Not Your Father’s Data Center - It’s All About the Power

Episode Date: April 10, 2023

The phrase “energy crisis” has been tossed around frequently the past few years, and not without reason: as demand for fossil fuels skyrockets, climate change speeds up too. So while the U.S. puts legislation into place to help mitigate the climate crisis, there’s another problem in the process - our energy grid might not be able to keep up with these environmentally friendly power alternatives. What happens now?  On the latest episode of Not Your Father’s Data Center Podcast, CRO and host Raymond Hawkins chats with the Vice President of Global Data Center Sales with Bloom Energy, Jeff Barber, about the energy crisis, ESG, and Bloom Energy.   The two discuss:  Why reducing your carbon footprint now adds time-value  How Bloom Energy offers an alternative that lasts into the future  What Bloom Energy is, how it generates power, and what their power-purchase agreement looks like  “So, Bloom energy is solid-oxide fuel cells that are on-site, at your facility, in a micro-grid form. So this is power-generation at your facility, without the centralized grid, with the centralized grid…Bloom is intended to create energy for the facility, to be the primary energy to that facility, 24/7 365,” explained Barber, noting that these are not batteries.  A California native, Barber has collected over 25 years of experience in the energy, sales, and data center industries. Though he first started working on network troubleshooting through the G-Tech Corporation as a Network Support and DEC VAX System Administrator, Barber has since held leadership positions at a number of globally recognized brands, including Hewlett-Packard, the EMC Corporation, and Oracle.  Prior to joining Bloom, Barber was a Partner & EVP Sales and Business Development leader for Prime Data Centers. He is a graduate of California State University at Sacramento, with a BS in Business Administration.  

Welcome to another edition of Not Your Father's Data Center. I'm Raymond Hawkins, your host. Today, we are joined by the Vice President of Global Data Centers of Bloom Energy, my friend, Jeff Barber. Jeff, how you doing, bud? Good to see you. I'm good.

Good to see you too. So I hear energy's a problem, right? Energy's a problem, but first I'd like to address the extreme cruelty of a 6.30 start. Yeah, that's fair. I know we need to get that out. We are recording early in the morning. We are recording early. Yes, energy, especially in our world, right? Data centers consume a

massive amount of power, as we all know. ESG, the environment is a critical, absolutely critical point. So what Bloom Energy does fits perfectly. It's a perfect storm for the company. We'll go into that a little bit. Yeah, great timing for you guys, what our industry is facing and what our planet's facing. So a couple of things. Before we get too far down the path on what we do together professionally, I'd love to have folks understand you a little bit. Where'd you grow up? Where's home?

Where do you live? How did you get to Bloom? Give us a little bit of your background first. Sure, sure. So I'm a California boy, born and bred. A little bit of a circuitous route to data centers. I actually started my career in technology working for the California State Lottery on mainframes, believe it or not. So I started my career living in the data center in a very cold room. This is well before hot L containment or anything like that. We actually had to be at the

console doing things. So I spent quite a few years there. And then my formative sales years, I was primarily with Hewlett Packard and a company called EMC, which was acquired by Dell. So EMC VMware. Did you work at IBM? I didn't know that. Did you? No, just on the mainframe side with mainframe. Well, mainframes, actually, they were many computers from a company called DEC. And then prior to that, Concurrent and Tandem, believe or not. So we're going way back. You're going way back. That's Tandem pre the compact acquisition. That's right. Well, the equipment wasn't actually new that I was on, so it had some age on it already. And you got an EMC stint in there, too.

Oh, my goodness. Yeah, great company. That's a professional selling there. 11 years at EMC, the toughest, best sales environment in the world. EMC was acquired by Dell. As you know, they own VMware and many other products. So ran global accounts for them for the the technology sector, and just had an absolute, absolute blast. So I went from working within the data center, like I said, very cold environment

with my contacts popping out, very dry environment, over to filling the data center on the IT equipment side, right? So servers and storage and software and all of that good stuff. After about 15 years in the IT sector, I went to work for a developer. So we were doing build the suits, we were doing powered shells, data centers, fit outs. So really learned the business from the construction side. So when I met you and Compass and all the other folks, and then Bloom Energy kind of popped on my radar with all of the issues right now with power.

Access to power, obviously supply chain issues. Most importantly, carbon footprint of data centers, right? Massive, massive power consumers. Started talking to them and working with them a little bit. Understood the messaging much better. I said, this is beyond a perfect fit for the data center world. So came on over to Bloom. So California the whole time, all your gigs in California?

