20:50
Welcome back to Equity, Tech Crunch's premier podcast about the business of startups. I'm Rebecca Bollan, and this is an episode where we invite industry experts to help us explore a trend in the world of technology and dive deeper. And with me today is senior climate reporter Tim Dechant. Hello, Tim. Hello, Rebecca. How are you doing? Yes, great. What are you focused on lately? Well, Anthropic is in the news again, and this time it's not just about their IPO. They have become another AI lab purchasing gigawatts of natural gas capacity for generators. So we should probably talk about how we power data centers.
You are right. You know, it seems like every week someone announces a new data center power plant. But there is a gap between what is announced and what is actually built. And huge capital depends on which side of this abyss you find yourself on. Last week's midterm elections made data centers the center of political debate, and the question is whether this gap will widen. Well, fortunately, this week we have someone with us who knows a lot about data centers and energy. So to help us separate the real from the hypothetical, we're joined by Brent Nelson, Senior Managing Director of Markets and Strategy
at Ascend Analytics, where he forecasts long-term energy markets for clients who are deciding which energy investments to invest in Brent, welcome to the show. I'm really glad you're here because I have a lot of energy questions, given all the drama surrounding data centers. So congratulations and thank you for joining us. Thank you for that. Thank you for inviting me. So, for our listeners who don't know you: you were previously a professor for 11 years. You spent two of them at the Ministry of Energy, working on R&D and strategy for commercializing energy efficiency and
load flexibility. As a professor, you taught at Northern Arizona University, teaching mechanical engineering with an emphasis on energy systems, correct? So you've been working at Ascend Analytics, an energy and analytics consulting firm, since 2019. And the focus there, as your entire career shows, was on forecasting and modeling energy markets. Now is the perfect time to have this conversation because last week on the show we talked a lot about the midterm election results and how everyone is a little panicked about how the blue wave will affect data center development. And I'm curious what you're hearing from communities and politicians about this backlash.
Yes, I think these are more of minor changes than anything significant. You know, one of the few bipartisan issues in the country today is the backlash against data centers and energy affordability issues So I'm not sure it was purely a partisan issue-discussing moratoriums on data centers and trying to emphasize affordability, impact on water resources, etc. We have indeed seen Democratic gubernatorial candidates campaigning on the basis of moratoriums, construction halts, delays, and so on. But I wouldn't say that it fundamentally changes what would have happened under other circumstances.
So, when we think about the end result, we think these are only minor changes: probably slightly stricter or longer moratoriums than they might have been under Republican state leadership. But still, we observed a slowdown. We have in any case considered the processes for determining energy availability. The only question is how severe and long these delays will be to get everything in order before further development is allowed. So, what do you think the consequences of this will be? After all, a few months ago, the Wall Street Journal reported that total investments in data centers and related AI infrastructure should reach $10 trillion between 2025 and 2032. This is a huge amount, but we
know that a significant part of it is based on debt obligations. What could be the further consequences for the economy if these data centers are not built on schedule? I think we'll see anyway.When I say "slowdown," I mean it's going to last one to maybe two years. So, looking at 2032, I don't think things will change much by then. 2027 is a completely different matter. Therefore. This only applies to moratoriums. The problem remains: will they be able to get energy at all? People I've talked to in the data center industry say that the biggest problem is actually that there are sites, but they can't connect electricity.
Yes, that's why there are many processes happening now aimed at obtaining this energy. So, there are people who view the metered solution as a temporary option until they can connect to the grid. There are programs built around consumption limits or conditional grid connection. Some talk about completely autonomous energy systems. I think this is talked about less often now, because those who are trying to implement it are starting to understand how difficult it is. But the main topic of the last year was the question: where will we get energy from? Will we be able to get it? But many processes are still ongoing. Many projects are still in the development stage. It seems to me that we are
underestimating the political risks for some projects under development. If your project is launching in 2030 and depends on a large-scale modernization of gas pipelines to service a new power plant, what are the chances that a Democratic president in 2029 will revoke this permit? I think they are not zero today. This does not mean that the president in 2029 will necessarily be a Democrat, but if that happens, your project must be sustainable enough to survive in such conditions. I believe that many things are still in motion, many things are in process, many things are still happening. In my opinion, these are only minor changes in the short term.
