I'm Dave Barden. I head up, Telecommunications and Communication Infrastructure Research for Bank of America in the U.S. and Canada, based here in New York. So thank you all for coming. We're really pleased to have with us today, Digital Realty, and their CEO, Andy Power, a Bank of America alumni, and, Jordan Sadler, a KeyBanc alumni, who heads up IR for, the company. And, so thank you both for being here. Appreciate it. Thanks for having us. So I guess, I mean, I'd like to start out by just talking about the kind of just overall demand profile. We'll get to kind of how we fund the business in a few minutes. But, you know, the stock has done amazingly well, you know, since the summer, you know, big 40% move up. And a lot of it has to do with, I think, a more clear-eyed view of kind of the fundamental base case demand, and then we can talk about some of the upsides to that demand. But one of the more kind of amazing things about the business is that, you know, second quarter, you had almost 7% re-leasing positive re-leasing spreads for the business, and that is a dramatic U-turn from probably the last 10 years. And so none of the AI stuff, that's not part of this. Like, something has really changed in the business over the last, you know, year to 18 months. Let's start there and talk about what is really going on in the data center industry. Sure. So I mean, maybe if I bucket into the two categories of demand, starting with like more enterprise-oriented, less than 1 MW interconnection, which, probably close to 1 year ago, on the Q3 call, we obviously were facing an environment which had economic clouds, potential recession, which does not appear to have come to fruition in 2023. I think I was an outlier and wrong in calling out concerns around that. Bank of America was, too. We've put up pretty strong results in that category, whether you look at our new-logo contribution quarter-over-quarter, or less-than-megawatt interconnection bookings, multi-market, multi-site, and a pretty strong quarter into the second quarter. I would say what we're seeing in activity just less than a month ago, it still remains robust and firm in that category called Hybrid Cloud Deployments. The hyperscalers, again, like, as you described, putting AI aside, have been continuing to grow their footprint in their availability zones, in addition to on-ramp locations, that's been doubled down in existing markets, that's been growing into new markets, and that's been a consistent, robust, and diverse demand source as well. All of this has transpired, after many years of the same song on demand happening in a world where supply's kind of run out of gas from multiple angles, whether it's power, generation or transmission issues, supply chains, environmental concerns, NIMBYism, moratoriums in certain countries or jurisdictions, and, that didn't show up overnight. I mean, we had some major events along the way. Just over a year ago, Northern Virginia and Dominion Power says no more power till 2026. That movie really hasn't improved in terms of their delivery times since then. Santa Clara, on the West Coast, has obviously had, like, a almost 2030, issue on transmission. That pendulum on pricing for both our categories has moved more and more in our favor. On the enterprise side, colocation, you've seen our cash margin mark-to-market adjusted now 4.8% in the last quarter. That's colo versus a trailing twelve-month of 4.2%. If you go back multiple years, they're closer to the 2.5%-3.5% type, mix. And then our greater than 1 MW contribution has had the greatest inflection, certainly anchored or pulled up by the markets that were most impacted. And the example I give, which is our largest market, is Northern Virginia and Nashville, in particular. Those rates had gotten pushed down to the $70s not too long ago, and now in the last quarter, we announced that we've been doing new deals and renewals with various customers in sizable multi-megawatt deployments in the $140s. Those kilowatts or megawatts are incredibly precious and, net net, I think this phenomenon is going to continue for some time because I don't see the demand abating. Pre-AI, which we can talk about, is a, I believe, a rising tide that will lift all ships in the data center industry, and these supply constraints are not dissipating quickly. Mm-hmm. Yeah, I mean... Especially with respect to the power, you know, it's. I like to think of it as there's this kind of geometrically compounding amount of demand for data consumption, but the amount of space and power available kind of grows in a linear fashion. And these two lines, for the first time in the years that I've covered data centers for the last 15 years, seem to have intersected in kind of a non-transient way. Would you agree with that kind of characterization? Power was not by and large, power was not a problem for our business. And now it is as crucial as permitting on land. As I don't know this all at the feet of the data center