All my gigs in California, but most of those at the airport. Gotcha, yeah. Right, so I'd call it home base. In and out of SFO a few times? Just a few times, just about once a week, yeah. All right, so technology the whole time, which is, I think, you and I, one of the reasons we connect so well, having spent a career filling the buildings full of data center, filling the building full of computers, understanding what needs to go on there. So as we switch and think about how do we get here, how did we end up where we have power constraints literally in every meaningful market, right?

There's some issue with how do we get power, how do we get it delivered, and what do we do to feed these very power-thirsty facilities and computers? So as you leave the construction business, leave the data center business, and you hear the Bloom story, talk to me there about why that, what about Bloom stood out, genius, said, hey, this makes sense, and i ought to be helping these guys grow this business yeah absolutely so you hit it on the head um any most popular markets have power constraints right now but that like most issues that's a multitude of factors that drove that um first and foremost you know being uh the environment the climate, carbon footprint, the desire to shift towards greener energy. That does put constraints on a centralized grid, which is kind of another contributing factor.

When the grid is centralized, it's difficult to transmit to some of these locations in the quantities that we need. We're talking tens of megawatts or hundreds of megawatts in some cases, as you know. The Bloom message is essentially there is time value to reducing carbon now. We're all marching towards zero carbon footprint. That's a fantastic goal. It's a goal I believe that will be achieved, but today it's extremely difficult. First and foremost, Bloom reduces the carbon

footprint of power at the data center dramatically over centralized utility. That's today. There is already a capability in the product to be carbon free by using hydrogen as a fuel source versus natural gas or biogas. So today, right now, we can cut the carbon footprint 20, 25% over standard utility. I think it's important to talk about what Bloom is though, and what it is not. So Bloom Energy is solid oxide fuel cells that are onsite at your facility in a microgrid form.

So this can be, this is power generation at your facility in a microgrid form. So this can be, this is power generation at your facility without the centralized grid, with the centralized grid. Can I ask some dumb questions? Because I think there's lots of people that think of, hey, this is just a big battery pack or this is something. It's not that, right? These are not batteries. We're not storing energy. We're actually generating energy inside the fuel cell, correct? That's exactly right. So we get that question a lot. We do have cathodes and anodes printed on each side of each fuel cell.

So that looks a little like a battery, but that's not what we're doing. Bloom is intended to create energy for the facility to be the primary energy to that facility 24-7, 365. But I mean, and for data center folks, if we think about, I've got a diesel gen sitting out back, right? It's generating power for me. I'm not storing, it's not a UPS. I'm not storing anything. I'm literally generating power. And that's what my Bloom fuel cell is doing, right? It's a power generation device. It may have some storage capability inside it, but it's generating power. And the same kind of concept, it's power when I need it, right? It's generation and it's flow and it's on, right?

There's no, we're not flipping switches on and off. We're not waiting for this thing to spin up in five minutes. It's either on or it's off. That's exactly right. It is, it does have load following. You can go, you can follow peaks. You can go up and down with the device. But most importantly, a great, great comparison with the generator. There's no combustion in our device. When I say fuel, I'm referring to converting molecules into electrons.

Okay. So I've got an engine. I've got no combustion, fully solid state, no moving parts other than some circulation fans, which are not mission critical to the devices and very redundant. So I've got a solid state device using an electrochemical process. That's the proprietary process for Bloom. Each device is putting out 55 kW per box. Those are aggregated into a common bus, into something we call an energy server, which is six of those devices. And from there, we can string them together

to provide unlimited power to the facility. So a standard setup's got eight, and I'm looking at 300, 400 kilowatts coming out of it. 320. 320. Each box is about 55 kW. It's a great point. Back to what did I see in Bloom? Looking at Bloom when we went public or even before that, when we kind of came out of stealth mode, the devices were putting out, say, 20 kW. So now we're at 55 kW. It's continuously

improving. That's moving up to spectrum roadmapped as well in the not-too-distant future. But we also were able to gain economies of scale as the company grew. So we have much, much, much more competitive pricing in many markets. Certainly in California, we have lower pricing than centralized utility. So everything is falling in line with the company as it relates to data centers. So we'll get back to the price conversation, because I do think that'll be an interesting one. But back up, you were hired to help them grow the data center business. Talk to me about where Bloom cut their teeth, because the

business isn't brand new. The business has been around a decade. How did the business start? What was it doing from the jump? What market was it serving? Because it certainly appears to me to be big enough now that it can handle what we're doing in the data center business, but it didn't start there. So back to the beginning, how do we in the mission critical space go, hey, I can trust this thing? Yeah, that's a very good point. So going way back, the device, the founder of the company and inventor, KR, our CEO, actually worked for NASA. And the device was intended to produce oxygen for Mars missions. Okay, well, those missions were canceled or delayed and KR founded Bloom because we can reverse that process.