Brent, one more question: how many of these projects are actually being built? If you are familiar with the energy industry, you will see that projects are often announced with the hope of getting in line for a connection or finding land, and as a result, a significant percentage is simply canceled. Is the same happening in the data center space Yes, we saw the same thing under the Biden administration in the area of clean energy development. Ignorant people often say: look at all these projects lined up. They were never ready for implementation. As a developer, you know: there are a million reasons why your project could fail. There could be a bad land deal, drainage issues, or a community that for some
reason has decided they don't like you. You may face problems with financing the project or with the costs of connecting to the network. I mean, there are a lot of different reasons why a project doesn't move forward. So, as a result, you are doing several things at once. In addition, there was a kind of speculative world where people engaged in development and resale, hoping simply to constantly reinvest capital. And when we looked at the clean energy project pipelines under the Biden administration, these were always numbers that were many times greater than the actual plans for what could be built. We're seeing something similar now with the development of data centers. There are some overlapping activities
because you need a certain margin of safety because you never know what will happen to the project and why it may or may not move forward. But on top of that, there is also a speculative mania where everyone is trying to get into this niche and everyone is trying to do it. And that led to bloated lines, and some of those numbers are just ridiculous, right? If you look at Texas, peak consumption in Texas is now about 90 gigawatts. And there's something like This is the whole state, right? State, more or less. This is ERCOT, which covers the vast majority of the state, not quite the entire state, but close to it. The large consumer queue for ERCOT is about 400 gigawatts, correct? And you're not going to increase the system fivefold
in 5 years, are you? This is simply an unrealistic number. You can't build 100 gigawatts of capacity per year. There are no 100 gigawatts of high-voltage equipment, there are no 100 gigawatts of turbines. It's just, you know, these numbers aren't serious. So when we look at any large consumer queue or load forecast, you have to be able to adjust them based on what you think is plausible. It is difficult to look at a specific project and say: this one will be implemented, and this one won't. But when you look at the big picture, there is both duplication and speculation. And I think it's important to note that both of these factors exist. Both duplication and speculation. In any case, what you
can actually build is much less than what we see in the queues of large consumers. How much less, what are we even talking about here? We see really big numbers here. So you're saying it's about a third, you know, or a half? Probably from a tenth to a fifth. Um, you know, and this but it depends, right? It depends on what number you're looking at. So some say: let's look at the entire queue of heavy workloads. Well, it's about an order of magnitude less. Some say: let's take this heavy workload queue and reduce it to realistic figures. Well, then it's perhaps not such a big reduction from that already reduced number. And so we have to be careful what we say when we look at high-load forecasts. You know,
if you have someone who has already done the due diligence and reduced everything to real numbers, then your attitude towards that number will be completely different than when you look at the "raw" data. Mm. So do you think developers intentionally apply for multiple projects to have a choice, or is it just the norm? Is this always the case? Well, a normal aspect of project development is working on multiple options at the same time, right? If you ask any developer, they will all tell you that every one of their projects is great and they will all be implemented. But that's what they
say. The realistic, real truth here is a little different. They know they have to have backup options. They know that you need a safety net because, as I said, you never know what could lead to the failure of a project. So there is always some duplication. But, as I said, in addition to this duplication, there is now a speculative mania where everyone is trying to get in. Everyone is trying to grab money from data centers. And so, you know, there are people who used to be in clean energy development and now they've become energy site developers, right? There are people who have never been in this field, and now they call themselves data center developers just because they happened to have land next to a gas pipeline, and
they say, "Oh, well, I'm a data center developer now." "I have a site for a power plant." So there is this speculation, a speculative mania of people trying to enter this niche. But this is in addition to the normal reality of duplication when developing projects in a high-risk world. This seems crazy because it looks like they are taking on debt for this very large bet. It's like taking out a loan from a bank and then going to a casino. Well, when you take out a loan, that's a different stage of development, right? So when you look at the big queue for connections, there are a lot of projects out there that don't have any debt yet, right? These are just, you know, early stage projects. When
you start taking out a loan, it means you start building. Um, and the world of debt obligations is a little more complicated. Um, generally speaking, if I want to get a loan, no one will give it to me until I can prove that I have the income to repay it. Um, and how do I verify this income? Well, if I'm a data center developer and I'm creating, for example, a cloud center that can have any number of tenants using my facilities, the bank won't give me a large loan until I have a signed
agreement with a reliable tenant, right? This is how I show that I intend to repay the loan, and this is what is necessary to obtain financing. Now, are these lenders doing their due diligence thoroughly enough? This is an open question . Um, but you can't just take out a loan without having some securitization on the other side. To what extent is all this, I mean, posturing or perhaps executive orders from governors? For example, in Texas, Governor Abbott called for an audit of large data center developers, which effectively became a kind of moratorium, postponing the implementation of some of these projects. To what extent will such policies or political actions help weed out those developers who are not serious?
