Digital Realty Trust, Inc. (DLR) Q3 2013 Earnings Report, Transcript and Summary

Good afternoon. My name is Theresa, and I will be your conference operator today. At this time, I would like to welcome everyone to the Digital Realty 2013 Third Quarter Earnings Call. [Operator Instructions] Thank you. Mr. John Stewart, you may begin your conference, sir.

Thank you. Good morning, and good afternoon, everyone, and welcome to the Digital Realty Third Quarter 2013 Earnings Conference Call. The speakers on today's call will be CEO Mike Foust; CFO and Chief Investment Officer, Bill Stein; and SVP of Sales and Marketing, Matt Miszewski. By now you should have all received copies of our press release and supplemental. In addition, we posted a presentation to the Investors section of our website, which is designed to accompany management's prepared remarks. You're welcome to download the presentation and follow along throughout the call. Before we begin, I'd like to remind everyone that management may make forward-looking statements on this call. Forward-looking statements are based on current expectations, forecasts and assumptions that involve risks and uncertainties that could cause actual results to differ materially. Such forward-looking statements include statements related to the company's future financial and other results, including 2013 guidance, preliminary 2014 guidance and the underlying assumptions. For further discussion of the risks and uncertainties related to our business, see the company's annual report on Form 10-K for the year ended December 31, 2012, and subsequent filings with the SEC. Additionally, this call will contain non-GAAP financial information. Explanations of such non-GAAP items and reconciliations to net income are contained in the company's supplemental operating and financial data package furnished to the SEC and available on the company's website at digitalrealty.com. Management's prepared remarks will be followed by a Q&A session. [Operator Instructions] And now I'd like to turn the call over to Mike Foust. Mike, would you like to begin?

Thank you, John, welcome to the call everyone. The highlight of our third quarter for Digital Realty was our solid leasing activity. Our business has taken a turn for the better and healthy momentum has picked up across each of our regions. Data center demand is growing in the second half of the year, and we're quite pleased with the leasing velocity. In fact, as you can see on the first page of our presentation, the dollar volume of leases signed represents the third highest quarter in the company's history and year-to-date activity already exceeds the full year of 2012. New leases include a healthy mix of new corporate logos, as well as expansions of existing customers to new facilities. As shown on Page 2, the third quarter was equally strong in terms of lease renewals. I'm sure most of you saw last week that we completed 5 early renewals with Equinix. This was a mutually beneficial transaction, providing long-term operational certainty to a strategic customer. And from Digital's perspective, these leases contributed to the positive mark-to-market on renewal activity during the third quarter. Cash rents on turn-key leases rolled down slightly, but represented an improvement over the past several quarters. We believe that rumors of data center rent roll downs have been greatly exaggerated. We estimate the mark-to-market across our Powered Base Building portfolio at 6% -- over 6% actually and the cash mark-to-market across our Turn-Key data center portfolio is roughly flat. We are encouraged by the recent trends in pricing, as well as the leasing velocity. Despite the leasing highlights, third quarter financial results were lower than expected. Bill will walk us through the drivers behind our revised earnings guidance in detail during his presentation. But suffice it to say that primarily the combination of the noncash straight rent -- straight-line rent adjustment, delayed lease commencements and the near-term dilution from the joint venture we announced earlier this month has led us to revise our guidance for 2013 and to reset expectations for future periods. We do believe that our revised outlook represents a realistic assessment and puts us in position to provide more accurate lease commencement forecasts going forward. Turning now to Page 3 of the presentation. We are encouraged by the pickup in leasing velocity that we've seen in the second half of the year, and pricing has largely stabilized relative to last year. Furthermore, the solid backlog of leases signed but not yet commenced represents contractual obligations for future rental revenue and sets the stage for healthy growth and cash flows over the intermediate term. As you can see from this chart, the backlog of signed leases translates to an incremental $52 million of NOI in 2014 and additional $38 million in NOI beyond that. We're also pleased to announce today that on Monday this week, we signed a 7-megawatt lease with a social networking company that had outgrown its existing colocation footprint. This one deal represents an additional $11 million of annualized GAAP revenue, and is tangible evidence that the momentum we saw during the third quarter is clearly carrying through to the fourth quarter. The healthy backlog aside, we are taking proactive steps to address the lumpiness of our leasing activity, as well as the signing-to-commencement cycle. And Matt will walk through some of these steps in detail during his presentation. In terms of supply and demand, as shown on Page 4, we currently are tracking an estimated 3-megawatt shortfall nationwide. Silicon Valley is the one major market in the U.S. at risk of significant oversupply, whereas demand has been particularly strong in Northern Virginia, as well as in Dallas. Amsterdam, London and Frankfurt are currently the healthiest markets in Europe, and supply-demand dynamics are generally favorable across AsiaPac. Over the past 2 years, Digital Realty has made significant investment in developing additional inventory to meet growing demand, and this accounts for some of the new supply you see represented in the chart on Page 5. We expect that this new inventory in our major -- active major markets will be largely absorbed by corporate enterprise and cloud services demand over the next 12 to 24 months. In the near-term, however, the delivery of this inventory will continue to weigh on our return on invested capital. Lease up of existing inventory and driving improved ROIC are the top priorities across our organization. We fully expect to generate attractive returns on these new completions as we lease up our inventory and maintain our focus on consistent improvement on return on invested capital. In the near term, new development starts will be more measured and lease up of existing inventory will likewise be our primary source of near-term organic growth. In late September, we formed the previously announced $369 million joint venture with Prudential's Core fund at an attractive cap rate. Aside from our return on this investment, which Bill will cover in his remarks, we believe that this transaction was an important milestone in terms of demonstrating the appeal of data centers from the perspective of a sophisticated Core investor. I would like to point out that growing acceptance of this asset class is a bit of a dual-edged sword. On the one hand, it has very positive implications for the private market value of our portfolio. On the other hand, lower return expectation on the part of Core investors do make it more difficult for us to achieve external growth through acquisitions of income-producing properties. Consequently, while we remained active and will continue to underwrite almost every deal and portfolio in the market, our near-term investment activity is likely to be more consistent with our pre-2012 history, until some of our expectations readjust consistent with our return requirements and our investment alternatives. During the third quarter, we introduced Digital Open Internet Exchange, which is running in conjunction with the Open IX initiative spearheaded by the Internet community here in North America. In short, we've committed to provide the data center environments for the new open Internet exchanges so that they can operate across all the targeted U.S. markets, for Open IX, as well as international exchanges that are expanding currently into the U.S. Our facilities will provide the foundation for a successful initiative, and we're very excited to be a part of it. On a related note, the Digital Realty Network ecosystem continues to develop on schedule. And we are now seeing new leasing opportunities with specific network requirements that our ecosystem and Open IX, allows to solve for our customers. It's too early to quantify the near-term financial impact from these initiatives, but we believe they will significantly enhance the appeal of our properties to a wider range of tenants and, in turn, our ability to generate future NOI growth and enhance the value of our portfolio over time. And with that, I would like to turn the floor over to Matt Miszewski for a deeper dive on our sales platform and leasing momentum.