You know, creating oxygen is taking energy to molecules. We can reverse that process and take molecules to energy. So the company came out of stealth mode. And then, you know, I would say the first deployments were early adopters in the classic sense. Lots of retail, some hospitals, some data centers, some data halls in data centers. Companies like eBay, for instance, in Utah, were using Bloom very early on with our friend Dean Nelson, who I'm sure most of us know. Yeah, yeah.

Yeah, running that over there. So to the current, just constant improvement, constant monitoring, which is important to the mission-critical point. So every Bloom device, every fuel cell stack is monitored by Bloom. Tremendous amount of predictive analysis going on. Your folks have insight to it as well through the BMS or however they want to look at it. So you say early on hospitals, excuse me, or retail, were they doing it from a cost perspective

or were they doing it from an availability? Was it because a retail site was far enough away and they needed a generation source? What was the selling point early on? I think a combination. So first and foremost, we can get you to market sooner if you don't have a transmission line or a substation there. These are devices on site. These are microgrids on site. So behind the meter. There's an ESG play there as well. Like I said, dramatically reduced carbon footprint. There are groups, as we know, within companies that look to the future, two, three, four, five years out. So some small deployments, testing it, understanding it,

putting it through its paces. Many of those, after years of deploying Bloom, zero issues switching wholeheartedly to Bloom as the primary power source. Are these hospitals or retailers, are those guys doing instances where they're not on the grid? They're not using- There are those. Wow. Okay.

Absolutely. Bloom's their primary source? Yes. If we take retail, for instance, we have a little bit different definition within Bloom of what we call critical. If I'm at a retail location, critical power may be half of the capacity. Maybe I light every other aisle. Right.

But I'm not bringing down the business. I can still run the cash register. I can still run things, exactly. In the data center industry, as you know, it's all critical. We call critical IT, but mechanical is just as important as the data hall. So, yeah, primary power, no utility involved, no substation there. Working with the utility, we can load follow. You can use the utility to shave the peaks. Your hot environment, for instance, in our world, if you're cooling like crazy.

Can you use the utility as your backup? You can use the utility as the backup. I mean, there are some economic challenges there. I mean, you'd be paying reservations. They're going to say they're going to have to make you take many load fees, yeah. But we do have data center customers that more look like a 2N than an N plus 1, where we have Bloom as primary power,

static transfer switch in the middle, still a generator involved, but utility is the first point of fail. Gotcha. Right. So we go bloom to utility to gens. Right. So that's a deployed architecture today. Gotcha. So as I think about it as a sales guy, I think, okay, I've got a green angle that I can sell this on. Hey, I'm taking combustion and I'm taking carbon emissions out of the equation. You can take it completely out if you want. I've got a time to market equation.

I probably don't have a huge, I mean, I have some markets, but I don't have in every market an economic win, right? There are some markets where I'm going to be more expensive, the bloom route. Is that fair to say? Absolutely. If the decision is being made solely on price, I'm going to have a tough time in Texas or in Phoenix. I think of hydro markets, right, where they're already a fairly green source and they're fairly inexpensive. So there are places where this makes a lot of sense and places where it may not be a perfect fit. Purely economics. Yeah, that is a reality. It's interesting though.

I'm personally and the company is seeing a shift where the economics are, of course, critical. But Bloom is still being the choice for primary power because of the ESG angle, because of the ability to switch to hydrogen when the distribution is there, when the production is there. So that's a carbon-free aspect of the device. Economics also take a little bit of a backseat if I can get you to market one or two or three years. Time matters a bunch, right? Time to power is massive right now. Yeah, let's stick a pin in that one. We'll come back to where we're at globally from a power availability perspective and time to market, because that one, I think, is a big selling point for the folks in our business. So we'll do that one in a minute. As I continue to think through the why am I selling this? Why are people winning? Why are you making hay with this? How big can we get? Because 300 kilowatts is great, but the scale of people are developing today, we're talking tens and in some instances hundreds.

So tell me about the success you guys have had in large-scale deployment so far. Yeah, Bloom, it's completely modular, the devices. So the energy servers, like I said, are installed on a common bus, hundreds of megawatts, hundreds of megawatts. I mean, there is... Individual installs that are hundreds of megawatts? Close to it. Tens at this stage. Oh, yeah, tens.

Okay. Definitely tens. All right. All right. Interesting. And for that, I'm guessing you're doing something special from it. Because I would think, as I think about these servers, if you had tens of megawatts with them, you're covering acres of land. So how are you guys handling that today?