I mean, they hope that there are many ways to filter and separate the wheat from the chaff, right? And anything that increases costs will help you separate one from the other. Um, so one of the things that they're doing in a lot of places is increasing the security deposits so that you have to put up more financial capital up front. Um, this makes it harder for one-day firms, but at the same time creates an uneven playing field for players with different capabilities . For example, larger operators have an inherent advantage in such a structure where capital requirements arise. But slowing down processes also separates the wheat from the chaff, doesn't it? Time is precious, so the people who are better able to
survive a recession are usually those with more capital. So, both of these factors together will help put everything in its place, but in the end, the most important thing is your ability to find a tenant, right? So the tenant-it could be a partnership with a hyperscaler, like a large tech company, or it could be multiple leases with different potential startups or something like that. You need a tenant in some form or another to get the money going. Is it difficult to find tenants? Judging by the way things are being talked about today, AI companies and hyperscalers can't find enough capacity. It all depends on who you're
talking to and what time period you're talking about, right? If you look at a hyperscaler trying to get compute resources for 2027, then yes, they can't find enough capacity. And when you look at 2030, where there are a hundred different data centers competing with each other, it's a slightly different scenario. Another thing I would add is that there are a lot of companies doing a lot of things in the AI space right now. This is somewhat similar to the dot-com bubble in 2000-2001, isn't it? The internet has obviously become something very big and important, but that doesn't mean there weren't a lot of ideas that didn't come to fruition, right? Therefore, many companies
went bankrupt. A lot of people lost a lot of money on those companies that went bankrupt, and there's no reason to think we won't see something similar now, right? This doesn't mean that AI won't become something significant. This doesn't mean that computing power won't be needed, but those who will need it in five years may be significantly different from those who need it today, right? There are many venture- backed startups doing a lot of different things in the AI space. How many of them will remain on the market in five years? Probably much less. And many of those who do good work will likely be bought out by one of the big tech companies or hyperscalers, and then you will have a reliable and solvent tenant. The deals with the best chance
of success, and the most reliable debt, are those backed by leases with large, well-capitalized technology companies. But even in such cases, one has to make many concessions. For example, I believe there was a deal with Clean Spark, which is building a data center for Meta, and Meta is leasing that center, but even they had a hard time closing the deal without significant concessions, right? I believe that there is a serious tightening in the debt market Yes. Although, when I talk to people from the debt market, they say that they are usually willing to finance data centers because it looks clean and simple. Because
they say: well, you know, I have a nice and simple one? Clean? What do we mean by the word "clean"? Because if we're talking about energy. clean in the sense of the contract, not in the energy sense, you know? So you have a large hyperscale tenant . It is well secured. This has a good collateral. They would be better off entering into such a deal than, say, a solar energy project with PPA risks, where you need to attract investors in tax equity, and it is a much more complex financial structure. So bankers tell us that it's better to work with data centers because it's a clean deal that's easy to document. Of course. Well, speaking of "cleanliness", I want to hear how exactly these data centers are powered?
Because now we hear about a real race for different energy sources. From what I've heard, perhaps you can give me some insight into this matter. It's not that easy to power data centers with renewable energy, unless nuclear power somehow works. In fact, it is simply not enough. That's why everyone is returning to natural gas again. Many companies and hyperscalers are trying to secure their own natural gas, just to have it in case communities say, "No, you can't use our pipelines," etc. So, can you describe the general situation? Yes. You know, if you go back a few years, all the tech companies were talking about was clean energy, clean energy. And then the political winds changed direction. And they seem to have stopped talking about clean
energy. But they still made clean energy deals, right? If you follow the news, you constantly see reports of deals with those tech companies. It's just, you know, they don't necessarily shout about it on every corner like they did three years ago. So, clean energy is still evolving. But the problem that I think many people overlook is that solar energy cannot simply be replaced with gas. These are not interchangeable energy sources. One is not equivalent to the other. And as we move into a world where, outside of California and the desert southwest, most regions are preparing for winter, reliability risks become the biggest concern. And the main challenge with winter reliability is that it usually happens at night, and the load remains consistently
high. So think about long, steady, high load. If this happens at night, solar energy will not help you in any way. This is usually accompanied by light winds-not only because the cold often comes in calm weather, but also because when there is wind, there are usually no problems. Your problem, by definition, occurs when there is no wind. So, there are resources left that are available at night during long peak loads. This is also a problem for storage devices, as such a long peak does not suit them very well, as you pay more and more for longer durations.