Thanks, Mike. Hello, everyone. Let me jump right into the sales activity for the third quarter. It was a strong quarter for lease signings, and we continue to see positive momentum. In fact, this was the third highest quarter for lease signings in the company's history. Our year-to-date signings are about 40% higher than at this point last year, and as you see here on Page 6 of the presentation, already surpass full year 2012 activity. As of today, we are at over 91% of the highest signings year in the history of the company with just over 2 months left in the fiscal, and all regions provided positive signings momentum in the third quarter across North America, Europe and Asia Pacific. And while this quarter's performance was solid, the future also looks bright. Our pipeline continues to see accelerating growth in qualified deals in North America, Europe and APAC with both healthy domestic and international components. We currently have over 2 million raised floor square feet requirements in our fiscal year 2014 pipeline. In addition, over the past 90 days, we've implemented a number of lead generation and nurturing tools to continue to build our already healthy pipeline. In our first test of this system over the past few weeks, we are seeing solid metrics supporting our new content marketing, inside sales and digital demand generation programs, including impressive open rates on new offers, healthy new lead generation statistics and encouraging lead conversion rates. In addition to this pipeline activity, we also closed a significant number of mid-market deals inside the quarter, including a rapidly growing cloud services provider in SingleHop [ph] in the Midwest, taking 1,500 kilowatts in our Franklin Park facility, as well as a well-known online travel advice company taking a total of 1,400 kilowatts across multiple locations inside our portfolio. In our colocation business, we continue to see mid-market penetration in the quarter, including one of the globe's top enterprise private cloud providers, as well as the country's fastest-growing mobile cloud platform company in one of our Boston facilities. We also continue to add new logos at multiple properties and in multiple segments. Our unique build-to-suit project with ARM Holdings highlighted our commitment to environmental innovation, while our new signings with a global leader in virtualization technologies highlight our global platform as they took space in both the Dallas and the San Francisco markets. We continue our new logo success in the financial services vertical with a national provider of diversified financial products in one of our Chicago facilities. And we also landed new leases with a true global software and cloud powerhouse in Australia and with one of the world's largest manufacturers of industrial engines, Cummins. I'd now like to cover some of the organizational adjustments we have made in our global sales and marketing program since the beginning of the year. Programs I'll describe for you in a moment, including our mid-market and global alliance strategies, are expressly designed to shorten the signings-to-commencement cycles, as well as addressing our premium pricing support. In addition, adjustments to our incentive compensation programs will focus on targeting existing inventory clearance to drive improved return on invested capital and continue to shorten commencement time line. Turning now to Page 7. With regard to our new global alliances program, our customers are asking us to provide them with insightful guidance across all of their needed data center services, not simply space and power. By providing them with a comprehensive global alliances marketplace, we will be able to satisfy our customers' requirements, while providing additional revenue opportunities to some of our closest, existing and exciting new partners. As pressures increase on our customer base, they need their data center decisions to be easier, and we hope to make Digital Realty their one-stop shop for a data center advantage. While Digital will continue to focus on its core strategic advantages, we will leverage our new partner ecosystem to provide high demand services, including managed services, cloud services, network services and more. In addition, we will be providing guidance to our customers from their perspective, initiating a vertical go-to-market process across multiple industry verticals, ranging from financial services, cloud services and manufacturing to energy, health care and retail. By focusing on our customers' needs as the primary focal point of our sales and marketing efforts, we believe we'll be able to more accurately fit their requirements and jointly create increasing value across the multiple segments from lead generation and nurturing into a growing level of customer loyalty, ultimately resulting in a solid renewal rate amongst our base. As shown on Page 8, we have also developed a new focus on the mid-market segment as an area that holds great promise in terms of shortening our sales cycle, keeping stable our pricing advantage, compressing the average time to commencement and broadening our installed base across our global platform. We have also begun developing muscle in addressing the small and medium-sized business segment, with a lower cost-of-sale model that will allow us to continue serving this segment by leveraging a more aggressive demand-generation program and a scalable new sales force. We are adding significant sales and marketing resources to successfully execute this multi-segment strategy. While we increase our focus on these 2 segments, it is important to point out that our large enterprise segment continues to perform incredibly well. Demand from that segment drove the healthy signings volume in the third quarter, including multiple purchases from one of the largest financial services firms on the planet now in multiple countries. We have also continued to see large expansions from global cloud services providers and have landed large enterprise logos as well, including a leading Japanese automobile manufacturer. We expect that this segment will continue to be strong for Digital. We simply view the mid-market and small- and medium-sized business focus to be additive to our longstanding large enterprise program. The longer sales cycles in large enterprise accounts can now be balanced with a more complete program to address the mid-market and small- and medium-sized business segment. We expect that this segmented strategy will continue to close the gap between lease signings and commencements and help to smooth out revenue from quarter-to-quarter. And as highlighted earlier, we also saw mid-market deals close across multiple verticals and across every geographic market we serve, including multiple new logos and deals that will commence in a much shorter time frame. And now I'll turn the call over to Bill.

