For our COBT viewers and listeners, it's Maynard, Mike, and Jeff here with something we are very excited about. Today, we're going to get to talk to KR Sridhar. He's the founder, chairman, and CEO of Bloom Energy. One of the most interesting experiences we've had this year is going back to earlier in the year we started getting interested in fuel cells in light of everything that was happening in power. Through good friends we connected with KR throughout the course of the year. We really enjoyed getting to know him, the team, saw the plant, have had a great experience understanding the product and its applications. So let me just pause by saying KR, we've been on a journey with you. We've learned so much. We're so excited about this and we're obviously pleased to be able to share this discussion with our friends about what's going on at Bloom and what's happening in the power markets more broadly. So, thank you.
Maynard, it's a pleasure to be with you and I look forward to our conversation.
Fantastic. Well, Mike, what would you tell us before we jump in with KR?
Yeah, Maynard, what I'd say is I think everybody realizes that tomorrow is the FOMC meeting. Markets are essentially going sideways. Bonds are going sideways. The equity market's going sideways. Everyone was prepared for what's going to happen tomorrow. Consensus is that the Fed's going to cut 25 basis points. And I think consensus is also they're going to be another two 25 basis point cuts before a year. And I think what's really important is not the 25 basis point cut that's expected. The really important part is what is the tone of Chairman Powell's speech afterwards. It's really going to tell you whether there's two more 25 basis points cut, whether it's going to be more aggressive from there and the markets are going to trade off of that. So we're clearly looking at that. Just from the standpoint, if you think about the last three rate cuts that were in the Biden's term, it looks like the average increase in bond yields a week later was about 10 basis points and the markets went up about half a percent. So my guess is, depending on what Powell says tomorrow, the market's going to react. But the last three rate cuts, it's more been sort of buy the rumor and sell the fact type of market. And I think we're going to probably see the same thing tomorrow. As it relates to oil prices, oil prices also are sort of in the sideways trading action. We're in the 62 to 64 range and it's probably not going to move out of this range. Couple things that we saw over the last week that were interesting. And Jeff, we'll talk a little bit about this in a second, but the IEA and OPEC both came out with their estimates for supply and demand in 2026. And IEA surprised everybody. Well, maybe it didn't surprise everyone, but they ratcheted up their oil surplus for 2026 to 3 million barrels a day, which we think is nuts. We do think there's going to be a surplus. Not as big as that. And so that's what the markets are dealing with right now. I say as far as energy companies or energy deals is probably the better word right now is that we've had two energy deals this week. We've had basically California Resources buy Berry Petroleum for $700 million. And we also had basically Court Energy buy some Wilson Assets from XTO Energy for about $550 million. And the only reason we mentioned this is because back to what we said with the IEA and OPEC and oil prices. I think there's a lot of uncertainty about where oil prices will be in the second half of this year and the first half of next year, but it's bearish, right? And so I personally think if you basically have a balance sheet, you have a pretty good long-term outlook for crude oil prices, you probably should take this opportunity to be buying companies. We probably see more consolidation and more asset deals. I think that's going to be a good deal. And these stocks have been rewarded this week for doing those deals. The last thing we'll do is because we got Bloom Energy on today, we're just going to give you a little bit of a perspective of how the stock has done here over the last year or so. Year to date, the stock's up roughly 225%. Over the last year, it's up 625%. To kind of put that in perspective, some of the darlings of the market right now are these small module reactor companies, they're up on average over the last year around 450%. IPs are up around 100%. So Bloom Energy has had a great run here. We're really looking forward to talk about why that is and why it's going to extend. So, back to you, man.
Thanks, Mike. Jeff, anything on the I wish we had the stat for how much time we spend talking about power, but what would you add jumping in with KR?
