WEBVTT - ICYMI: AI Powers Green-Tech Surge 0:00:02.520 --> 0:00:07.040 Bloomberg Audio Studios, Podcasts, radio News. 0:00:08.200 --> 0:00:11.920 You're listening to Bloomberg Business Week with Carol Masser and 0:00:12.000 --> 0:00:14.400 tim Stenoveek on Bloomberg Radio. 0:00:14.840 --> 0:00:16.239 One thing we like to watch is what's going on 0:00:16.280 --> 0:00:17.959 in the green space. Green firms, by the way, in 0:00:18.000 --> 0:00:21.160 the US have found something of a lifeline in artificial 0:00:21.160 --> 0:00:24.599 intelligence after being bogged down by high interest rates, shrinking 0:00:24.640 --> 0:00:28.800 funding and more recently President trump sharp rollback of support, 0:00:29.080 --> 0:00:31.639 and our own Bloomberg New Energy Finance team. They have 0:00:31.840 --> 0:00:35.200 recently pointed out that clean tech companies that have INK 0:00:35.280 --> 0:00:38.000 deals to support data centers have seen their stocks sore 0:00:38.080 --> 0:00:40.800 this year, outperforming the S and P five hundred, and 0:00:40.800 --> 0:00:43.559 that includes the share price of our next guest. We're 0:00:43.560 --> 0:00:46.360 talking about fuel sale provider bloom Energy. It is up 0:00:46.479 --> 0:00:49.279 seventy six percent year to date today trading at an 0:00:49.280 --> 0:00:52.640 all time high, rising another eight point six percent. Jon't 0:00:52.640 --> 0:00:54.200 ne us to talk about the impact of AI and 0:00:54.280 --> 0:00:57.960 just the business overall. Is doctor K. R. Schreider, Founder, chairman, 0:00:58.000 --> 0:01:00.960 and chief executive officer of Blue Energy. It's a nine 0:01:01.000 --> 0:01:03.279 and a half billion dollar market cap company. By the way, 0:01:03.320 --> 0:01:06.320 he was a former advisor to NASA and so much more. 0:01:06.520 --> 0:01:08.200 He joins us from San Jose. We could go on 0:01:08.280 --> 0:01:10.720 and on about your background, but we really want to 0:01:10.760 --> 0:01:12.560 talk to you. Good to have you here, it's been 0:01:12.600 --> 0:01:17.040 a while. Talk to us about how AI is specifically 0:01:17.120 --> 0:01:19.320 impacting and driving growth at your business. 0:01:20.880 --> 0:01:22.560 Carol, it is such a pleasure to join you. 0:01:22.959 --> 0:01:27.959 So this is an amazing time ya, in human history 0:01:28.000 --> 0:01:30.840 where we have not seen the need for power growth 0:01:32.000 --> 0:01:34.920 anywhere close to what we're seeing today. Let me put 0:01:34.920 --> 0:01:38.760 it in perspective, the amount of power that the big 0:01:38.840 --> 0:01:42.160 hyperscalers are saying they need between now and twenty thirty 0:01:43.280 --> 0:01:48.200 is equal to all the power every single family home 0:01:48.400 --> 0:01:51.559 in the US put together would need about eighty five 0:01:51.560 --> 0:01:55.400 million single family homes give or day. All the power 0:01:55.440 --> 0:01:58.440 that they use is going to be needed in addition 0:01:58.560 --> 0:02:01.480 to the power we have today to just power the 0:02:01.480 --> 0:02:06.120 AI data service. That's how large this opportunity is. And 0:02:06.280 --> 0:02:11.200 very clearly the old model of centralized power plants with 0:02:11.320 --> 0:02:16.280 transmission distribution simply cannot meet that demand in a timely fashion. 0:02:17.240 --> 0:02:22.560 That is the opportunity, and Bloom was purpose built to 0:02:22.680 --> 0:02:26.200 provide on site power for the digital revolution, and we 0:02:26.240 --> 0:02:28.240 are so happy to be in the right place at 0:02:28.240 --> 0:02:29.480 the right time well. 0:02:29.400 --> 0:02:31.959 Help us understand how the company works. And you mentioned 0:02:31.960 --> 0:02:34.840 that on site power, so what that looks like. Of course, 0:02:34.880 --> 0:02:37.679 there's been so much conversation recent months about perhaps the 0:02:37.760 --> 0:02:41.560 need to invest more in nuclear power or fortify the grid, 0:02:42.040 --> 0:02:44.280 bring more power plants online. How do you see your 0:02:44.280 --> 0:02:47.360 company as maybe not a substitution for that, but a 0:02:47.400 --> 0:02:50.359 compliment to the kind of growth and develop we're likely 0:02:50.400 --> 0:02:52.160 to see here in the power grid going forward. 0:02:53.400 --> 0:02:55.720 I think we're going to need all forms of power 0:02:55.760 --> 0:02:59.399 in very large quantities going forward to meet our ratios, 0:03:00.040 --> 0:03:02.919 you know, like energy needs. However, in the short term, 0:03:02.960 --> 0:03:07.040 if you take five to ten years, natural gas to 0:03:07.200 --> 0:03:11.360 electricity is going to be the single largest contributor to 0:03:11.400 --> 0:03:16.639 this additionality what Bloom does. It takes natural gas without 0:03:16.680 --> 0:03:20.760 burning it, without combustin. We extract the chemical energy in 0:03:20.800 --> 0:03:24.680 it using a fuel cell and make electricity at the 0:03:24.760 --> 0:03:31.080 highest efficiency without air pollution, without noise in our generators 0:03:31.120 --> 0:03:35.040 because they have quiet, solid state and a very high reliability. 