Thank you, Matt. Good morning, and good afternoon, everyone. Let me start by addressing the straight-line rent expense adjustment that Mike mentioned in his remarks. We booked a $10 million noncash true-up in the third quarter because we discovered that we had not previously remeasured the straight-line rent expense related to one of the handful of leasehold interests in our portfolio when we executed a 10-year early extension and modification of this leasehold in September of 2010. The $10 million adjustment that we booked during the quarter represents a catch-up of the noncash straight-line rent that should have been recorded from the fourth quarter of 2010 through the third quarter of 2013 at a run rate of approximately $830,000 per quarter. The entire prior-period adjustment has been included in reported FFO, whereas only the current period of $830,000 has been included in core FFO for the third quarter of 2013. Let me emphasize that the straight-line rent is always and everywhere an add back to FFO in the calculation of AFFO. So there is no impact on AFFO, regardless, for any period involved. Since the annualized straight-line rent expense adjustment represents an immaterial percentage of net income in each of the affected periods, we believe that it is not material. Consequently, we currently believe that it is unlikely to give rise to a material weakness. We believe that our internal controls are adequate. We have since reinforced our existing policies and implemented supplemental policies to address this specific issue. We believe that we've taken the right steps and have fixed the problem. If we can better help you understand the nature of the issue or its resolution during the Q&A portion of today's call, we'll be happy to do so. Moving on to our results. We reported third quarter FFO per share of $1.10 and core FFO per share of $1.16 compared to the consensus analyst estimate of $1.20. Starting on Page 9 of the presentation, let me walk through the various components and attempt to bridge the gap from our third quarter results to the $1.22 we reported in the second quarter as a proxy for the consensus estimate. The first item to note is the $10 million straight-line rent expense adjustment for 111 Eighth Avenue in New York. The entire $10 million adjustment, or a little over $0.07 per share, is included in reported FFO. In contrast, only the current period portion of $830,000, or less than $0.01 a share, is included in core FFO for the third quarter of 2013. Higher property taxes costs us a little over $0.01 per share this quarter, primarily due to additional county personal property taxes assessments in Santa Clara and in Texas. We believe these personal property taxes estimates represent double, and in some cases potentially triple, taxation since the counties charge personal property taxes but do not deduct personal property from a real estate assessment, and our tenants are taxed on their personal property as well. We will, of course, appeal these aggressive assessments, but in the meantime, we have accrued for the higher rates. The second quarter also included $0.04 from a gain on an insurance settlement, as well as an extra $0.01 of catch-up from the first quarter for the change in capitalization policy announced on the last call. Finally, G&A accounts for $0.02 of favorable variance in the third quarter since our independent directors are awarded a fully vested equity grant in the second quarter of each year. Adjusting the future run rate for each of these items gets you to the clean quarterly run rate of $1.14 to $1.15 or $4.60 annualized. This represents the base from which future growth forecasts should be framed, in our view. Moving on to Page 11 and turning now to our outlook going forward. We are revising our 2013 FFO per share guidance to $4.60 to $4.62, down from $4.73 to $4.82 previously, and revising guidance for 2013 core FFO per share to $4.65 to $4.67, down from the prior range of $4.74 to $4.83. The primary variables relative to our prior forecast include $0.07 per share for the straight-line rent accrual on 111 Eighth, $0.01 per share due to the lower-than-expected acquisitions. The joint venture with the Prudential Core fund is $0.11 dilutive on an annualized basis, so we will see $0.03 of that in the fourth quarter and another $0.08 next year. Finally, delayed lease commencements account for $0.06 per share. In hindsight, our sales forecasts for recognized revenue have been too aggressive. Our signings have been on target, but revenue recognition has not lined up with our projections. Nevertheless, as Mike pointed out during his remarks and as you can see, if you refer back to Page 3 of the presentation, the backlog of leases signed but not yet commenced does represent an extremely strong tailwind of contractually baked-in NOI that will come online over the intermediate term. We are still in the process of finalizing our budgets for next year, so we are not providing formal guidance for 2014 at this time. But we have outlined the broad brush strokes for next year in the earnings release. Clearly, the pace of our earnings growth has moderated somewhat as our sector has matured and the size of our asset base has grown dramatically. Even so, we expect to deliver respectable mid-single digits FFO per share growth, even on a $13 billion asset base. And now I would like to turn to our funding strategy and capital markets activities. As previously announced, we refinanced our global credit facilities in August, upsizing the line of credit from $1.8 billion to $2 billion and the term loan from $750 million to $1 billion. All-in pricing for the line of credit was reduced from 150 basis points to 130 basis points and all-in pricing for the term loan was reduced from 145 basis points to 120 basis points. In late September, we formed a $369 million joint venture with Prudential Real Estate Investors, as Mike mentioned, at a 6.7% cap rate. Since this transaction has already been announced, I'm not going to dwell on too many of the details here on Page 12. But I would like to highlight 2 numbers, mid-teens and 20, that I think have been largely overlooked with respect to this transaction. Specifically, we achieved an unlevered IRR of over 20% on the sale of our interest to the joint venture and we expect to earn a mid-teens return on our residual equity interest going forward. I also think it's important to note that the fee-stream figured into the pricing of the transaction. Said differently, the cap rate unquestionably would have been lower had this been structured as an outright sale rather than a joint venture. We believe that this transaction represents very solid execution and achieved our objectives of maximizing the menu of capital options available to us, while minimizing the related costs and simultaneously demonstrating the appeal of our data centers to a core institutional real estate investor. Proceeds from the joint venture have initially been used to pay down short-term debt. And as you can see from the credit stats on Page 13, we maintain a flexible investment grade balance sheet. We do not expect to need to access common equity in 2014, barring any unusual portfolio or M&A transaction activity. We do, however, expect to raise preferred equity sooner rather than later, and may also revisit the corporate unsecured bond markets as well to take advantage of what we believe is an attractive low rate window. The front loading of these anticipated long-term capital raises is one of the drivers behind our below consensus outlook for next year. But we believe it is the prudent course of balance sheet management. I would also like to point out that our exchangeable debentures are expected to convert into common equity in April of next year and that conversion will improve the debt-to-EBITDA by 0.3 turns. As indicated in yesterday's press release, our Board of Directors recently authorized a $500 million share repurchase to give us the flexibility to opportunistically buy back stock when the public market values our portfolio at a meaningful discount to the private market value. At present, however, we believe the double-digit yields that we are achieving on development still represent the best risk-adjusted returns available to us. In closing, I would like to draw your attention to several enhancements that we've made to our quarterly supplemental package in direct response to feedback from the investment community on our disclosure practices. As you can see on Page 32 of the supplemental, we've added a new schedule with details on our land holdings, including acreage, as well as our cost basis. In addition to the leasing activity typically disclosed, on Pages 29 and 30, in our traditional schedule, we've modified the format and added the following new metrics: the initial stabilized cash rent per square foot for new leases; the change in cash rents on renewal leases; tenant retention; and rolling 12 months in addition to the current quarter's activity. The presentation that we posted on the website to accompany today's call is a further step in the direction of improving our communication with The Street. We aim to continue to improve the transparency of our financial disclosures over time and we welcome additional input from analysts and investors in that process. This concludes today's prepared remarks. I would now like to turn the call back over to the operator and we will be pleased to take your questions. Operator?

[Operator Instructions] First question comes from the line of Jordan Sadler with KeyBanc Capital Markets.

So my first question is really regarding sort of the guidance and how it is sort of not necessarily parallel to the leasing that was done during the quarter. So maybe just drilling down on leasing a little bit, can you maybe characterize what you think the leasing -- what will happen to the leasing volume over the course of the next 5 quarters? You mentioned the momentum continued in 4Q, but is it expected to slow materially next year?

The signings are not expected to slow, but what we've been experiencing is a lag in commencements from signings, and that's really a function of the large enterprise customers. They're taking significant amounts of space. And there's a natural, I'd say impediment, to their ability to absorb all that space immediately based on just their needs. So that's taken down over a period of time. And that's the point of that backlog slide, is a lot of this business has been signed up but it will not commence for as much as 6 quarters away.

That's right. So especially with the large tenants, like the one we signed earlier this week, where they're taking multiple pods, they'll phase in 1 or 2 pods at a time and that may take place over a 12- to 18-month period. And so then you have a delay in commencement for each of those phases as they lease in. But the income is there as you can see in that NOI build up. So the asset value and future -- near future cash flows and NOI will go up pretty significantly.

Can you think about bracketing what NOI growth might look like for next year?

Well, if you look on Page 3 on the accompanying slide deck, we break that out for you, and I think in a conservative fashion. This is our -- and this is our backlog. So this slide represents actually signed leases and the contractual backlog that is already baked. So new leasing will be on top of that.

Okay. So, right, so this is no -- this $952 million for 2014 includes no new leasing?

That's correct.

But that's not necessarily your assumption?

Of course, not.

That is correct. That is not our assumption. That is the minimum baked contractual without any of the good leasing that Matt and his team are doing.