No, we're excited to have KR on today, talk about really how the story at least in public market is transition. But I wanted to talk just for two minutes on something Mike mentioned. The world and consensus and research, the research community is quite bearish crude oil, but I think it is notable some of the pivot you're seeing in terms of long-term perspective from some of the policy organizations here in the last week or so. Just the industry and the worry about the industry going ex-growth, meaning oil and gas demand over long-term plateauing, is a pretty dramatic overhang on the sector. Because ultimately what gets rewarded in the stock market and why Bloom's been rewarded is growth and profitability. And so if the sector is not growing, that is a meaningful overhang. But you've had first you had Javier Blas at Bloomberg about a week ago preview or write a story that previewing that the IEA is likely to introduce a new long-term scenario for oil and natural gas demand where it's showing growing over the long term. So that ex-growth sentiment from an organization that has had some meaningful bearish commentary seems to be pivoting. And then just today you had another IEA report out just talking about underlying decline in oil and natural gas fields and what the industry is going to have to reinvest in order to offset that. None of that should be surprising, oil fields decline, etc. But it is really important for sectors to be perceived as having a growth path, which power obviously does. The power value chain has meaningful growth expected of it. But just as we think about the rest of the energy companies and rest of the energy world, the overhang of that lack of growth may start to lift. It doesn't mean the near-term can't be bearish or cautious or oversupplied, but you are having some of the tides turn a bit.
Awesome. Well guys, that was a lot, KR. Sorry to hit you with a lot, but there's always great and interesting things happening in energy. I referenced it, KR. You've been one of the real bright moments of our year because you were gracious enough to you and your team spent a lot of time with us and we've really enjoyed getting to know you and the product. As I mentioned, you founded Bloom way back when. So you've seen 24 years of technology and of improvements and the way the markets come and go. And you're right at the center of the power revolution that all of us are experiencing. So let me just thank you again and ask you to start anywhere you'd like as you think about what Bloom Energy looks like today and what's on your mind.
That's such an open question. I've never, you know, I've never, it's either the worst or the best I've ever had. So, I think that's a great place to start because let's just pause for a moment and think about what has happened in the world in the last 20, 30 years, right? We have clearly moved from a mechanical industrial age emphasis to a digital age emphasis. You know, we've talked on and on about digital transformation, but digital transformation is everywhere. A young child almost learns its parables and fairy tales from an iPad, watches TV, plays with a computer, learns things all the way to healthcare for an elderly at home. It's all digital. And the amount of digital penetration into every aspect of human life means there are no longer technology companies, quote unquote technology companies. Technology really is a tool in a toolkit and pretty much every industry, every aspect of life uses technology. And what does technology need? Technology runs on electrons. Electricity ultimately powers them all. And so I don't think there has been a moment since Edison where the need for electricity, its abundance, its affordability, its access, its reliability, and its responsibility as a sustainable electric source all converge together in a way that we are sitting out here. And we are so glad that we are in the business of providing electricity at such a historic moment in mankind's history.
Would you mind because it is a fascinating from when you started Bloom in 2001 to today, like what a dramatic change. In 2001 we were just finishing what you would call maybe phase one, the internet revolution. We were just about to begin the natural gas renaissance in the US. Within a few years, we'd be talking about the Barnett shale. But we also in that period went from power, we were oversupplied with power for much of that period. And the question was just how to allocate it, how to deregulate, to now where we are today, which is like oh gosh, we're short and we need it fast. And we want it to be as low impact as possible. It's an amazing 20-year period that you launched at the beginning of that. I think I'm going to try to build a fuel cell power company. Maybe take us back to the days of founding Bloom and what you were thinking about at the time.
Sure. So 99, 2000 is when the idea of Bloom was entering into my head and I was writing the business plans for it. And here is what I saw happening in the world. What I saw happening in the world is 50 million people were coming from abject poverty to lower end of the lower middle class. $100 a year per capita income to $1,000 a year per capita income. 50 million people roughly. This is China, the seven tigers of Asia, predominantly that being the engine. And it was happening for almost 20 years from 80 to 2000. And I was watching that in awe. But then looking at when you go from that $100 to $1,000, if you actually see the graph, it's a log-log graph, which means if you try to draw it on a linear piece of paper XY axis, it just shoots off the chart faster than you can think. When you go from $100 to $1,000 a year in per capita GDP, your marketed energy consumption goes up thousandfold, not tenfold. And you look at those charts and then you look at the population growth that was going on where we were about to hit somewhere in the next few decades a 10 billion people planet. And I'd been working on Mars missions and knew how difficult it is to rehabilitate or to build a habitat in an alien planet. And here's the one pale blue dot. And I was wondering, 10 billion people on a planet. We want everybody's standards to rise and we want their incomes to go up. We want to eliminate poverty. What happens to that energy needs? And is the old way of doing things still relevant in that modern future if we want our children and our grandchildren to have a great planet left behind for them? So that was the thought process.
