or the buyers that might buy that.

So it does get a little more -- a little more difficult in -- Sacramento is a good example.

We're going to pair an asset that we've held for sale with another existing well-leased asset in an effort to kind of get a little more bulk there and attract some other buyers.

Okay.

And I guess as a follow-up on that same point, I'm just kind of curious which type of building is commanding a higher cap rate -- or I'm sorry, a lower cap rate today, kind of a power-based building leased to like a credit tenant on a long-term basis or just kind of your traditional multi-tenant wholesale data center building?

Yes.

I think that PBB, the long-term lease PBB kind of net lease assets are still easy to buy and a little lower on the cap rate spectrum, just given there is a universe -- a much larger universe of buyers.

You can buy a lot of -- find a lot of passive buyers who would clip coupons.

So the multi-tenant data centers, you have to consider what you're going to do from an operational standpoint.

But I will also say, as a general statement, there's been -- while those cap rates used to have a few hundred basis point spread between them, they -- the spreads get compressed quite a bit.

And in some cases, they are kind of right on top of each other.

The next question will come from Jordan Sadler of KeyBanc Capital Markets.

Not sure if you touched on this yet, but I was wondering if you could offer an update on the COO search.

Sure, Jordan.

I can do that.

We engaged Heidrick & Struggles to do the search.

I think that we're -- we've -- they've certainly sourced a number of highly qualified candidates.

I've met with several, as have members of the management team here.

So I think we're making good progress.

But I think what's important here is that we're going to take as much time as we need to, to find the right addition to the team.

And I feel comfortable and confident that things are being handled appropriately in the interim.

We have Chris Sharp, who's now part of the leadership team, in charge of his old area plus product and SEs, and then the ops team -- the data center ops team is reporting in to Andy.

So things are in good hands while we pursue the search.

Okay, that's helpful.

And then just on the interconnection side during the quarter, it looks like you added 1,400 cross-connect sequentially.

Was that entirely organic, so no contribution from DFT whatsoever?

Correct.

No material contribution from DFT.

And the last question will come from Lukas Hartwich of Green Street Advisors.

Can you provide some color on market rent growth across your portfolio?

Lukas, Andy here.

Happy to do that.

I would say if you look at -- maybe start with North America rental rates in the Ashburn market, where we're seeing the largest amounts of demand, but also the largest amounts of competition and supply are relatively flat.

Conversely, if you go to the other coast to the valley, we're continuing to see a modest uptick in rents, certainly on a year-over-year basis really due to the supply constraints in that market, limited capacity coming online by ourselves and our competitors.

And there's much less competitors kind of in between Chicago, Dallas.

But maybe kind of put them certainly in between those 2 bookings for you, flat to slightly rising rates.

Broadly speaking, I think despite the robustness in the overall volume of demand, the fact that supply is rising up to just try to intersect with that demand is something that kind of puts a bit of a lid on rates from spiking.

And you also have a phenomenon where the buyers are buying in bigger and bigger quantities, hence, commanding better pricing on each of those buys.

We remain optimistic that these rates stay flat to slightly increasing.

Kind of going -- leaving the U.S. for half a second, I think we're seeing similar trends in our major campus-oriented markets, be it Frankfurt, Ash, London, Amsterdam and Dublin and same thing in Singapore and Sydney.

That's really helpful.

And then one last quick one.

The straight-line rents that they've been "declining over time." What -- can you remind us what the main driver of that is?

Two drivers.

One is just obviously just natural as you move through the lease, the delta between cash and GAAP rents changes.

But I think the more material one is really a firming up over several quarters, not just a quarter-over-quarter change, but several quarters and even probably years of the market where really less and less incentives, free rent or ramps have been given away as the market tightened and came out of trough.

And hence, we were able to command firmer pricing and less incentives.

That's more of a better quality of our earnings.

And at this time, we will conclude the question-and-answer session.

I would like to hand the conference back over to Bill Stein for his closing remarks.

Thank you, Denise.

I'd like to wrap up our call today by recapping our highlights for the third quarter, as outlined here on the last page of our presentation.

First, we closed on the acquisition of DuPont Fabros, a high-quality, highly complementary portfolio concentrated in top tier U.S. metros in accretive, prudently financed transaction that expands our relationships with a blue-chip customer base.

Two, we delivered solid current period financial results, beating consensus estimates by $0.03.

And we raised the midpoint of our full year guidance by $0.03 as well.

Last but not least, we further strengthened our balance sheet by refinancing DuPont Fabros' high-yield debt with attractively priced long-term capital and exchanging the DFT common shares units and preferred stock for DLR common and preferred equity.

We finished the quarter with debt to EBITDA below 5x and fixed charge coverage above 4x, both pro forma for a full quarter contribution from DuPont Fabros.

As I do every quarter, I'd like to conclude today by saying thank you to the entire Digital Realty team whose hard work and dedication is directly responsible for this consistent execution.

Thank you all for joining us, and we hope to see many of you at our Investor Day in December.

Thank you, Mr. Stein.

Ladies and gentlemen, the conference has now concluded.

Thank you for attending today's presentation.

At this time, you may disconnect your lines.