Okay. And then in terms of the guidance, you said that you're looking to front-load some of this issuance? Any of the other levers where there's been sort of a conservative treatment so we could understand where the potential upside is here. Because obviously, being 7% below consensus with this initial guidance, it seems a little bit shocking to everybody, the stocks are reacting in that fashion. But it does seem like you guys have reset the bar somewhat. So any other sort of leverage you'd point to where you've tried to take the conservative route besides the sort of, that front-loading the debt?

Well, there's a reduction in capitalized interest that reflects the reduction in capital expenditures. But to the extent that leasing demand merits it, there will probably be additional CapEx towards the end of year or there could be, which would affect the capitalized interest. Right now, there's about a 20% reduction in capitalized interest. I think the other thing to keep in mind, Jordan, is that the JV adversely affects the -- 2014 to the tune of about $0.03 a quarter and obviously, that's in 2013, but it's not in the next 3 quarters of 2014.

Next question comes from the line of Robert Stevenson with Macquarie.

On Page 34 of the supplement, you guys list about $561 million of construction in progress. How much of that is going to need to be spent in 2014? And if I'm understanding your comments from earlier, you guys are intending to fund this with debt and preferred, or are you going to need a second JV or some sort of other equity in order to fund that?

I think that most of that's going to be spent next year. And we have enough liquidity on our revolver to fund that without any additional equity. You'll see in our guidance that we're assuming a preferred issuance. On the JV front, we are assuming up to $400 million of additional JV proceeds. But we don't need a JV to fund that construction program.

Okay. And then, Mike, if you guys were to do another JV, is it also likely, given what you've demand-wise, to be Powered Base assets? And so sort of can you contrast institutional investor demand in a JV format for Turn-Key assets versus Powered Base given the sort of longer-term triple-net leases of those assets?

If we were to do another tranche with a core investor, likely it would be -- it would have some Turn-Key in it but the majority would be Powered Base, likely. It's potentially that we would look at, at other types of investors who may be more core plus, who would look at Turn-Key assets. But right now, we have a limited amount -- number of properties that we would even contemplate at this point putting in a joint venture, and they'd be more geared toward core with a little bit of Turn-Key.

Rob, let me add to that. The way we see ourselves using private capital next year would be to supplement our acquisitions program and enhance the returns on our retained equity. So if you think about the Prudential JV that we did with an 80-20 structure, just think about applying that to a future stabilized acquisition where we retain 20%. But instead of a market cap rate that is, say, call it 7-ish, just for the sake of argument, but it could be under 7, through fees we may be able to boost that into the low double-digits, if not higher, as we did with this JV. So obviously, making the opportunity far more accretive.

Next question comes from the line of Michael Bilerman with Citi.

Manny Korchman is on with me as well. Bill, just to start with you, can you just sort of go in a little bit more on the development side just so we're really clear as to what's happening currently and into the future? You talked about $800 million dollars of projects that are expected to deliver in '13 at 10% to 12% yields. And I want to understand how much of that will be in the numbers by the end of the year in terms of that run rate? But then the disclosure for 2014 talks about $600 million to $800 million of spend, not necessarily what's getting delivered. And with about $561 million left to spend on the current development pipeline, that would indicate that you're really not starting any new -- or much new development next year at all. And I want to make sure we're thinking about that the right way.

I'll take the second question first. So the capital that's highlighted for 2014 is the spend in 2014. But it's also likely to be close to deliveries as well. There may be some carryover into 2015, but predominantly 2014. And you are right. Most of what's been highlighted for 2015 is either to complete existing projects or it's to honor commitments that we've made to customers for leases that they've signed. So it's really to finish projects that have signed lease commitments. And as for the capital that is in the guidance for 2013 with respect to deliveries, virtually all of that is in the numbers except for whatever is to commence this quarter.

Okay. And then maybe just sticking with commencements. Mike, this is not the first quarter where you sort of had a delay in commencements. If you go back to last year, there was, I believe, 2 quarters where you pushed things out. And I'm just curious what internally from a function perspective are you doing that you can't get ahead of this, because it's a pretty material changes each time in terms of that delay. And while I respect that leasing and the numbers are large, you do have a large portfolio suite, we would hope it would be growing. But why -- is there something internally that was off in terms of tracking this stuff? And I guess, why hadn't you gotten further ahead of it that we're now talking about further delays in that cycle?

That's a good question. And I think part of the challenge is a lot of our customers -- it's not internal methodology and how we're leasing. It really has to do with the decision-making process with our customers. And these large customers, for the most part, this is their first outsourcing and the first of many. So we're working with them in a pretty close partnership. And frankly, in many cases, what their time frames for when they want to complete a lease and they want to get up and running, it has slid more than they thought, more than we thought. And clearly, we haven't been conservative enough in putting in more wiggle room in the time frames for when these deals get actually signed. And then that gets compounded a little bit because on these larger deals, the ramp up, when someone is taking 3, 4, 7 megawatts, you have typically an 18- sometimes even a 24-month ramp on that long-term lease. So I think it's that combination of really spearheading this growing trend for outsourcing, as well as the ramp up. And as Matt went into detail, our mid-market and SMB program, we look to bring in more customers in that middle range, where they'll take space and commence space much more closer to signing.

Yes. And Mike, this is Matt. As you think about the topics that Mike just discussed, it's a natural maturing of the market. When we initially entered the large enterprise segment, we saw a little bit tighter lease-to-commencement time line. And we're starting to see that large enterprise segment mature and take a little more time in their deliberative process as they move forward. But the maturing of the market, especially in the mid-market accounts, is starting to bring that same pattern that we initially saw in the large enterprise now in multiple transactions in the mid-market as well. So we expect to be able to bring -- tighten up the spread between lease signing and commencement inside the mid-market space the same way that we initially did in the large enterprise space.

And I think what will also continue to help that acceleration, the fact that we're doing more and more with existing customers as well. We're expanding in multiple markets, multiple continents, and those tend to be more good size, more bite-size solutions where they are commencing much more closely to lease signings. So I think we'll see that tighten as we do more and more of these expansions with existing customers.

Mike, I'm going to add one more item to that, which is to the extent we are doing business with a customer that hasn't outsourced before, their process may not be as well defined and refined as some of the existing customers that are accustomed to outsourcing. So there can be fits and starts in this process and it could be -- and it is, frankly, a situation where the fellow with the company that's got the data center requirement's talking to our salesperson, and he is absolutely sincere that the requirement needs to be filled within a reasonably short period of time, and our customer -- and our salesperson believes that, too. But then it gets kicked around inside these companies to several different groups and several different levels. And you find that instead of 1 RFP, there might be 2 RFPs or even 3 RFPs, as the requirements go from IT to corporate real estate and back.

But the good news is we are getting these deals signed and we have a good backlog, and that backlog is going to grow significantly over the next 1.5 years -- 1 year, 1.5 years.

The next question comes from the line of Stephen Douglas with Bank of America.