industry. We have been in a robust economic run for numerous years across numerous sectors, and this is all again happening with the CHIPS Act and other manufacturing, industrial deployments of capital that all pull on the same resources, whether it's power or water resources, or talent pools, or manufacturing slots, and data centers have certainly been a demonstrative piece of this puzzle. But it kind of makes sense that the demand has eclipsed the supply. So- ... kind of thinking it in a P times V type of context. So let's focus on the P, the releasing spreads. So you kind of talked about 4.8% for the sub-megawatt kind of enterprise colo business that, for lack of a better term, kind of the Equinix kind of part of the business. And then but having delivered an almost 7% releasing spread for the total business in the Q2, obviously, the greater than 1 MW is doing a lot better. And that has been a source of volatility historically for Digital Realty, sometimes unpredictable. But you've said that you think that you can assure investors that it'll be at least non-zero releasing for the year. But given that we are running high single digit releasing spreads for the H1 of 2023, why am I not more optimistic than that? Because it sounds like you might have to take your guidance up. We just bumped the overall. We guide on overall, so the mix of both of those numbers and plus 4%. The less than a megawatt colocation, interconnection, rich grade or numerous data points in that pool of activity, as I mentioned, has inflected. Not as great of an extent, but that's been a positive for a lot, much longer, almost forever, type of a category. Listen, we're doing our best handicapping renewals for bigger deals because we don't control when those get signed, right? And they don't show up on the day prior to the expiration saying, "Hey, I'd like to renew today." Right? So we don't know what mix comes in any given quarter. We did, I believe, call out on our earnings call or prior conferences that our Q2, greater than megawatt, had some episodics, about a short-term renewal, like 2 years. That was at a leading the way premium- Mm-hmm ... for that flexibility we afforded the customer. But I think my comment is, it feels more and more firming that in that category, in any given quarter, it's gonna be on the positive side, when I've certainly been prior answers to that question. But it'll build in a lot more wiggle room if you never know when the negative could come through. Mm-hmm. That's because these market rates have run a fair bit and continue to run. Go ahead. Sort of layer in and in the context of your question, you know, the historical volatility associated with the greater than a megawatt bucket is really downside volatility. Right. But we couldn't predict that. I was being polite. Yeah. Exactly. And, and there's been an inflection this year, obviously, in fundamentals, and what you saw in the Q2 was really, you know, that inflection, which is really- Mm-hmm ... you know, the upside volatility. So to your point, what are you going to see in the H2? You're not going to see the same degree of upside or amplitude, but normalization, but on trend in a positive manner. Yeah. So I would imagine that you've got enough visibility into the portfolio and into the maturity of the various contracts. Now, obviously, these companies can come to you at any time and say, "I want to get ahead of this." But the pricing floor for almost every market around the world has come up to a point where, is there any real legitimate risk of a downside surprise, or just not maybe as big an upside surprise as we've seen in the past? We're just dealing with small sample sets in this category. So far, yeah. Right? I mean, we just don't... These are chunky, chunky contracts, so but we could be doing hundreds of renewals in one bucket, and we could be doing handfuls in the other. Mm-hmm. Right? So that's the, that's the only when you zoom in to a quarterly basis, but it does not feel like... It feels like we got a pretty good shot of exiting this year, at the very least, it would've been positive in that category. Okay. So switching to the V side of the equation, I guess. So first of all, I think you guys have maybe cautioned that, you know, just because of a lack of new supply, Digital Realty supply coming to market in the third quarter, that we shouldn't expect necessarily, you know, amazing bookings, because there's no new stuff to sell. But it is coming. I just want to kind of make sure, is that a, like, third quarter isn't necessarily the bookings quarter we should, we should hang our hat on. We've got new supply coming in the pipeline soon. I wouldn't be that cautious on the third quarter signings. I know we got weeks to go. Third quarter, like, will not be the record for Digital Realty, but everything was shaping out to be a pretty darn healthy quarter, in my opinion. And so just in terms of that V, again, can you kind of remind us, what's your vacancy or