And what became very clear to me is two billion people didn't have access to marketed electricity, another billion could not afford as much electricity as they needed for a better life. So if the existing way was good, we wouldn't leave that many people behind.
And juxtapose that against two revolutions we witnessed during the same time. One was the computer revolution and the other one was the communications revolution. The computer revolution. In the early 80s when I left India, there were probably two mainframe computers in the entire country and we were fortunate enough to have punch cards and go get 15 minutes of time in a computer. 1999, the beginning of Y2K, India was a back office and every street corner had people sitting with their laptops trying to figure out how to serve the entire world to make sure that things don't melt during Y2K. Access affordability to compute. Then when I left India in early 80s, getting a landline telephone you would have to put your name and wait for 10 years if you're in Mumbai, I'm not talking about villages, I'm talking about a city. And you go in 99, people are walking around with cell phones. So you look at access and affordability coming in compute and in communications and I said, 'Wait a minute. Isn't that what we need in electricity if we need to go where we want to go?' And what's common to these two things? They went from centralized to distributed. That doesn't mean the centralized network doesn't exist. It still exists today. But the distributed becoming the impetus, the growth vector was the critical factor. So I said, shouldn't that happen in electricity fundamentally in the world? Can the world ever get access, affordability, and ubiquitousness of where it can be with electricity? And if you need to do that, distributed is absolutely necessary. So that was my thesis. I believed it strongly then. I believe it even more strongly today.
Mhm.
Mhm.
And we've been with you when you've done this. I know so much work has happened at Bloom between 2001 and the present, but you have been, these are our words, but you've been refining and improving and tweaking and just driving up the resiliency and driving down the cost. Is there a way to talk about that 24 year period and just what you've done to the product?
Sure. So again just for your listeners, Bloom Energy does solid oxide fuel cells. These are high temperature fuel cells that simply take the chemical energy in a gas molecule and without burning it, without making it run in a mechanical device and create motion, without capturing that motion to make electricity, it just sucks the chemical energy from the natural gas and makes electricity without combustion, without burning, without mechanical parts. So it's a solid state power generation device. Now this concept, the basic physics of that is an 1890s patent, 1890s not 1990s, patent by William Nernst, a German, and it was bought in the US by Westinghouse. Now it's existed, the knowledge of such a material existing and potential to make fuel cells with it has existed that long, and every major corporation in the world that's been involved in making power devices has tried this at one time or another for decades and given up because the material science challenges were so hard.
So young entrepreneurs or slightly older entrepreneur in my case, the one thing that you have in common is you don't know what is hard or why it is so hard, so you just dive into it with all the enthusiasm. And that's what we did. We said we think we understand why everybody else failed. We think we know how to solve those problems and we think this is the right time to do it. So that's how it got started.
So what happened the first 10 years was purely technology development, making sure that a fuel cell can actually produce electricity in reasonable quantities for reasonable amounts of time. Right? And the next 10 years I would say was both early commercial adoption and manufacturing. Can you manufacture it? Can you build a supply chain? None of that existed. So we had to build the machines that can make the fuel cells. We had to create the processes that created the materials for the fuel cells. And then we need to know how to apply all these things. And then you put them in the field. Lab is very different from the real world. It can fail in so many different ways. We had to learn every time it failed why it failed and fix it. So we were flying a plane as we were fixing it, is what I would call it. And it was a concurrent experience and we had confidence. And now I'm going to use another metaphor not to confuse your people, but now you're ready to land your plane, which is you're going to go out in big scale, but you're still building your runway, and you're hoping by the time you land the runway is long enough so you can land it properly, which is you are putting enough fuel cells out in the marketplace and expecting that by the time you have to replace those fuel cells in 2 years, the life will become 3 years. By the time you have to replace it in 3 years, the life becomes 4 to 5 years. So economically that model can sustain itself. So that was the highwire act that we had to play having confidence in our people, our technology and execution all put together so we can make Bloom Energy work.