Two, if I could. First, I'm interested in how you're thinking about the pacing of the buyback and kind of what the plan is there to balance that with the leverage? And then within the 2014 guidance, to what extent does that contemplate any buybacks? And then second, on the pricing front, I see the new disclosures on per kilowatt basis for the new leases signed. And I guess I'm wondering is there any kind of context you can put around kind of what the pricing for maybe last quarter would have looked like on that basis.

Well, I'll answer the buyback question. There is no assumption with respect to a buyback in the outlook that we've given for next year. And as I said in my remarks, we consider the returns on our development deals, the 10% to 12% unlevered, to be superior to what the return would be in repurchasing our common stock. Now another point, quite frankly, is that we have -- basically, our capital is earmarked for our development program at this point. So our ability to repurchase common would be a function of raising alternative equity. As to the leverage, we obviously are very keen and focused on maintaining our investment-grade debt ratings. You saw on the statistics that we provided that we are in good shape in that regard. But I think if we were to embark upon a common stock repurchase program, it would be funded with either another preferred issuance or potentially a joint venture.

And in terms of the lease rates as reflected for the per kilowatt, probably last quarter and this quarter are pretty consistent. I think we've seen pretty consistent pricing between the 2 quarters.

Next question comes from the line of Steve Sakwa from ISI Group.

I guess, 2 questions. I guess, first, on kind of the leasing front, if I think about the returns, Mike, that you're sort of talking about and Bill, this 10% to 12%, I guess, it's clear from kind of the way the numbers are penciling out now and the growth going forward that those returns are clearly not being achieved in kind of year 1 of these deliveries, which may have been sort of the old expectations. And I guess, I'm really trying to figure out is this a situation where year 1, these are 4% yields. Year 2, these are sort of 8% yields. And maybe by the third year of stabilization, these are hitting that 10% to 12%. Is that how we should be thinking about it?

Well, yes, when you look at a project, typically it's going to be generally a 36-month process to reach stabilization. Sometimes it might be 40 months. So on an individual lease basis when we do our analysis lease by lease, we fully load all the costs and carry costs into that calculation on a lease-by-lease basis. And then that translates when we're stabilized on that asset, like any development project. In year 3, generally, you're going to be at stabilized or very close to, and that's where you're going to achieve your 10% to 12% unlevered returns.

Okay. I think maybe part of the problem with estimates is people are giving you kind of these returns kind of day 1 when the project delivers and, clearly, it's taking kind of multiple years to get there, which I think just may be a modeling flaw. I guess, secondly, I just wanted to come back to the kind of the leasing, and you talked about the signings, but obviously, the commencements are falling short. And I guess I'm just trying to think through if a large office tenant kind of took a lot of space and somehow was phasing into it, or got free rent in effect, they would still start the lease kind of day 1 and would recognize ultimately the free rent as kind of a free rent payment. But it seems like in this case, the tenants are phasing into their space and the recognition isn't really starting until they are taking that down. And I guess I'm just trying to think what's different about this versus kind of maybe an office lease and kind of why wouldn't you be starting the leases but then have a larger free rent period?

That's a really good question, and most of the time, what happens with these larger deployments is that because we're building pod by pod, the rent will commence on the pod that's built and being utilized by the customer. So in those cases, it phases in as they occupy pod-by-pod, as opposed to free rent, where customer is taking the entire facility and is getting free rent up front. So it's a little different nuance and the accounting treatment requires you to phase in pod-by-pod for those multi-pod customers.

Just to add to that, Steve, so we would -- we commence upon both the tenant taking possession and Digital's obligations to the tenant in the construction process being met or fulfilled. So as Mike said, if there are multiple PODs, we need to finish the second, third, fourth POD, whatever it is, in order to commence the rent.

And that reflects -- that phasing reflects the fact that for large customers, they just physically -- it's just physically not possible to take down and populate a multi-megawatt data center in 6 months or even a year depending on the size. It's just not feasible for anyone.

And Steve, that's part of why we made a decision to focus an increasing amount of effort on new market opportunities, because while we see the large enterprise exposure as something very positive to the long-term health of the company, we think that in the mid-market space, where the sizings for an ideal customer land is around 1 POD, it fits more directly with being able to commence within the fiscal.

And your next question comes from the line of Vance Edelson with Morgan Stanley.

So another question on underestimating the potential delays with the more complex customers. How do we know this is just a timing issue that surprised both you and perhaps the customer, versus any actual slackening of the need on the part of the tenants or at least the delay in their needs versus just the delay in implementation? So how do you know there's not some sort of temporary pause that some of your tenants are seeing in their own needs that's causing them to delay?

That's a good question. And there's 2 -- actually 3 items that give us comfort that demand is really growing and the trend for outsourcing for both enterprises and cloud providers is growing. And first of all, one is the -- that Slide #3, that actually shows you the contractual backlog. I mean, those are signed leases that are commencing over the next short period of time. And over on top of which, new lease revenues will be added. Also on top of that are the signings and the growth in lease signings. So those represent real deals, real requirements that are moving ahead on a contractual basis. And I'd say third is our funnel of prospects that we're addressing directly today continues to grow at a pretty good clip over the last 2 quarters. So all those things give us a lot of confidence that the market for our facilities in our major markets is pretty healthy.

Got it. Okay. And then lastly, regarding the straight-line rent adjustment, I guess what I'm not clear on is why was this one building ever treated any differently than the others? And how do we know there aren't more out there that could require adjustment?

Most of our buildings are owned in feet. So this is 1 of 4 buildings where we have a leasehold and then we have another 7 ground leases. And I think what triggered this -- I mean, the reason we discovered this was that there was a cash bump in the modification in 2014. So cash rents paid didn't change from the original lease term through the initial period of this modification until 2014, which was when the original lease was scheduled to expire and then we added another 10 years to 2014 at an increased rent. So that's how it was discovered. And it's just -- we just don't have that many of them. So our team is obviously very accustomed to processing leases where we are the lessor and our customer is the lessee. Those leases, there are procedures that address that situation and the individuals in asset management, property management deal with those situations every day. This was a different situation. The procedure is still there to deal with the leasehold, but our individuals -- this is one of just a handful of situations where you would have an extension that needs to be reflected in our books and records appropriately, which it was not, and we admit that. We understand why it happened and we've taken additional steps to make sure that something like this doesn't happen again.

Your next question comes from the line of Michael Knott with Green Street Advisors.

Two leasing-related questions to start. First, do you think the delayed commencements further suggests that tenants have the advantage here in negotiating and setting the economic terms of data center leases? So that would be one. And then two, couldn't one argue here that lease signings as a data point just lost a lot of value as a leading indicator for your business? And if it's potentially a false-positive like it seemed to be in this case.

No, I don't think so, Michael. Because if you don't have signings, you don't have future revenues. And the signings are very tangible, contractual-based cash flow that one can underwrite. So I don't think that's the case. And as we mentioned earlier, with the outsourcing being relatively new to a lot of these large customers, it's really the decision-making process and the fact that it is so important, they're entrusting critical facilities to Digital Realty, that it just has taken a longer time than our customers have contemplated and certainly than we estimated. I think we're in a much better position now to estimate and project when these leases will commence. But it certainly does not reflect a reduction at all in demand, and actually, our demand statistics and deals that we're tracking and pursuing have actually increased with Matt's group.