what's your occupancy rate within the portfolio right now? So the portfolio, the stabilized portfolio is called the mid- to low 80s. And that, I'll speak to that. I think we've made progress in releasing that capacity. And by the way, that same draw pool gets adjusted every time we joint venture or sell anything that's 90, 95, or 100%- Mm-hmm ... leased. So, and the other thing about that pool, we're actively managing this pool with a full spectrum of customers. We often have customers that don't renew from certain capacity, 3 MW-5 MW. And rather than releasing that like for like, we re-productize and move that towards our colocation mix, because it's a much higher, better use and economic return. So that could be an interim hit to occupancy, but in the end of the day, you have much higher ROI or incremental NOI. We've got almost 400 MW of capacity that are actively underdeveloped, shovels hitting the ground, record or are going to open up soon. Just over half of that percent will be leased.... And then beyond that, we have land capacity of, call it, north of 3 GW. That is all, call it, within our control agreement. Let me go back. So these are things I wanted to kind of check through. So the 50% pre-leased within the 400 MW that's in the development pipeline, would you say that that's a normal percentage? Is that a higher than normal percentage? Is that drifting northward as people try to take larger deployments sooner and to get ahead of this consumption curve and this power availability curve we talked about? It ebbs and flows over time based on what gets added and what gets subtracted. By and large, but unless it's colo, which you're not doing a lot of pre-leasing into, and by and large, anything that's not colo is pretty highly pre-leased by the time the doors open- Mm-hmm ... and exits that pool. I would say the trend is more and more towards pre-leasing. Mm-hmm. And some of that is just the broader supply constraints, more urgency in the storm of, "Hey, I'm going to grab this before somebody else does." And then some of this is obviously new use cases, and AI applications come on the scene, and they have fewer options where they can put those workloads, needing large, contiguous capacity blocks. And we're not going spec on a 200 MW shell, right, or multiple suites in that size. So that means the leasing is usually ahead of the actual, call it, entering that part of a development cycle. I would say I've heard that from other data center providers as well, that the upfront need or is getting more expansive. But one of the pushbacks would be that, you know, in industries, not to pick on Google industry, for instance, where in 2020, 2021, they saw 30% higher than normal broadband net adds, but then there was a hangover effect in 2022, and the broadband just didn't grow at all. So we've seen, you know, maybe some of that in wireless, where, you know, we had 9 million postpaid phone net adds a year for a couple of years, and then that's kind of gonna fall to 7 million this year, maybe 5 million or 6 million next year. Are we at all concerned that what we're watching happen is something that's transitory in the sense that IT managers are getting more anxious about availability, they're buying more than they need, and so there's going to be an air pocket where they fill up what they've already kind of pre-leased and growth slows down and pricing power slows down. Does that happen, or does it not happen? I don't think on the enterprise front, if you parse it between the two major buckets, I don't think on the enterprise front, we're seeing that much of any pull forward or type of degree. And then enterprise colocation customer, they're feeling less of a pinch point on these inventory constraints because their needs are so much smaller. Mm-hmm. Right? And on the larger footprint side, I can tell you, we've not been signing. Our signings, we've reported, and not even likely in the third quarter, do not have a massive bubble pull forward related to the new use cases of AI, quite honestly. Mm-hmm. And I think which speaks to we've now gotten numerous quarters with consistent call it growth in new signings. We're gonna enter a phase here where this new use case is a buyer of big, contiguous capacity blocks, and that transition to AI as an incremental demand source will put positive volatility. But we'll do our best to frame the different pieces when we communicate. And listen, there's a long tail after the beginning of AI. We're just at the first innings on the training piece, let alone to go on to inference. Mm-hmm. I don't think this volatility or this concern of pull forward is necessarily a tremendously bad thing for the industry. Okay. Just a couple more questions on the power. Just the 400 MW in development, do you have the power to actually deliver 400 MW if you had that demand? Yes. Okay. By and large, all but call it 800, 900 of the 3+ gigs that I mentioned in land capacity- That was where I was going next, yeah. We have the power in a realistic, normal time, and that last caveated piece is really anchored to this Northern Virginia pinch point- Right ... on power. You think that's about 700-800 of the 3 gigs in the... Is that the Digital? 