So that first 10 years, the research focused part of the journey. Do you mind talking about the funding and kind of how you sustained the business during that, what's a hard period to do all that research and before you start manufacturing in sort of a second phase? You know, long-term capital using a venture model is kind of hard to do.
Yes. Right. Because a venture has its own cycle of returning money for its customers, showing its returns what it needs to do. So it was pretty obvious to me that any venture firm that's not seasoned and has a great reputation, has built enough ROI for its LPs, is not going to invest in this. So you can only go to the best of the best because they can afford to take a small portion of their portfolio and invest in something that they fully understand the day they invest is not going to be their typical three-year cycle to return money, is not going to take their typical 10 to 100 million to return money, but probably billion plus and it's going to take 10 years plus as opposed to a three year plus. So having that awareness up front to say this is going to take long, this is going to take money, but it's worth doing it because the payback will be there someday that is commensurate with the time and money that was spent. And there are still people in Silicon Valley that believe that making life better is one of the goals that we have.
Is this an interesting, I mean you're headquartered in San Jose. We've gotten a chance to tour the factory and the premises. We'd love to talk about that. But was this an advantage to being out there because, as you say, the traditional energy community, lots of people had tried this, but you tried it with a different group, a very innovative group, a VC oriented group. Was being in California the advantage you needed to do this?
I think so. I would it would be hard for me to think of any other location where I could have built a company better than Silicon Valley. Let me tell you why. When you ask a bank robber why they robbed the bank, they say that's where the money is. Okay. And so for us it's risk capital and this is the risk capital of the world. So that's one reason to be there, but that was not the only reason. In Silicon Valley you have engineers who are problem solvers. They are innovative problem solvers who are willing to work on problems for which you don't know ahead of time whether there's a solution or not, and they've done it in various fields. So that is more a mindset and a frame of mind that you absolutely need to do something that we were doing because there wasn't an existence proof of a solid oxide fuel cell working for somebody else and we were trying to make it better. The existence proof didn't exist. Right? So when you're in that place, the question becomes why should somebody give up a very lucrative career doing whatever they're doing and come and join you and be part of the journey. You will find those people in Silicon Valley. It's much harder to convince somebody working in a large energy company.
I think the other thing, and maybe this has gone on with others besides us, but for a long time fuel cells were associated with the hydrogen economy and that's where they seem to gather the most attention. And then you know what struck us earlier this year and that helped us recalibrate is we said wait a minute, let's think again here. Let's just think about taking natural gas and making power and doing it in a way which is advantaged and is attractive. When did, how do you think about the hydrogen chapter? Was there a point where things started changing or was it always part of the fuel cells your mentality that hey we could do a lot with this, let's don't get particularly focused on hydrogen? Maybe talk about that piece of it for us.
So between my co-founders and me, we had done hydrogen fuel cells for underwater for submarines, on the land, low earth orbit for space shuttle, Apollo mission, and I had landed something on Mars. So we had just covered the gamut, right? We were not starting Bloom to do a science experiment. We're building a company. And what was very clear to us were two things. Number one, you have to use natural gas as a fuel and not hydrogen because hydrogen is not a native fuel available on the planet and it's very expensive to make it. And if you make it, converting that to electricity is probably not the smartest thing to do with it in the short period. Somewhere 50 years from now, that'll be the case. But not today, right? That was number one. Number two is the low temperature hydrogen fuel cells all had some kind of a noble metal, precious metal in its bill of materials. For us in Silicon Valley, that's anti-Moore's law, which means the more you make of it, the more expensive it'll get because a limited quantity. So if you needed to go to a fuel cell that did not need precious metals, you had to go to high temperatures. If you're going to go to high temperatures, then we said, let's crack the hardest code there is to crack or go home. Let's not go for sub-optimal solutions. What did we say? On the fuel cell, we can strip the hydrogen off the natural gas because that's what's available in the pipeline. And let the fuel cell make its own hydrogen and eat its own hydrogen from natural gas. So we did steam methane reformation right on our fuel cell without needing a reformer outside. Very, very difficult thing to do. People thought we were nuts to try it. But I said if you don't solve that, you're not solving a fundamental problem.