Storage tanks and solar energy complement each other perfectly for summer peaks. Summer peaks occur on a day when the sun is shining and a little after sunset before bedtime, which is a few hours of load. Storage drives do a great job of this , but it becomes a much bigger challenge at night in the winter. That's why everyone is returning to gas. And if you understand that clean energy sources are cheap and available, but they don't quite meet what is actually needed . That's why everyone chooses gas as their main option. The argument we at Ascend have been trying to make for some time now is that gas is not as good a solution as it seems, for many reasons. First, you need gas generation equipment, which
is currently in short supply. Additionally, you need to supply the fuel itself, which is not an easy task. You need gas pipelines, which is also not easy and carries political risks. Um, you also need to have the fuel delivered during those critical winter storms, as they are the ones that are typically associated with gas outages. So, if there is an outage in the gas supply system and that is what is causing problems in the power grid, what are the chances that you still have gas at your facility? Hmm, maybe not very big. And so, you know, the gas solution is not as good as people think.
Moreover, if reliability risks in winter are the biggest problem or the most stressful period for the network, it actually doesn't happen that often. It's not like the summer heat that happens every year. This happens about once every three to five years, when there is a winter storm or some kind of major storm that disrupts the gas supply. And if this happens infrequently, then perhaps a connection with the ability to limit is not so bad, especially if you add energy storage. So my argument, which I've been making for a long time to anyone who'll listen, is that plug-in with the ability to limit, combined with batteries, is a pretty good solution, maybe as good as gas, maybe even better. Um, it's
clean, might be cheaper and faster. Um, actually all of the above. This doesn't mean everyone agrees with me, but that's what we stand for. Connection with the possibility of limitation. what is the solution? Connection with the possibility of limiting consumption. Of course. OK. So, if you're on a connection where you're powered by the grid most of the time, and then there's a winter storm that causes a power shortage, you're temporarily out of power. Um, but if you have storage that can cover most of this problem, and the problem itself only occurs once every three, four, or five years, maybe it's not that critical
. Um, that's what we're trying to prove. This doesn't mean that people are listening to us. So people still continue to cling to gas. Um, nuclear power is still a long way off, but gas is the standard that most people are looking at. We just need to change course. You mean nuclear energy, not nuclear weapons. Got it. Sorry. Yes, nuclear energy. Atomic Does this storm story seem to be a regional risk? Sometimes there are very large storms that cover a large part of the United States, but they never cover the entire country. It seems to me that one solution could be better network connections, perhaps this would help redistribute energy better. Perhaps we should briefly consider how the electrical grid works.
I'll probably give the answer first, and if we want to go deeper, we can do that. You know, it's like a magic bullet solution-to build huge power lines across the country to transmit power from California to New York. Hmm, not necessarily realistic, right? Building new transmission lines is very difficult. It is difficult to find a place to place them. It's hard.When you go interstate, getting permits becomes incredibly difficult. So it is practically very difficult to create the level of power transmission across the country that would be needed to really solve the problem, because winter storms usually cover huge areas. And to get energy from somewhere else, you need really long transmission lines. I'll probably
stop there. This is very, very difficult. If you could wave a magic wand, that would be one thing. But these would have to be really long transmission lines. Would a new bill on the permitting system, if it were passed, which is unlikely at this time, solve any of these problems? Wait, Tem, describe this bill on the permitting system for our listeners. Sorry. So, there was a new authorization bill, the "Bipartisan American Affordability and Jobs Act, " which seems to have just passed the Senate. It aims to speed up the issuance of
permits and approvals for facilities such as new power plants and high-voltage transmission lines, and also obliges data centers to cover part of the costs if they benefit from it Ahem. So, it will give FERC the authority to reserve power lines. It would help, but it would still be difficult. The Federal Energy Regulatory Commission is a topic for a separate conversation. But, essentially, FERC has permitting authority over pipelines. They currently have no permitting authority over power lines. One of the things in this permitting reform bill is that they will have more power over energy transmission. This is a bit simplistic, but it will help somewhat, although large interstate transmission lines are still
difficult to build. This also assumes that you can choose the right location, closer to the consumer. You know, there is also the problem of connecting to local networks. So even if you have these big power lines, there are still local problems that need to be solved. And people who have more experience in electricity transmission could probably tell you more about it. But it's just not that simple. So I think it helps with the reserves, but it still remains a big challenge. I know you've talked about this before, but what do you think about the permit system reform? Why does this matter more than people actually realize?