Yes, and Mike, I think that on your question about who has the advantage on economics, we certainly do see a differential from market-to-market, and we don't see a connection between delayed commencements in the large enterprise segment and sort of tenant advantage in terms of negotiation. We haven't seen that come through in the -- in this fiscal year and we don't anticipate that coming through in 2014.

And your next question comes from the line of David Toti with Cantor. David Toti - Cantor Fitzgerald & Co., Research Division: I hate to ask another question on lease signings, but we understand there's pretty significant pressure in the market for tenants broadly to sign new leases even when the lag in signing and commencement is long. Are you concerned that at this point you're pulling forward future demand that could result in weaker lease signings in '14 to '15?

Yes. So David, thanks for the question. We do see that pressure. We don't believe, given the 4 quarters looking forward pipeline, that we're looking at, that we're pulling forward deals into either Q3 or into Q4. We have a very healthy pipeline throughout the next fiscal year. David Toti - Cantor Fitzgerald & Co., Research Division: Okay. And then as part of that, and maybe I missed this earlier, but I'm having trouble connecting, I guess, the occupancy declines that we saw in the quarter to the sort of historically high signing. And, obviously, that's a function of the gap between the lease signings and the commencement. Do you expect that to close at some point given the sort of high volume in the quarter?

Yes, definitely. I mean, we've seen a bit of a blip, in that we've delivered completed Turn-Key space into inventory that will be leased over the next year, as well as we had a couple of office leases in non-data center buildings that expired and we're at the very beginning now of re-leasing those in Fremont, California.

One other thing I'd like to mention is that the 2012 third quarter same-store occupancy data included occupancy from the JV assets, which are obviously not in the Q3 2013 since we've -- those are now owned by the venture. I think in the future, we'll be adding a footnote to the disclosure to make that clear. But that was -- that contributed to the decline as well.

Yes, those were all 100% leased buildings.

Right.

Next question comes from the line of Jonathan Schildkraut with Evercore.

If I may, first, in terms of the commencements and backlog, and I do appreciate the incremental disclosure on Page 3, but could you give us what the commencements number was for the quarter from a revenue perspective? And the backlog, what it is from of revenue perspective? So that way we can sort of tie it to the leasing activity and also the backlog that you provided on prior earnings calls. And then secondly, if I can sort of take us off the leasing path for a minute. I'd like to talk about your interconnected data centers and that initiative, and just give a little sense as to what the services that you're going to be providing? Are you just go to be providing dark fiber or is your intention also to deliver lit services?

For the commencements in the quarter, we're looking at a little over $32 million of annualized GAAP rent and a backlog of almost $91 million of annualized rent.

And could you spread that out for us Mike? Last quarter, you gave us what was expected to commence for the remainder of this year and then '14 and then '15 and beyond. Can you spread that $91 million for us?

Well, we have the actual lease rental revenue realizations that will be contributed on Page 3. We can probably provide something further on that. I don't have that table available right now, but I'm sure we could have more definition around that on kind of what the future revenue streams. Because these only reflect on Page 3 what is going to be realized in that particular year and not the long-term.

We can give you that data at our Analyst Day in a couple of weeks.

And also, Page 3 here is also NOI, it's not revenues. So we're really taking it down to the actual operating income at the property level for you in that, getting down to the bottom line. So obviously, those revenue numbers will be much larger, representing the future cash flows from each of those leases that will be commencing. You also asked about the network ecosystem, the Digital ecosystem. And that initiative is moving ahead actually quite well. We expect to have our major metros' dark fiber installed between our assets and back to the peering points, by the end of this year, January, February of next year. So we're very close to having those networks completed. And in some cases, providing lit wavelengths for customers that they can utilize as well. What this really allows us is for customers to utilize fiber themselves. It also, probably more importantly, allows the different carriers, metro and long-haul carriers, to land efficiently and effectively in our buildings to serve the customers in those buildings, and tied with the Open IX initiative, really allows all of our buildings in a metro to be connected directly with the major peering points and interconnection facilities, whether they're in a digital facility or whether they're in another location. And that's proven very effective already in customers in London, Silicon Valley and Virginia. And we expect that's going to really create more tenant demand from much wider range of customers than ever before.

Your next question comes from the line of Tayo Okusanya with Jefferies.

So first question just on pricing trends. When I do look at the pricing data for leases signed during 3Q, especially on a per square foot basis, it looks kind of lower than what we've seen historically. Just kind of curious whether that's all geographic-mix-related this quarter or was that more of a power density mix that's kind of making those numbers look somewhat lower than historical levels?

Yes, I think if you look on a per kW basis, they're pretty consistent quarter-over-quarter. I think some of it's market mix as much as anything, and the fact that some of our tenants are taking not necessarily lower power densities, though in some cases that is true. But they are taking ancillary space along with the data centers. So the per square foot number was a little bit lower because they're taking staging space, a little bit of office space and it all gets wrapped into that lease.

Okay. That's helpful. And then the second thing, just when I take a look at the 2014 guidance and the capital raising activity, I'm still trying to get a sense of why the decision to try to do all the capital raising very early in the year, number one. And then number two, the whole idea of once all that's in place, leverage ratios continue to go up. You're still probably going to be below your self-imposed target, but how comfortable are you kind of moving above those targets and adding more leverage to the overall balance sheet?

Well, the reason we're front-loading the longer-term capital, Tayo, is because we're concerned that we're in a -- that rates will go up at some point next year, that the long-term end of the curve will increase. I mean, it's come down about 50 basis points in the last month or so, and we think this is the time to strike before it goes back up. As far as the leverage ratios are concerned, so the preferred brings the leverage down clearly. The conversion of the debentures to equity, as I said, helps our debt-to-EBITDA by about 0.3. And then the growth in EBITDA as well gives us that capacity. So our forecast shows us right around 5.5x, give or take, over the course of the year. And we're comfortable managing to that. I would hesitate to commit to the equity, that we can take the leverage up until we have the appropriate discussions with the rating agencies. So I'll defer on that question.

The next question comes from the line of Bill Crow with Raymond James & Associates. William A. Crow - Raymond James & Associates, Inc., Research Division: Could you give us the spreads on the re-leasing activity x Equinix?

I don't have that x Equinix... William A. Crow - Raymond James & Associates, Inc., Research Division: We're hearing the Equinix is a big number.

On the Powered Base Building, they generally reflected kind of the midpoint of the Equinix. We had a couple that went down and a couple that went up in addition to the Equinix ones. But the ones that went down were not material overall. William A. Crow - Raymond James & Associates, Inc., Research Division: So I'm sorry, if we pulled Equinix out, the numbers would be similar to what you reported?

Yes, I believe so.