800+ of Digital Dulles, at least. Digital Dulles, yeah. Is a 2026 arrival or later, potentially? Demand will grow between now and then, so We think so. Okay, so that's super helpful. So then maybe two more questions on this. One is, as we think about the AI of evolution, right? There's a sense that AI is only gonna be 40 kW cabinets driven by GPU chipsets and liquid-cooled, and if you haven't engineered your data center to 1,000 watts per foot, you're just not a part of the game. And so there are companies that are coming to market, private companies primarily, that are trying to create these bespoke types of data centers. So could you talk a little bit about, you know, is Digital Realty building in that 400 megawatts of development? You know, does it-- has it contemplated or can it contemplate or can it accommodate, you know, these heavy workloads? And is the rise of these private companies, these bespoke builds, is that a threat to your position or a recognition that the opportunity is just simply growing at a rate you can't keep up with? ... The, I would challenge a private company telling you that they're leasing entire buildings and been designing historically to that highest power density and actually doing any business. Because I can tell you, we do a lot of business with the biggest cloud service providers, and they have not pushed to those type of infrastructure limits yet for all the compute they're doing around the world. Me, I've been doing a few things. Obviously, modifying our design for our new builds, which I believe is gonna be the lion's share where this goes, because the use cases and the applications and the CPUs in our existing builds are we're not emptying out our 300 data centers to refill them with GPUs. And the innovation is really making it water-ready. So, we're not talking liquid immersion, we're talking about replacing blowing air with water closer to the rack and chip. We've done this before with customers. We've done this in buildings that were not engineered at the outset to do that, so retrofit. I think this is gonna be an evolution. We're in conversations with AI customers that are able to take the infrastructure we have today, be it a building that we built years and years ago, and deploy AI with call it, bringing, call it, that liquid closer to the, to the, to the servers. We're in conversations with customers that want to do this in our new builds, where the ship—like, the ship has sailed on capacity that's gonna be delivered anytime soon. You can't, like, go back to... If you want the capacity near term and say, "Ah, you know what? We're gonna totally redesign this building,"- Mm-hmm like, on the fly. But they're able to call it, make the modifications to meet their design objectives. So I don't think that this is. I think the infrastructure, the physical piece of this is being a bit overstated in terms of what could be potentially proprietary, what could be utilization for different types of data centers. And I think there's you're gonna see just like you tour our data centers that have been supporting hyperscalers, enterprise colo, network dense, more compute-oriented. There's numerous different iterations and flavors of these that call it, look about heat containment, cold aisle, hot aisle, cooling features, infrastructure redundancy. So I think the where this is going is there will be mashups of GPUs and CPUs used by enterprises. There'll be more GPU-oriented firms for some of our large hyperscalers, and there'll be everything in between. And I think we have a demonstrative hand to play in supporting all shapes and forms of that. Just by line, the last question. Have you had or been approached by, say, the hyperscaler guys for, for these kinds of builds? Like, in, in incremental to the 400 MW, which are probably some cloud-centric builds already. Has there been conversations about that other 2 GW of power that you have out there, how to use it and, and exploit it in new ways? What we know about these new use cases, be it AI and training, is the one major factor is the seeking of large, contiguous capacity blocks. You cannot break these up over multiple buildings, campuses, metros, or parts of the country. That is a key attribute that they're really seeking, and you're only gonna find that in large parts of this whole hand, where we haven't already started leasing and selling to other customers. Mm-hmm. Right? So, this dialogue with these customers, it goes back to, call it, very beginning of this year. I think it ratcheted up in its importance to these customers, call it, late spring, early summertime. And we see numerous customers saying, "Maybe I really need this. It's important to my strategy, and that big contiguous capacity is crucial." And