This is important because, you know, we've seen what can happen when a presidential administration decides it doesn't like a certain type of energy. And I don't just mean the Trump administration, right? Many, especially in the clean energy space, don't really want to talk about what the Biden administration has done in energy, including "undermining" Keystone-I say " undermining," sorry, I don't mean to imply that they physically blew it up, but you know, they revoked the permit for the Keystone pipeline when significant funds had already been spent. For years, they maintained a "pause" in LNG exports, so to speak, which everyone in the industry knew was a moratorium, not a pause. And it was kind of the same strategy: I don't
like this kind of energy, so let me just stop it at the federal level using executive authority. And then Trump seemed to say, "Well, anything you do, I can do better." And now, you know, wind power was completely stopped by the Trump administration. Solar energy has slowed down, although it has been moving more actively recently. So, you know, the executive branch's ability to significantly influence infrastructure development is quite significant, and we've seen that, especially during the Trump administration, but not just during the Trump administration. We have seen what can happen when a president decides he wants to support or discredit a particular energy resource. On top of that, you know, most environmental
groups have spent the last 50 years trying to stop various processes. And now there is a very well- developed strategy for preventing something from happening. And so, you know, one of the main things that people were talking about during the Biden administration was, " We have all these clean energy projects, we want to build all these clean energy projects." But whether it's a conservative group or something else, you know, there are many ways to stop projects, whether it's a power line or a solar farm or anything else. So, NEPA became problematic. NEPA-that's.I'm getting confused now. It seems to be the National Environmental Protection Act.
It seems like it. But this has become a way to get stuck in " permit hell," where projects are reviewed over and over again and ca n't seem to get off the ground. We have problems at the local level and at the federal level. So the ability to streamline permit review, speed up processes, and provide greater certainty-this is really important for project implementation. And certainty is better than uncertainty, right? Even if you get rejected, it's better to hear "no" nine months later than four years later, right? How much do you think resistance would decrease if data centers were able or willing to pay for more of the necessary upgrades?
I think this is a necessary but not sufficient condition. When we talk about building data centers, I believe that cost isolation is a minimum requirement. If you don't have a plan to contain costs and protect taxpayers, you won't achieve anything. But it's not just about that, right? In agricultural areas, water use becomes a big problem, so a water consumption plan is also needed. Noise is a problem for those who use gas. There are concerns about air emissions. Some local communities are strongly opposed to this. Others don't, because they see the jobs that a gas-fired power plant will bring. So it's not necessarily a simple story. But in addition to all these very real and understandable logistical issues, there is
also a general negative reaction to AI. For example: AI will come and take our jobs. AI will take over the world. AI will destroy humanity. Or simply: we don't like data centers. This is a certain undercurrent directed against all this. So, even if you solve the problem with costs, water, noise, and air, it doesn't necessarily mean you'll get community support. Yes. Well, examples are starting to appear -I forget where exactly, but in some communities they're saying, "Okay, we'll pay bonuses to teachers." And they just start bribing people and, you know, paying communities for things that they could, well, pay big taxes for. Um, and essentially, you know, give a financial incentive to communities; Because, you know, you
can't really fool anyone with the phrase "it will bring jobs." Well, yes, of course, maybe a few jobs in the short term, but it's not a long-term employment opportunity. But as a way to generate tax revenue, it could be possible. Yes, and this is not news. I mean, it's the same scenario as in solar, wind, and energy storage manufacturing, right? You know, they say, "Okay, we'll come to your area." Yes, there are a few construction vacancies during construction, but not much afterwards . Um, but we will become part of the tax base. We will pay for new schools, for a new fire station, for new parks, and so on. And this, in fact, is the entire scenario for the development of such projects.