Bill, let me go speak to that for one second. We do disclose the Turn-Key renewals on Page 30 of the sup and we disclosed that both on a cash and GAAP basis. And we have Turn-Key and Powered Base. Equinix is all PBB. And I think one point that I'd like to make about these Turn-Key renewals is if you take a look, the Turn-Key cash renewals are down 2.7%. But we show a rolling 12 quarters -- just on a cash basis, we show a rolling 12 quarters down 7%. So if you think about those 2 numbers, that would, to me, indicate a positive trend on lease spreads. William A. Crow - Raymond James & Associates, Inc., Research Division: Okay. And then sort of beat the horse one more time here on the commencements. When you sign a lease, you contractually assume there's a commencement date at some point. And it may be 1 year out or 2 years out. I guess I'm trying to get my arms around why on a very short-term basis, it would cause you to have to miss guidance in the quarter, for the quarter. I guess that's what I'm trying to get my arms around.

Sure. So part of it is leases getting signed late in the quarter or getting signed in the next quarter. And frankly, we did not forecast conservatively enough to take into account kind of that slippage in timing. So while the signings are good -- are very good, we definitely miscalculated how quickly those deals would get signed within a quarter and if they slip from 1 quarter to the next.

And your next question comes from the line of Gabe Hilmoe with UBS.

It's Ross Nussbaum, here with Gabe. Two questions. The first is what are you anticipating the incremental operating expense and G&A spend might be with regard to the expanded effort to target mid-market and smaller tenants. I mean, how much -- how many more bodies do you have to put in place to execute on that initiative?

Yes, so we've taken a look at the amount of increase that we would need on the sales portion of that expense. And the number of heads is roughly 1/3 growth over where we are in the space today. So we are very wary of not trading too large a burden on our SG&A, but we do have to adjust significantly in order to meet the market.

Do you have a rough dollar estimate as to what that does from a dollar perspective to the payroll?

No, not yet, because we're also in the midst of a market that has -- is resetting its salary requirements. So we don't have the detailed information right now, until we get done with the -- with acquiring the asset.

Next question comes from the line of Vincent Chao with Deutsche Bank.

Just wanted to go back to the development commentary for next year, and just it sounds like the incremental dollars' best spent on development. But if I heard the earlier discussion correctly, it does sound like 2014 spend is mostly winding down the existing pipeline. I'm just trying to marry that with the very positive commentary we've been hearing all call on the signings and just the overall demand. Just why would the starts be -- why would there be limited amount of starts if that was the case?

Well, we're in a good position of having sufficient inventory now in the major markets that we've targeted. And so we've gone through the last couple of years really ramping up our development program to meet that demand. And now going forward in 2014, 2015, 2016, we're in a really good position to lease up that space that we've been developing. And so we're ready to accommodate those customers because we have made the investment in our inventory.

Okay. And then just the second question, just on the commencements and some of the sales initiatives that you're implementing, which sound positive. Just curious though, it sounds like the process itself is just simply more complex and that's causing things to be delayed, that's more for the larger users. So is the expected improvement in that commencement timing, is it more just the mix of the type of leases where you're getting some more of the mid- and SMB-type customers, which just naturally brings that down because they're easier deals? Or I'm just curious how the sales force really can impact that decision-making process if it's simply the complexity of the deals that are being undertaken.

I don't want to suggest that they're necessarily the easier deals, but the structure of the companies that we're targeting in the mid-market and SMB space means that there's a higher demand in those markets. So they will have -- they will be incented internally to commence quicker because there's a larger amount of external demand in their business space. And then the sizing of the individual deals makes it sort of a perfect fit for our POD architecture to be able to serve in relatively short time frame. So instead of taking out 8 megs, 10 megs of opportunity, if you're taking out an 11.25, we can roll that in relatively quick.

I want to go back to Ross Nussbaum's question. Just to clarify on one point. So the incremental -- a substantial amount of the incremental compensation from the additional sales personnel will be capitalized in the form of leasing commission. So base salaries are relatively modest, and most of the compensation to that sales force is success-based. And that shows up as capitalized leasing commissions.

Next question comes from the line of Jon Petersen with MLV & Co. Jonathan M. Petersen - MLV & Co LLC, Research Division: Going back to leasing on your Turn-Key Flex space, the renewals, the weighted average lease term was only 2.9 years this quarter. Just wondering if we should read anything into that? Historically, it's been about 7 to 8 years. Is this just tenants renewing early and just tacking on a few years at the end of their leases, that really don't already expire for a few more years, or is the new reality is that tenants are only leasing 3 or 4 years at a time?

I think -- well, you have 8 deals, so you just don't have that many deals. So I think that could skew it as well. If you look back on a rolling 4 quarters basis, it's 4.5 years with a larger data set.

And you have a few colo deals in there as well that are more of a, by their nature, are shorter term when you have the colo renewals.

That's right. Jonathan M. Petersen - MLV & Co LLC, Research Division: Okay. And then in terms of the delay of commencements, just one more question. Just curious if you guys have the flexibility when you sign a lease and realize it's not going to commence until maybe 12 or 18 months from now, can you guys slow down the development spend on that? Or is it going to get developed at the same time anyway, and then you're just carrying it on your balance sheet until it gets occupied?

Yes, Jon, the structure of the POD 3.0 architecture does allow -- and our normal development time line does allow us to match up the actual deployments with the split of schedules as we know them. So if something is an 18-month planned rollout, we have the ability to rollout at staggard terms.

And when you think about it too, it's not that the whole lease doesn't start for a year. You may get -- oftentimes, you'll get a portion of lease starts immediately or close to immediately, and then it phases in over 18 months. So you'll get portions commencing as you go through that 12- to 18-month time frame.

Next question comes from the line of Art Massimiani with Federated Investors.

My question relates to FFO versus AFFO. Looking at the slide deck, you do a pretty good job of highlighting the new reset FFO rate. Whenever I look at the supplemental and you do adjusted FFO from there. First of all, you don't give guidance on AFFO, is that correct?

That's right.

Okay. So if I look at it and just do a back of the envelope calculation, is the calculation for AFFO for the third quarter, would that be a good run rate, so about $0.89 a quarter kind of starting from this point?

I think that the difficulty in providing guidance on AFFO and just using that as a starting point, relates to capitalized leasing commissions and capitalized leasing compensation, because that's going to be a function of what's signed in the quarter and what's commenced in the quarter. So I would discourage you from using the AFFO line at the very bottom of the page.

Okay. That being said, it looks -- could you then relate the lower guidance to what you guys -- how you guys are viewing the dividend going forward?

Yes, the dividend is going to be a function of taxable income. So -- and I've said on many calls and with many investors that our dividend policy is to pay out 100% of taxable income. And the best proxy for growth and taxable income is AFFO. It's not FFO, because AFFO is cash. Having said that, we revisit our dividend level annually at the next board meeting. And we have a -- when we reported the JV, we indicated that we had a sizable taxable gain on the sale of those assets. And the IRS allows a REIT to in essence take some of the distributions made in a subsequent year and count it towards a distribution in the current year if taxable income comes in higher than might otherwise have been forecasted. So this year, our taxable income, because of the sale of assets, is higher than what we thought it was going to be when we set the dividend last year -- or earlier this year. So the implication of that is that we're going to be pulling some of our January 2014 dividend back into 2013 to satisfy the 100% distribution requirement. So having said that, what that means is that our 2014 distributions are starting out light because you can't count it twice. The only way we can meet our 2014 distribution requirement will be to raise our dividend to a level that gets us to a 100% payout of taxable income. We haven't determined exactly what that is, but that's our situation.