There's multiple options there, right? Yeah, if the land is available and easy, we have some data centers in Napa, for instance, where they have enough land. If it's constrained, Asia, for instance, and definitely in California, we have a designed power tower, we call it. Essentially, it looks a little bit like a parking garage, but the devices are stacked. Gotcha. Right? So it is another structure, but again, it's replacing or supplementing the substation. So there are multiple ways to install it. If it's a brand new build, you can even put them

on the roof. I mean, they're heavy devices, but you can do it. So thinking if I don't have a substation, can I get a power tower in about the same footprint as a substation? The same, about the same footprint. Yes, footprint's the right term. It would be higher. Yeah, it's going to be bigger. It'd be higher, three or four stories, let's say. How are the AHJs handling that with you guys? How do they look at it? The power tower and just your footprint all together. You guys having challenges? Because I think about it, you're not generating, so you don't need to get permits from an emissions perspective. Do they look at you and go, hey, no problem from an appearance perspective? It depends. It depends on where you're at in the country and in the world.

But great point. Do they think of you as the yard? I mean, I'm just wondering how, because you're not really generating, but I also got to think you're not a... It depends on the utility. It really depends on the utility. Some welcome us with open arms, some not so much. Bloom works through all of that with you. We have folks that work through the legislation, that work with permitting, work with design. So yeah, we help you through that whole process.

But I'm thinking of it from a sales perspective. Hey, yes, there's a green view to this because I don't have emissions. But I also think from a planning and permitting perspective, right? If I don't have to, because there's lots of, especially here in the US where we have trouble getting sufficient permits to do generation, this could be another reason why I go, hey, Bloom makes sense because I don't need emission permits. Yeah, this is true. I'm assuming, right? You don't need any emission permits. No NOx and SOx. If you're in California, there's no CEQA process to go through. The amount of particulates is almost immeasurable. There's essentially none, knocks and socks. We do as a function of the electrochemical process today, if we're using

natural gas or biogas, we do release some carbon dioxide. That's typically 20, 25% less than a standard utility would. That's back to the time value of doing something now, so we can reduce now. And again, the hydrogen roadmap is already there. Yeah, you mentioned hydrogen earlier. So what are we using in a fuel source today? And tell me what's the hydrogen roadmap look like? I think we're using natural gas, right?

Natural gas or biogas. So the device is actually based on hydrogen today. So within the bloom box, we have what's called a steam reformation process. So we take in the natural gas or the biogas as an example. We're extracting from that essentially the CH4, the hydrogen, and that's what we're using to produce the electrons. So that is much lower emissions, like I said, the hydrogen roadmap. So Bloom is essentially using hydrogen today. But if we go up to a 20% blend today, if you have hydrogen available, along with the natural gas, the product that is all hydrogen, it requires different fuel rails, for instance, different materials in the fuel rails. Each fuel cell has fuel coming up each side of it. Those ports or ducts need to be a different size. Hydrogen is less dense.

Based on what you're using. Yeah, based on what you're using. But it's not a lift and replace. place. So during the normal maintenance cycle with the energy server, we would take one device offline, which still leaves you with your peak load. We size them so you have that redundancy. We essentially remove the guts of the devices and put in a hydrogen. I got you. So I got natural gas guts. I take those out. I put in hydrogen guts and I'm back up and rolling. And again, those are going to be different fuel lines, stainless steel lines, things like that. So as a developer, I'm thinking about that. And I think, okay, the power infrastructure is really

clear, right? I understand distribution lines and generation and substations. And I got that infrastructure really well. I think I get the natural gas infrastructure fairly well, right? I think it's pretty well built out. I mean, I think we'd like more, but pretty well built out. I have no clue. What does the hydrogen distribution system look like? I have no, I mean, is that a thing? Can you run hydrogen lines to your site? I mean, does that happen? It does happen. Not so much in the U.S. yet. Places like Singapore, there's quite a bit of production. No, you hit it on the head. So that's what I mean by the time value to lowering carbon now. Hydrogen distribution is just not there yet.

Hydrogen production- There's not a, hey, where's the nearest hydrogen point? Oh, it's just two meters up. All we got to do is hit the distribution line. That's not a thing, right? Yep, that's not a thing. So are you bringing it on trucks or what are you doing? No, actually today we're natural gas and biogas.

Where it makes sense for hydrogen is really, I would say, three to six years from now, if everything comes to fruition with distribution and generation, and of course, the economics of hydrogen. You're ready with Bloom today. So when and if that occurs, you're generating power today, you're powering your critical facility, four years down the road, five years down the road, during a normal maintenance cycle, we can replace the natural gas-based innards with the hydrogen-based. Yeah, I got you.