No matter what you're developing-a data center, a solar or wind project, or a gas pipeline-when you come into a community and say, "We want to build something here, " you take steps to gain support by explaining, "Here's why you should let us do this in your neighborhood." And usually, you know, providing a tax base and funding some public infrastructure becomes part of that. Yes, by the way, I just remembered that it was Louisiana. Um, teachers were getting $50,000 bonuses for the Meta data center. Meta is just throwing away its own money. So, okay, if you're a founder or investor and you're looking at
AI infrastructure right now, what would you be wary of assuming before you embark on your next venture? I would advise caution with the assumption that a gas data center with a long implementation period will receive all the necessary approvals if the political winds change again in a couple of years. Hmm, this must be a tough time to build a data center, right? Or for trying to do it? That's what worries me the most, right? That is, will there be support for new gas infrastructure in my jurisdiction? And if the answer is yes, then you're probably safe. And if the answer is "probably not, " then you should seriously consider the political risks of this project.
Another point that concerns me: who are the tenants of this data center? Because if the data center has no tenants or it depends on a bunch of startups that might go bankrupt, I would be very worried about the ability. You mean Neo Clouds? Not all Neo Clouds are the same. Some of them have an anchor tenant, well, like an anchor tenant in a shopping mall, you know? A grocery store or something, right? And around it is everything else, where tenants can change. But if there is a very reliable anchor tenant in Neo Cloud, the situation is completely different than when there is none. So, you know, how.
This is a double-edged sword for Neo Clouds, because, on the one hand, having an anchor tenant makes their development more stable. On the other hand, it's a risk for them -keeping, so to speak, all their eggs in one basket. Yes, you have to look very carefully to see what kind of basket it is, right? Yes. So, you know, diversification is always good, but stability is also good. If you look at large hyperscalers, the probability of their bankruptcy is almost zero, right? Even if someone else develops a better AI than theirs, they'll just buy that technology, you know? And then they will swallow her. So I'm not too concerned about whether hyperscaler will exist in 5 years. Perhaps
mergers or acquisitions will occur in the meantime, but it is highly unlikely that it will disappear. I think you need to be very careful about whether you're investing in infrastructure or being a lender-you need to carefully assess who will pay the rent. As you said, these are very interesting times. I think there is a lot of uncertainty ahead for the next 6 months about how companies on both sides will operate. Um, do you see a way to resolve this before the change in Congress? Well, you know, I think the best opportunity for reform of the permitting system is right now, during the " lame duck." Um, I know it's not ideal, and I know a lot of
environmental groups don't like it, but, you know, if you're an idealistic group, you can't like anything because nothing fits your idealism, right? Hmm, I think now is the best opportunity to reform the permitting system. I am very worried that Democrats are overestimating their strength and thinking they will have more leverage in negotiations in the next Congress. Um, it does n't really matter if you don't get the president's signature, and you know, if you can do it now, I think you should do it now, because there's no guarantee that you'll have the president's support even if you get everything you want. I think, I think, that now is the best and only opportunity to reform the permitting system. I'm not a Washington insider, so, you
know, take my opinion for what it is, but yeah, I think this is the most important thing that can happen before Congress changes. Well, on that note, um, we're running out of time. Once again, thank you very much for joining us. I feel like I could ask you anything and you would give an eloquent answer. I have many thoughts. That doesn't mean they're all good. This is great. We may have to invite you again. So where can our listeners check out your work online if they want to hear more of your thoughts?
Yes, www.ascendanalytics.com is our website. We publish many white papers and analytical materials. We publish regularly. Of course, we also provide paid consulting services. And, you know, I'd be happy to chat on LinkedIn. Good. Tim, what about you? Where can people find you? You can find me almost everywhere: on LinkedIn, X, BlueSky, I'm everywhere. And, of course, on techcrunch.com. Yes, I have the same thing. You can find me anywhere . And Equity can be found as Equity Pod on your favorite platforms. See you next time. Equity hosts-TechCrunch senior reporters, producer-Teresa LoConsolo, editor-Kel. Subscribe on YouTube or wherever you listen to podcasts, and
find out about upcoming events at techcrunch.com/events. Thank you very much for your attention, see you next time.