lastly, what I would say, we are not following these customers out to unproven markets just for big deals without a competitive advantage to our model and our pricing and our returns. We're fortunate that the big customers have shown preference to do this in their core markets, where they have Availability Zones, where they have infrastructure, where they have teams, and where the future they don't know will unfold. They have other use cases that could fill the capacity they're seeking, be it their traditional Availability Zone Compute. Mm-hmm. Right. It helps to be the incumbent in all the places where computers exist. All right, so kind of over the course of the last, you know, roughly nine months, you know, we've seen positive releasing spreads and the sub- megawatt and the better than one megawatt. In the absence of a recession, the conversation about, you know, the pace of digital transformation hasn't really changed. You know, that seems to be a corporate priority for everybody, including Bank of America. Then we have, you know, this kind of, you know, aha moment regarding, you know, the kind of opportunity seems to have really crystallized over the course of the last nine months. The thing that Digital Realty faced at the beginning of the year was how to harness that and fund it. And how do I take, you know, that top-line growth and turn it into profit? And so, you know, again, round numbers, we've got about, what? $1.5 billion in operating cash flow. We've got a $2 billion CapEx budget. We've got a $1.5 billion dividend, which leaves us with kind of a $2 billion hole per year. It's a little bit bigger this year for some reasons, development reasons. So in order to kind of fill that hole, you know, as rates have kind of moved up, we've had to get more creative. So if we kind of take incremental debt off the table, because your intention is to go from the current post-divestiture 6.3 leverage, you kind of did a few things to get there, right? You raised $1 billion in equity, $2 billion in stabilized JVs, about $300 million in dispositions. You've talked about a desire to do another couple hundred million in dispositions, maybe $750 million in development JVs at the margin. So can you tell us, you know, our progress on those two remaining kind of items? And when we get it done, have we only just plugged the 2023 hole, or we made a dent in 2024? Well, there's a lot there, sorry. The—I think two things come to mind, and I'll get to the funding and the balancing piece of this in a second. But the, what recent experience, I think, is further and further validating are two things about our strategy, one of which is m- newer. But one, we're out to capitalize on this AI opportunity, not as a trade, but as a sustainable business. So like I said, you heard me picking out spots- Mm-hmm. -where we go and where we don't go, but also benefit from the next legs of this, when we're past the training models, when we're on to inference, when we're on to private data sets, when we're on to enterprise being the major consumers, not the consumer. And I think that speaks to the merits of having this full product spectrum, because if this moves in that direction in location or latency sensitivity of any shape or form, the cloud and its availability zones lives within Digital Realty. The enterprise's hybrid IT lives within Digital Realty. If it gets as far as needing to be proximate to the networks, that is also within Digital Realty. Two, something that wasn't. It's been in our DNA, that I think really crystallized for me more than anything, is this opportunity, be it cloud and AI and everything around it, is so big that the best way to maximize value is both a combination of public and private capital working in tandem. And we, as a company, who've been a long-term public company, have been wading into the private capital world for many, many years, be it strategic joint ventures with financial joint ventures on stabilized assets, another separate publicly listed vehicle for stabilized hyperscale assets. And this year, we obviously needed to delever the balance sheet, and we did that by activating the private parts of our capital sources when times were not great for our public source of equity. We were very fortunate through the month of July, to call it, be able to transact on largely 2 massive hyperscale JVs with financial partners for bringing us to, of course, nearly $2.2 billion of capital and getting led to leverage. The next leg of that, not the last, but the next leg of that, which we've also been talking about for some time, has been the development joint ventures. Yes, with that, you're sharing attractive returns or upside in projects with private capital partners, but we view this opportunity so large and expansive, so long run, and the merits of doing that to lighten our balance sheet, to drive greater efficiency. So our organic fundamentals, we've finally arrived, the pricing power, the at, and the mark-to-markets, the