And the maybe the power piece of this because we should get into, we had a good friend tell us because we're all fans of natural gas around here for all those reasons you're saying. It's nothing's perfect, but it's a great fuel. But this friend shared, people forget natural gas, not combusting it and using it chemically like you're using it, that kind of helps you get the best of both worlds. You can have the portability and the energy content, but you can avoid the emissions. When did you see at Bloom, hey, with the power revolution that's taking place, our phone's going to start ringing because here's the breakout use of our technology.
So there were two things, right? So we apply Moore's law in our own way in the fuel cell business and what we have done since we started commercializing the product and shipping it out to now, on a yearly basis we roughly have a 10% cost reduction. So early technology like the cell phone, like the laptop that we talked about before, is very expensive. Every year it gets better as it becomes slightly more affordable. More people can afford it. That scale allows you to go to a lower cost. That's the virtuous cycle. That's Moore's law. So we said that's what we need to do. And we have been reducing our cost by double digits every year. So about 3 years ago it was very clear to us we now had, on a non-GAAP operating income level, a cash flow positive business from that perspective. Okay, non-GAAP operating which is what we need from a business perspective. And therefore now we can scale and sell. And the reason we are able to do that is in certain pockets and niche places in electricity market that's trillions, is still hundreds of billions. So nothing to sneeze at. We can go compete in that market and we can grow. So we were very confident of a 30% growth and that's what we are targeting, building the infrastructure.
As far as electricity is concerned, when we look back you'll say pre-GPT and post-GPT. Once ChatGPT came out and the large hyperscalers said we are going all in, pretty much that changed the game for electricity for the foreseeable future.
And then much like us, a lot of people are figuring out, well certainly in the last year, how competitive your product is for someone who needs power. They need it fast. It has to be reasonably priced. Do you mind just talking about if I had a data center and I showed up and I said I really need this and I need 50 megawatts of power, I'll just pick a number. Maybe just talk to us about your product and how it meets that need so well.
So let me preamble that with something very small so your listeners can understand the context and then I'll build on exactly the question that you asked me. So what's becoming very clear is these AI data centers, both training and inference, are going to be consuming so much power in such a concentrated area. These are AI factories. They are manufacturing intelligence and manufacturing intelligence is probably the highest value added product we will ever make with electricity. Okay. So these are big factories. Very similar to other power intensive industries, you know them all, whether it's steel and metals, whether it's refining, whether it's large chemicals like BASF, whether it's plastics, whether it's glass, cement, they all have their dedicated power plants. There is no choice but for these AI data centers to have their dedicated power plants located where they are. Yes, we have a shortage of power through transmission and distribution. That's all true. That adds to the problem. But independent of that, we have never in the history of mankind done such power intensive factories to just be able to take power off the grid when they want it and not take it off the grid when they don't want it. It doesn't work that way. So you need on-site power. So now let's talk about Bloom as an on-site power source and how do we compare with alternatives.
Someday, you know, small-scale nuclear, other things may come along, but in the short term, at least in the US, there's only one choice: it's natural gas to electricity. That's the molecule you need to convert that to electricity on site. You have turbines, you have engines, and you have fuel cells. Everybody knows turbines and engines. Very few people know fuel cells.
Okay. So if you're doing on-site power, then it's capex, the cost of the equipment fully installed, the reliability that you need at least grid level reliability, which is 99.9% — what we call in the industry 3 9's reliability. Many will require 4 9s and 5 9s reliability. So you need — and on the operational cost, it's predominantly fuel cost.
Okay. Bloom in the US will be comparable to what today's on-site power generation prices are for turbines and engines on a dollar per megawatt basis, dollar per kilowatt basis. Capex installed, post our ITC. Okay. Anywhere else in the world outside, we will be higher in pricing on the capex side compared to turbines and engines.
And can I — for, would you mind if for our viewers I think roughly those numbers let's say $2,000 — you know. Yeah. Yeah. Yeah. 23, $2,400 installed.
Okay. And then that's as you say that's after the ITC, right?
After the ITC.