Same technology, though. Nothing's changed. The system works the same. It's just different emissions. Do you get any performance boost from hydrogen? No, we wouldn't get performance boost. Does my 320 turn into 450 or anything like that? No, no, no, no. Not a boost. Hydrogen is much less density than natural gas. So essentially

the same, maybe 1% or 2% lower. I don't know that exact number, but you're carbon free. You're carbon free day one. Before we talk about the global power situation, I'm still fascinated by the Bloom technology. What's the roadmap look like? Because 320 is great, but our industry keeps taking down more and more power. What's the roadmap look like? You mentioned at the beginning it was a 20 kilowatt in a single device. What are we looking like? What do you guys, do you have a published roadmap? What does it say today? I don't know that it's published. Good question. I would say that you will see some 20% plus boosts in density.

Our technology is showing. Talk about published roadmaps. Yeah, yeah, yeah. Let me tell you what we're going to do to our chips in 2027. I'll go back to my software days and I'll give you a month. That's right. That's right. In June, we'll have this.

That's right, yeah. Q2 of 27 is our target. That's our target. No, it's yeah. We will see some significant capacity boosts, like I said, 20% plus in the future, not too distant future. Okay. From Bloom. Yeah. I got you. And if it started at 20 and we're at 55 and you're going to get another 20 percent, are there are there advances? And I think about, you know, in our world, chips and microprocessors and more, you know, is the technology having that kind of advancement or can you guys keep doing it? Or is it, hey, we look at these things and this thing is going to be a 500 kilowatt thing. And that's kind of what we're aiming towards. And it's never going to get bigger than that. Or are there advances happening?

That's a great question. No, there are definitely advances happening. It's the base product remains the same. The solid oxide fuel cells remain the same. There are things we can do with different types of cathodes and anodes that we actually print on to the electrolyte. You know, is it infinite? No, I wouldn't say that it's an infinite growth it's not like a moore's law of uh of power generation but i think the real advantages come around the economics involved right as as bloom

grows the economies of scale become more and more and more favorable yeah so it's a it's a great point we're manufactured 100% in the United States, in Delaware, and in Fremont, California. We're ramping to be able to hit, avoid any backlog issues. We're ramping up production right now. Yeah, I would imagine backlog's coming for you guys, the way that we'll get to the power conversation

in a little bit. Yeah, but I could see how you guys have, people are starting to ring your phone, I imagine, a pretty good bit. Yeah, absolutely. Yeah, yeah. Before we get to the power conversation, how big are we talking?

One of these power servers is eating up how much space? And the other question, how heavy? I'm just interested, can I put this somewhere or do I need a foundation for it? Tell me how big and how heavy the thing is. Yeah, so approximately one meter squared from a footprint perspective. For one power server. Well, and power server is actually eight Bloom devices. So eight of those, eight one meter squares. Okay. Yeah. Eight one meter squares,

essentially. But two of those eight are things like inverters or things like fuel filters and some of the microcontroller equipment that we need, six of those are producing power. Okay. Right? So square meter, about seven feet high approximately. These are in ruggedized. A big refrigerator.

A big refrigerator. Yeah. Maybe a double door type of thing. They're ruggedized. They're meant to be outside. We can deliver them skid mounted. Otherwise, they're on a pad. They can be on heavily compacted ground.

So you can deliver them skid mounted. Otherwise, do you bring me eight units and I set it on a pad? Is that how it normally goes? Typically. Okay. Typically. And then we would obviously string those together onto the bus and get to the capacity you need. We always size for failures.

It's a highly redundant and very granular system. So we're ready to take one offline to do maintenance. We can handle a fuel cell outage if we had that for some reason. Like I said, no moving parts to speak of. Who's monitoring that box? Is it tied into my BMS? Are you monitoring it?

How am I tracking what goes on inside that thing? Bloom is absolutely fanatic about monitoring for several reasons. Obviously, the break-fix aspect of it, of powering critical facilities. We need to know if something is wrong. Predictive analysis is huge for Bloom. Using ML, AI, looking for future failures or future warning signs that there could be a failure. Now, that being said, your folks in the SOC, your folks on site, your ops folks, yeah, we have Modbus integration set points you can view. They can see if it went from green to red to yellow and see what's going on. Bloom will obviously let operations know when we're coming

out to do something. So the answer is both. We have a tremendous amount of monitoring, auto alerts, all of the above, predictive analysis, but your folks can look into the box as well from an alert perspective. I know you don't want to talk about bad things that happen, but can you give me a perspective on what does, in our world, we lose power, failure, we're managing the power, we're managing redundancy. What does bad look like in a bloom world? What does something really terrible happen and that system went offline?