same-store growth, getting that to flow to the bottom line, to call it grow our core FFO per share, which obviously, hopefully accelerates and improves our cost of capital on the equity on the public side. So I can tell you, I had no pleasure in issuing stock back at the end of May, beginning of June, at the prices we did. I still think it was the right risk mitigation move to keep us in the realm of opportunities we saw ahead of us. Since then, we've not issued any other equity on our ATM. That's not saying we're taking that off the table, but I think it speaks to the success we've had on the private capital, on the JV side, and the non-core dispo side, and my commitment and conviction on the development side. I can't promise you that we're going to announce one or two of these by the next earnings call, like last earnings call. I thought it was going to be today. Apologies for that. But quite honestly, this would be the easiest form of capital that we raise. Mm-hmm. Just the return opportunities, the opportunity, I'm proud of our company, and this obviously, even put bias aside, this is a tremendous opportunity for these private capital partners to invest inside us. Right? This is a privilege, and we are being thoughtful, finding the right capital partner that brings more to the table than just dollars and cents to be our partner, and it's a great opportunity. And the last thing I'll say, the activity we're doing today, be it the hyperscale stabilized JVs, the development JVs, those were necessities of urgency we had to do. Pleased with the valuations, the structures, the partners, the ones we've announced, the ones we will announce. But the end game here is, whether it relates to our hyperscale, we still believe we can monetize 3+ gigawatts, what many of that will be hyperscale, and attractive ROIs that are profitable, and stay away from chasing business that's not profitable, and then move to a model that benefits from the access to the private capital in almost a fund-like structure, to essentially bring other folks into funding our business and allowing to preserve more of the organic growth and pricing power in our, in our public company. So I want to talk a little bit about that. So the strong demand drivers and what we talked about for the first half of our conversation have been driving kind of a high single-digit top-line growth. The funding vehicles that you've deployed, the equity, the stabilized JVs, the dispositions, have a dilutive effect on those strong top-line fundamentals and kind of drive low single-digit AFFO per share growth. On a go-forward basis, either that could continue to address these $2 billion a year deficits, or the tool that you haven't used yet, this development JV, could be accretive. And it, you know, it depends on how you decide to use it. If you use the JV to attack the same volume opportunity, but at a lower CapEx rate, you'll generate more cash flow, but it'll kind of have a, you know, slowing effect on the top line down the road. In the alternative, you could try to mine that 2 GW or 3 GW of opportunity faster, and use the development JV as what I would call force- multiplier, where you spend the same amount of CapEx, but now you're attacking the opportunity faster. You could drive your top line faster, and that will fall to the bottom line faster, and that'll drive more incremental AFFO per share growth. I don't know which one's more important, is to kind of attack maybe the deleveraging, limit the CapEx, or maximize the growth, or both. I would say it's both, and this is not going to hit your questions straight on. I'm not trying to avoid it. I just think about it a little bit differently. One, we, with urgency, delevered this balance sheet this year from close to 7x to closer to 6x, and we are still not done yet, and the destination is closer to 5.5x, but we're not going to do it overnight, and we believe we have the grace to do it through calendar year 2024 in a prudent fashion. So that's a priority, right? That's where the capital that comes into the company is first going, right? At the same time, we look at this development JV as not only contributing to the delevering, but in general, something that is gonna remove a demonstrable headwind to our bottom line growth by airing, shouldering the weight of non-income-- of hundreds and hundreds of millions of dollars of non-income producing value, that we've been having as a yoke around our, our value creation, path. Mm-hmm. Right? And we're doing that with the view that we've got to get the organic fundamentals that we're seeing in the business to flow to the bottom line, and get to an earnings growth that's, call it, mid to high single digits. And I'm of the view that when we get to that arrival and feel confident that arrival stays in that earnings growth profile, that gives us the flexibility as to the quantity we share with partners and don't share with partners, right? Because for every dollar we joint