Next question is a follow-up question from Jordan Sadler with KeyBanc Capital Markets.

A quick one, hopefully. On inventory, we've referenced this a number of times in terms of what's available for lease up. In some of your disclosures, obviously, occupancy disclosures by property, you referenced this. But I guess I'm trying to pinpoint a different number, which would just be total available inventory of Turn-Key versus PBB space that are available for lease up today. So completed developments. I know you point us to what's been completed in the quarter, but is there a spot in your disclosures where you have total available inventory for lease up?

Yes, it's on Page 27 of the supplemental. At the very top line for each product type, it says, available. So for example, at the end of this quarter, we had 421,251 square feet of Turn-Key available and 161,000 square feet of PBB available. Do you see that, Jordan, on Page 27?

I do see that. That's perfect. And so anything -- and that obviously excludes anything in the development pipeline?

That's just what's available today.

And your next question comes from the line of Michael Bilerman with Citi.

Two quick follow-ups. One was just can we just get a little bit more specific on the capital drag. If you think about almost $1 billion of capital raise, call it, a midsize rate when you blend preferred and the debt, you are talking about a $0.38 drag before you use that capital. So I'm just curious what you're putting in, in terms of the return on that capital, knowing that you have the $600 million to $800 million of spend and about $200 million of debt maturities. Clearly, the timing of when you do that and when you capitalize will offset some of almost the $0.40 drag of interest expense and preferred expense. So can you just be very specific in terms of what's embedded into 2014 guidance for that? And then the second question relates to this $90 million of NOI backlog. And it would be very helpful to understand what that backlog is attached to. And so there's a handful of things that are out there. You have $800 million of deliveries this year. You have a $1.1 billion current development pipeline. You have vacancy in your core portfolio. I don't know, I can't remember if it includes renewals, I believe, it doesn't. But can you sort of break out where that $90 million is attached to so that we can better understand the dynamics of what's left?

Michael, let me handle your first question. So we are assuming that the preferred and the debt is in place for the full year 2014, for the purposes of the outlook that we gave you. We are not assuming any acquisitions. That capital is being used to retire short-term debt at basically a very low rate given where LIBOR is today. For planning purposes, we're assuming a preferred at 7.5 and a debt at 5%. So a little bit higher than what you have in mind. I mean, if you were to think of something that's a little more normalized, you might say well, maybe we would do these financings midyear rather than at the beginning of the year but then of course rates, we think, would be higher as well. I'm not sure if that addresses your question, but we're not deploying that capital incrementally into acquisitions and really, the dilution is a function of retiring short-term debt and replacing it with what we think is a more prudent capital in terms of its duration.

And in terms of breaking out where the commencements and NOI is being delivered, it's really primarily from -- for 2014, from the product that's been delivered and is work in progress that will be delivered in the first half of the year. So that's primarily in large part from space that's built in inventory or currently under construction.

Next question comes from Michael Knott with Green Street Advisors.

Guys, just curious on your mindset on capital allocation and your cost of capital. I think with today's share price action, this will take you certainly below your own, what I would guess is your own estimate of NAV. I'm just curious how -- even if that discount is not enough to make share buybacks attractive at this point, does it temper your enthusiasm for deploying capital into developments? Just curious how it affects your cost of capital versus capital allocation opportunity mindset?

Well, the capital that's being deployed next year, Michael, is to finish projects that are very near completion. Some of them are physically complete, but we just need to finish the billing, and then it's to build out projects for which we have signed leases. So there's really not much discretion in that capital allocation.

And it's going to generate, we think, still attractive returns relative to other uses of capital.

Your next question comes from the line of Jonathan Schildkraut with Evercore.

I was just wondering in terms of the backlog, are there any contractual outs for the customers that have signed up for these commitments? Are there circumstances where they can push out the start date or they can pull it forward for that matter?

Really, there are no outs, other than in some cases if it's a build to suit, where if our development takes longer than planned then the commencement would happen later. But we don't anticipate that occurring in any cases. And so there are no outs in any of these leases.

And your last question comes from David Rodgers with Robert W. Baird. David B. Rodgers - Robert W. Baird & Co. Incorporated, Research Division: I guess, first on cost of capital, but thinking more from your customers' point of view. I know historically there's been a lot of discussion about oversupply in the market or competitive pressures out there, pricing at below market pricing. But I mean, what's been the impact since, let's say, 2009 and the financial crisis, of your customers' cost of capital getting substantially better and the impact of that on your returns versus kind of some of these other inter-market factors like new supply? And I guess, to tie into that, do you think that pressure, if in fact it's there, is abating at all?

Well, the customers' decision to outsource data center facilities, certainly, they are looking at their cost of capital and they are realizing in many cases, if not most cases, that their investment opportunities are better investing in their own businesses rather than in the expensive data center infrastructure facilities that we specialize in. So it's really where customers set their internal hurdle rates and their internal return requirements. And I think we're seeing almost across the board, except for the Googles and Microsofts of the world, almost everyone else is saying, "Hey, our internal hurdle rates are such that it makes sense for us to outsource." David B. Rodgers - Robert W. Baird & Co. Incorporated, Research Division: Okay. And then as a second question, just to visit a topic I guess that was too often discussed earlier in the year, but with regard to obsolescence of the data space. I think a long time, your average lease term for first gen space that you delivered was about 11 years. It seems that most of the renewals that you're seeing, as you said, are somewhere between 2.5 and 4.5 years in total. That's kind of capping you out at maybe a 15-year total combined lease term, or 15 to 17. I don't know, is there any reason why there's kind of an inability to push above that on some of these second generation spaces? Can you give us any recent examples of combined first and gen -- first and second gen spaces that may be pushing beyond that hurdle to give us more confidence that there's not a big kind of capital spend in year 15 or so?

Well, yes, certainly. I mean, there are 2 things -- I think you're looking at a pretty limited base of customers there from the Turn-Key just because of small number. If you look at combined Turn-Key and Powered Base Building renewals, those have gone out -- the Powered Base Building especially lately have gone out 10 years, 15 years. So those are -- have gone out, in many cases, very far. And what's interesting to know is that our cost on those renewals, even on the Turn-Key this past quarter, our cost per square foot was only $7 on Turn-Key and less than $2 a foot on the Powered Base Building, and -- which means, as a landlord, we're contributing very little capital. And on the Turn-Key, I can tell you, it's all leasing commissions. I mean, really none of it's going into TI or refresh, which really shows how long-lived these assets are, otherwise our customers would be asking us for a refresh capital, and that's just not happening.

There are no further questions.

Great. Thank you very much, everyone, for taking the time today. And I want to thank our team for the hard work and everybody's attention. And I just want to emphasize the fact that our business is healthy, demand is growing and we expect to see really increased returns for our shareholders going forward here. Thank you very much.

Thank you. And that concludes today's conference call. You may now disconnect.