What does that look like? You know, to bring the entire system offline is pretty much going to be lack of fuel. Okay. Okay, we lose a major natural gas pipeline. I was about to say, but I'm running off of a pipeline that's underground. Exactly. Okay. We, we lose a major natural gas pipeline. I was just saying, but I'm running off of a pipeline that's underground. Exactly. Exactly. You're, you're not up on aerials, like with power where storms and car accidents, yeah, they're not gonna have a truck crashing in my substation or something like that, or the utility substation. So much, much more stable. There are many examples where utilities have been down and Bloom had absolutely no impact.

But yeah, lack of fuel would do it. We're converting molecules. If we don't have those molecules, I can't give you electrons. The modularity of the system, though, we have some pretty cool pictures of construction equipment falling over and crushing a couple of devices. One or two of them went offline, but the other ones remained up. They're all independent. They're just talking on a common bus.

You're making me think of my AS400 days, right? Yeah, there you go. The box is half full of water and it's still running. It's that kind of stuff. It's that kind of stuff. Yeah, I got you. Exactly. Interesting.

All right, so let's talk a little bit about, you've mentioned a few times, how fast you can get us power. Because today in our business, right, in the data centers, we have, you know, living through COVID, you know, this huge spike in demand for our customers. They all built tons of capacity. We all went onto the grid and committed a bunch of power. I mean, I think everyone in our space, pretty familiar with what's going on in Northern Virginia, right? Novec and Dominion struggling with, hey, how do I get sufficient distribution out?

So that's driving a lot of us to ask the question, OK, if I can't get it off the grid, where can I get it? And I've got to believe that's making your phone ring. So talk to me about, you know, in markets where I'm getting told three or four or even five years for power, how big can you be? How fast can you be? How fast can you be? How you managing that? And is that starting to drive activity in your business? Yeah, absolutely starting to drive activity in the business. Bloom has been planning for this, not this type of event, but this type of uptick, right? So we're bringing more and more manufacturing capability online all the time.

I think we started with an annual ability of about 200 megawatts, and we're gigawatts now of production capability out of Delaware and out of Fremont. How long does it take to get on site? So, you know, I would say months versus years. The reason I say that is it's not difficult to get the devices produced and shipped. And I'm asking a scale question. So if I want 30 megs of production and I make the call and say, I need 30 megs and I need it in South Texas, what's it look like? You know, six to nine months, I think we'd be on site, installed, ready to rock. Some of it depends on permitting, obviously. So there's outside influences that can cause delays, right?

But it is months versus years. And we're doing it natural gas, right? That's the way we're doing it today. Natural gas or biogas. Yeah, okay, gotcha. You got it. Nine months compared to, I mean, cause that's the problem we're running into right now, right?

Is everybody's, you know, in most markets, we're measuring everything in years. So what does the agreement look like? Am I buying the device and I can generate all the power I want? Am I paying just like I'd pay a utility? Do I have a meter? What does it look like when I go economically, how are we doing this? Yes to all. So it can be a capital acquisition, Of course, you buy the gear, you buy a service contract for maintenance and things like

that. You're procuring the natural gas from the distributor, from the utility. That's a perfectly viable model. Most common for Bloom is a power purchase agreement. So hold on, I want to make sure I understand that one. I contract with the natural gas provider. He's running a meter for me. That's feeding my equipment. And what I get from electricity out of the equipment is mine. And you guys don't care because you sold me the gear. That's right. I'm paying him for the gas.

I bought the gear from you. And what I get out of it, I get out of it. That's right. And I could get the maximum throughput or I could get half. I mean, it doesn't really, at that point, you guys don't care, right? No, we don't care in that example. And I say don't care, meaning economic. I know from a service customer provision, all that stuff, but I'm paying the gas provider for the gas utilization and I'm getting as much energy as I can get out of my box. That's right. That's right. We're behind the meter at that

point. Okay. And I can do that in a capital arrangement with you. I can write you a check for it. You can absolutely do that. Okay. I got that one. What's another way I can do it? Most common is a power purchase agreement arrangement where Bloom looks like a utility. We're charging you a fixed rate. You're charging me per kilowatt hour? Yeah. exactly. Exactly, and the reason that's attractive is it's, we can do things with futures on fuel, for instance,

we can help you level out, so it becomes a very predictable cost. And you're talking about from a source, from a gas source perspective, okay. And that's not the only reason, it's OpEx versus CapEx, depends on your model. As you know, some managed service providers may prefer an OpEx model to a CapEx model.

Or if you're dealing with the end user directly too, right? He may want an OpEx model. That's a great point. The tenants in our world are absolutely influential in this. They care about the environment. They're driving these terms in the leases, as you probably know. They wanna understand the carbon footprint of the environment. They're driving these terms in the leases, as you probably know. They want to understand the carbon footprint of the building.