The current tax law that's deductible. We have that through early 2030s.
Exactly. Exactly. So, so that's equal. However, we are 99.9% or 99.99% reliable with just a 10% overbuild, which is what the data center needs. Engines and turbines will need 30 to 40% overbuilt. So if you look at that reliability that they need at the end of it, we are actually cheaper on capex, number one. The second part of the electricity cost is fuel cost. Engines and turbines will be in that small scale — I'm talking about now aero turbines, 50, you know, 50, 60 megawatts, or engines, you know, small engines — they will be in the 35 to 42% range. We are in the 54% range. So we are a lot cheaper in the fuel cost, so we are comparable in capex, we are lower in opex. Now take that 20% plus 20 to 30% plus efficiency gain, or fuel gain, that you have with Bloom, and take it to a place like Europe where you're paying an obscene amount of money — you know, like a dollar per mm BTU — we don't need the ITC, we will still compete there. So on a cost basis, we're at least as comparable, if not better, on a TCO basis.
So this is really, just to add a little to that in case as people are digesting this high gas price environment, the more competitive you are. But this piece — and Rob West has written about this — this piece about reliability, and if you really want absolute certainty that the power is going to be there, and you've got these numbers in a paper that just came out, KR, that you put out which is quite good. I'll include it as part of this distribution, but the necessary overbuild to make sure that you have the reliability — it's just much less with the fuel cell. And so the per unit pricing is misleading because in the reciprocal and the turbine example, you're going to have to buy more equipment, which is more equipment.
Quite fascinating.
Exactly. And there's another very important point to make here, which is the following: may not, today, we are only focused predominantly in the public domain about these large gigawatt-scale or hundreds of megawatt-scale training data centers. The value really is going to come a few years from now from the inference data centers that are the 10 to 30 megawatts that are located close to where people, machines, and processes are — highly congested cities is where they're going to be located. And it'll be almost 90-10, 90% of all the power needed for AI will be in inference, less than 10% will be in training.
Now think about those areas. Air pollution with combustion technology versus no air pollution with fuel cell technology means in non-attainment zones, we are the preferred choice.
This was what someone shared with us, KR. They said to this point, look, we have other ways of making power, we can honestly make it a little cheaper, but when we get in an area where we've got to get the air permit, it would be great to utilize the Bloom product because even though the power might be a little more expensive, we can get the air permit faster because there's no NOx and SOx, as you say — there's no, you're not combusting the natural gas. So the footprint is so much less. And so that was what struck us, because you do have these small units and you can scale them to the size that people need, so it can become part of a portfolio quite easy.
Exactly. And you know, in the past, people used to worry about the footprint, and again over the last 15 years, you asked about the learning process. We understood that footprint is super important. Today we compete on footprint with either a turbine or an engine. So you don't require any more land, and the fact that we don't have a smokestack or we don't have anything hot coming out of there and we don't make noise means that you can locate this much more efficiently, you know, land use efficiently, compared to other systems.
Okay, our — in our lead, and we talked about stock being up 225% this year to date and 625% over the last year. And I guess my question is, I think investors realized that you were getting more cost competitive and realized that the performance was improving, scalability was there. I guess my question is, what was the aha moment for people to finally say okay, something really interesting is happening here? Was it a set of investors coming into your stock? Was it current investor? How did things play out that we're seeing as performance? Now, what is that moment? Is it a speed-to-power moment? What are the main factors that have, in your mind, basically attracted investors to the story?
I would never try to venture a guess on why investors do what they do. That's not my area of specialty. But, you know, here are a few things that happened within the company, right? Number one is we were winning over customers on a consistent basis and having gotten the AE deal where a utility is now coming and saying, we don't build nuclear power plants, we don't make gas turbines, we don't make fuel cells, but if your solution is the best option for our data center customer, then we want to adopt you. So we applaud AE for taking that position and satisfying their customer and saying, why should you disintermediate us from our customer? We would rather buy from you and serve our customer. That's the right model. So number one, that happened. And what did they do? Their first 100 megawatts are going to two different data centers, AWS and so on. So there you go. Then the second thing is, okay, again the skeptics, maybe that's just one. And then they saw the Oracle deal.