venture, we're giving them a profit of an ROI project that's in double digits unlevered, right? So that's how I think about the algorithm that we're going through, which I know that we're complicated, and we're trying to simplify this, but that's what we're driving towards as we exit 2023 and go into 2024 and on to 2025. And the quantity of these things we're doing are—like, we're doing a massive amounts of JVs this year because we had to delever a whole turn this year, right? Mm-hmm. They're gonna be a smaller and smaller piece of our, of our pie and less of the headwinds to our bottom line growth. And again, we're trying to do this with a framework where the decisions we make about our partners today don't close doors as to what our partners could be in the future at the same time. The one of the vehicles that was really successful for you was Digital Core REIT. Obviously, they've been going through some challenges with the Cyxtera bankruptcy. Do you foresee that when that's over and done with, that that comes back on the table as a funding vehicle for you, or is it kind of matured? I've been pretty consistent on this point, go back numerous months, and I'm sure I faced a lot of chuckles as people heard my answer to that question and left the room. Absolutely. I described that... I've said that vehicle is down, but it's not out. Yeah. You've seen recovery of the stock price, not all the way to where it should be, but recovery's happened. We're seeing a scenario where it's a large customer, as I'll call it, appears to be on the final stretches of its bankruptcy proceedings. The facts and circumstances you can read from the public filings seems to be a much rosier outcome than the draconian states people had assumed. I believe the finality of that solution, which is we're talking within months, not years, it will clear the runway for that vehicle to have its cost of capital reset. And when that happens, I believe it has the path to actively grow its business, and we have a tremendous amount of pipeline that it could participate in, and it'll be an incredible partner. Very similar to the deals we've done, announced in July, but obviously a vehicle that we're even closer with than these new partners we've kind of onboarded. I think at the Q2, you adjusted guidance about $0.05-$0.07 AFFO per share. Were you draconian in that sense of setting that bar? Well, at least two of it is not likely coming back. That's-- that was called pre-petition claims. So we had to handicap when this thing comes out of bankruptcy, which leases of Digital Core REIT get rejected, and that impact us? What leases at Digital Realty that are rejected? And then if a lease is rejected, what do we-- well, how do we make out on the end customers in this capacity? I mean, one of the-- we got into the colocation interconnection business with a view that servicing the whole-- two views. One, servicing the whole customer spectrum was the right way to maximize value in this space and really tap into a market that really was looking for an incremental global competitor and giving the customers of this world another choice. And the ancillary benefit we found is we already were dealt with a hand of these reseller rework, like, sandwich lease situations, that by being in their business, we can mitigate when things go wrong, right? And we've seen this movie before, and we stepped in to support the end customers in the event a lease was rejected. This scenario, I'm not sure. Well, I'm not sure we were draconian in that. I think we did our best to handicap. I think we're very well prepared to face whatever outcome happens from this bankruptcy. So if we kinda start wrapping up, what I hear is Q3 leasing, very happy with it. Year to date, re-leasing, running well ahead of expectations. Cyxtera outcome could be better than expected. Non-dilutive new funding, joint ventures with partners coming up in the development side. Possibility that Digital Realty becomes the funding vehicle again in 2024. Kinda sounds like things are going pretty well. We're blessed with an industry that has got incredible demand drivers with new ones popping on the scene. The value prop, supply is constraining and driving the pendulum towards us in terms of pricing, rate, and returns, as is our value proposition strengthening. And it's nice when the fundamentals are great or improving, but it's also happening at the same time when the capital is exiting legacy asset classes and wants to invest alongside of us. So I think I don't think we ever warranted the negative lo- the negative perspective we wore from the beginning of this year. But I think we're seeing some recovery, and we've been a show me story, and now we've been showing you, and I can promise you we're not done yet in this calendar year, for sure. That sounds like a great place to leave it. Thank you, Andy, and thank you, Jordan, for coming. I appreciate everybody for being here, and thank you for listening. Thank you. Thank you.