They want to understand the carbon footprint of the power generation. That is just another great benefit of Bloom. We check all those boxes today. So I can capitally buy it, and I pay the gas company, and you guys have a service contract with me to monitor my gear. I can have a power purchase agreement with you and I'm assuming all my service and all of that is included. What happens to the gas company in that model? Are you guys contracting with them and it's all covered in my rate or do I still contract with the gas provider? Or I can do it

either way? Either way. Okay. I got you. So if I just pay you a price per kilowatt hour, you can take care of everything for me if I, for lack of a better word, a turnkey solution, right? You can do, and you'll handle the gas company negotiation for me. We will. We'll be right there with you. Absolutely. Yeah. It typically will look like perhaps almost like a joint venture on this power purchase agreement. And then exactly, we charge you per kilowatt hour, fuel can be included in that. We can work on hedging if it's a volatile environment for fuel. It's essentially just an economic model. How do you want to see it? Bloom is very flexible. Got it. We tap dancing around a little bit. I've got customers that want to be in my building and they drive a lot of how I do it, right? They have pretty strong opinions about how I do it from an ESG perspective. Today, how have you guys handled the customer base from a, you deal

directly with the customer. So I think of a hospital or a retailer you're providing and you've got your customer and it's really clean. In the service provider, because you mentioned MSP or guys in my business that are building buildings, real estate guys, how are those three party, for lack of a better term, transactions? How are those going? Have you done a lot of them? How is that looking for you guys? Is it something you're good at yet? Where are you at in that process? Because I get when you go to a retailer and he wants power, there's two parties and it's pretty straightforward. Yeah, it's pretty straightforward. But I think of a service provider or a developer, either one, I'd like to understand how's that going for you. Yeah, someone walking into a hardware store to buy a hammer is not really

going to be concerned with how that hardware store is receiving their power. Not true for our world. Not in my business. My business, they care. Yeah. Yeah. No, I think that you hit on the head exactly why Bloom has this vertical, the data center vertical. There are multiple influencers, right? There are tenants. There are design shops, right? You've got the Corgins and others. You've got GCs like Holder and Turner and Hitt and everybody that builds the buildings.

Then you've got the landlord, the compass of the world. Then you've got the landlord, the compass of the world. Then you've got the tenant and you will be influenced by what the tenant says, right? You care about that lease. So that's exactly why we have a vertical for data centers. So it's absolutely number one top priority for me to articulate the Bloom value to the tenants,

to the design shops, to the GCs, even to the operators. Are we good at that today? No. No, we're not. The world, the tenants need to know that we're here and there's another alternative. There is a way to reduce your carbon credits that you're buying today, right? And that's with us. So that's the number one priority for me personally it's messaging and branding to the influencers um the the the data center world is is fairly consolidated as you know we have you know 15 to 25 big players and then it drops changes dramatically after that yeah yeah after that and it's very very global is is jeff is is it too much for me to say is that this is part of what bloom is building this division for is to get in that conversation in that, I hate the word ecosystem, but I'll say it, in that data center ecosystem where you've got our developers and our design firms and our customers and our providers. Is that why this division is being stood up as Bloom says, hey, this is a real market for us?

I think that's a very big part of it. And I appreciate the humility. You say we're not good at it yet. You guys are starting to figure this part out. Hey, this business needs us. That's a very big part of it. And I appreciate the humility. You say we're not good at it yet. You guys are starting to figure this part out. Hey, this business needs us. We need this business and let's figure out how to do it as a group. Exactly. We have a tremendous team today and our team is expanding very quickly. So I'm bringing in some true data center veterans, some true pros that can speak across the spectrum, right? You hit it on the head again. It exists in large part, this vertical exists in order to begin articulating the message,

branding the product, fixing some of the misconceptions. We're not a battery. We're not a generator. We don't instantly turn on. We don't replace a UPS. There are instances where you could probably do that, but in general, the data center world is somewhat conservative with standards, right? So the company, I think, has some great foresight in trying to get to the influencers. And yes, that's a big part of why the vertical exists. The other part is- Hold on a few seconds, Dean. That's why they're looking at you, Dean. Right?

Leave that in. Yeah, yeah. Dean's going to get to hear his name again. That's right. Oh yeah, Mr. Dean. Oh yeah. I'll be with him, I think, later today. So also you need to, with these types of customers,

whether it's semiconductor manufacturing or whether it's network manufacturing or whether it's network gear or whether it's data centers, your team needs to speak the language of the customer. In the case of data centers and most other verticals, they're global, right? Centralized decision-making typically in the U S but not always, but centralized. They want to know what you're doing in Dublin, Ireland, what you're doing in London, what you're doing in Germany, places where power is even more constrained than the U.S. by far.

And the pricing is not stopping. It's getting atmospheric at this point. It's taking off like crazy. So we need to look like the customer. That's the bottom line. They need to know that they have a central place they can go to. They have a group that speaks their language. They have a group that they're not going to need to educate every time they speak to someone new. I know what you mean when you're talking about PUE, and I know what you mean when you're talking about medium voltage gear and everything else that we talk about. We get it. We get it. So one last thing. Let's hit on the story I think is fascinating. I think I get a pretty good understanding of why it's getting popular and why you're having all the conversations and frankly, why you're here.

Yeah, absolutely. What does the competitive landscape look like for you? I don't think you're a competitor to a utility, right? You're an augmenter or you're an enhancer or you're a collaborator. I don't think you're really a competitor with a utility. So who are you a competitor with? You know, I'll hit the utility one first. I'll take that in reverse.

It depends on the utility. So Bloom has a fantastic focus, similar to the data center world, on the utilities, right? We have a group of utility professionals that are working with utilities to supplement their power as within the primary grid, right? So not behind the meter, in front of the meter this time. So Bloom can actually get the utility there more quickly. It depends on the utility whether or not they see us as competitive or a risk of some sort when we're behind the meter. Very much depends on your location. From a fuel cell, bloom is the oldest with

the most experience i i don't i don't i don't i can't think of anyone with the scale that we have you know hundreds of hundreds of megawatts gigawatts deployed globally global support all of that so the competition i'm not asking you to mention your competitors but i mean is that legit you guys are i know you've been around more than a decade. You guys have got hundreds of installs. It's the most mature and thousands of them. For sure. Okay.

For sure. Absolutely. No one has been doing this longer than Bloom, which is why we have such a head start, in my opinion, from a manufacturing, from a scale perspective, from a global support perspective. Follow the sun, maintenance and support. Like I said, 24-7, always there. A competitor in our world might be gas turbines, might be diesel gens, if I'm looking at Africa, or something like some other form of generation with a massive carbon footprint. You guys see any traction in Africa? I got to think you're infinitely more reliable than their grid. Is that something that's a play for you guys already? Certainly conversations.

Yeah, nothing formal inked yet whatsoever, but some very, very scalable designs that we're working on with a particular provider called Kasi Cloud, K-A-S-I. Dean's involved with that as well. Dean, we see you again. That's through Cato working with Kasi Cloud. Massive demand.

I mean, in the Lagos region, there's 15.4 million people. Yeah. Right? The hyperscalers need to be in there. They want to be in there. So like I said, nothing formal inked yet, but certainly designs underway for how we could generate 100 megawatts of power for Kasi with a much lower carbon footprint

and five nines of availability versus no nines. Versus like three sevens. Yeah. I think it's two eights. Two eights, all right. I think about the first time I went to Africa, I was fascinated by how many people carried cell phones and it was still fairly early in the cell and having conversations with people there. They're like, look, we didn't build a phone network the first go around.

We didn't build a wired phone network. We just skipped over that phase and we just went straight to cellular, right? Because it was a much easier system to distribute. And I kind of think the same thing, right? It's not a terribly sophisticated power generation continent or country. And I could see skipping over that and just seeing you guys. That's a fantastic example. That's a fantastic example, right? No, you're spot on. And there is a true desire to start being green day one. That's right, right from the beginning. This is Greenfield.

We don't need to build a coal-fired power plant. Let's do something else because it wasn't here 80 years ago. So we don't have to start there. Right, yeah. Interesting. Dean's organization through Cato is giving you zero carbon compute

as a managed service. Kasi Cloud is paying maniacal attention to the sustainability of that building, the way in which they cool it, the way in which they build it, obviously, and the way in which they power it, most importantly. So start from day one with a much greener solution. Like I said, sitting there ready for hydrogen. If if it gets there, if we need to go there, it's not a problem for us.

And that's exactly how they're looking at the cell phone example is a great one. It's a great one. I say, well, Jeff, I really appreciate you coming and hanging out with us, even though it was 6.30 in the morning. 6.30. I know, that's brutal. 6.30.

We really appreciate it. Let me mention it one more time. Yes, it's 6.30 in the morning. But thanks for talking with us about Bloom Energy. We're excited what you guys are doing and grateful to see you in our space and see your team starting to understand our space and help us solve the problems.

Because at the end of the day, we get asked all the time about power. I'm like, hey, you keep wanting to order food off of Uber Eats, right? I think you still want your cell phone to work. You need us to be able to be behind it. So we really appreciate it, Jeff. Excited for what the future holds for you guys. Thanks, man. Thank you. Take care. Good to see you, man. Good to see you. Take care.