0:03:35.480 --> 0:03:40.880 For those reasons, if you are in a neighborhood that 0:03:41.000 --> 0:03:43.720 is going to have a data center, we are your 0:03:43.720 --> 0:03:47.080 best choice because we don't pollute your air, we don't 0:03:47.160 --> 0:03:50.880 use water, and we don't make any noise as we. 0:03:52.400 --> 0:03:54.400 Generate this reliable power, so you. 0:03:54.400 --> 0:03:59.280 Can have the economic advantage that a data center brings 0:03:59.680 --> 0:04:03.480 to the local community. At the same time, you're not 0:04:03.560 --> 0:04:04.920 compromising on your health. 0:04:05.160 --> 0:04:07.200 So one thing I want to ask you, actually, our 0:04:07.240 --> 0:04:09.360 Will Wade does who really covers this space, and we 0:04:09.360 --> 0:04:11.320 were talking about you coming on and he said, well, 0:04:11.320 --> 0:04:13.520 wait a minute, let's talk about these fuel seals for 0:04:13.680 --> 0:04:17.640 data centers. Can fuel cells really work for this? What 0:04:17.680 --> 0:04:21.440 evidence have you guys seen that it does really work? 0:04:22.600 --> 0:04:26.360 So, yeah, that's a great question. And this is really 0:04:26.400 --> 0:04:27.039 smart dude. 0:04:27.080 --> 0:04:28.920 I'm just going to put it out there. He's amazing. 0:04:29.000 --> 0:04:30.080 But anyway, go ahead. 0:04:31.440 --> 0:04:32.080 Absolutely. 0:04:32.279 --> 0:04:38.200 Look, the issue is we are progressing so fast in 0:04:38.240 --> 0:04:42.080 our technology at ai speed that some people that have 0:04:42.200 --> 0:04:44.480 looked at fuel sALS ten years ago and have that 0:04:44.640 --> 0:04:47.960 notion of who we are don't understand where we are 0:04:48.000 --> 0:04:52.160 today as a company. We already are powering more than 0:04:52.240 --> 0:04:56.320 half a gigawat worth of data center because our total 0:04:56.360 --> 0:04:59.640 installed capacity for data centers is more than half a gigawat. 0:05:00.000 --> 0:05:02.159 So we're not a concept, we are not a PowerPoint. 0:05:02.360 --> 0:05:05.800 We're actually producing power for these data centers. And this 0:05:05.920 --> 0:05:10.599 year alone, we have increased our capacity to one gigaw 0:05:10.800 --> 0:05:14.000 a year, and we announced in our earnings call that 0:05:14.120 --> 0:05:16.720 by end of next year we can do two gigawards 0:05:16.760 --> 0:05:19.640 a year. That's two nuclear power plants worth of power 0:05:19.680 --> 0:05:24.320 coming out of our factory next year. And you know 0:05:24.360 --> 0:05:26.599 how long it takes to build a nuclear power plant 0:05:26.600 --> 0:05:30.479 and actually commission it. So we are no longer a concept. 0:05:31.120 --> 0:05:35.400 We are a mature technology that is affordable and can 0:05:35.480 --> 0:05:39.200 compete with other technologies that don't have as good attributes 0:05:39.240 --> 0:05:40.919 as we do. Street. 0:05:41.000 --> 0:05:42.640 I wonder if you could talk a bit about your 0:05:42.760 --> 0:05:45.120 business model. I know that you've announced deals with with 0:05:45.200 --> 0:05:48.520 other companies, most recently with Oracle. Of course, the company 0:05:48.520 --> 0:05:51.320 that I know is building investing a lot in its 0:05:51.400 --> 0:05:53.080 data center infrastructure. 0:05:54.360 --> 0:05:55.040 Could you. 0:05:56.480 --> 0:05:58.479 Carisould you just give us a sense here sort of 0:05:58.480 --> 0:05:59.840 of what those deals look like, what that deal with 0:06:00.480 --> 0:06:01.200 like in particular. 0:06:02.279 --> 0:06:07.920 Yeah, David, So our customers, here's what happens. Once we 0:06:07.960 --> 0:06:11.520 sign a deal with the customer, they either buy the 0:06:11.560 --> 0:06:16.000 capital equipment from us, and then they also sign up 0:06:16.000 --> 0:06:19.120 for a service contract. It's one hundred percent attached rate 0:06:19.200 --> 0:06:22.320 because we run those units for them and we maintain 0:06:22.440 --> 0:06:27.320 that and we offer them the promise guarantee of performance. Now, 0:06:28.040 --> 0:06:31.440 if they don't buy the equipment as cappex, we create 0:06:31.480 --> 0:06:35.000 a special purpose vehicle using financing and we simply sell 0:06:35.040 --> 0:06:39.240 them power just like the utility would, except this is 0:06:39.320 --> 0:06:43.400 on site and highly reliable and clean. So either way, 0:06:43.520 --> 0:06:46.560 when we create a special purpose vehicle, that financial investor 0:06:46.560 --> 0:06:50.600 in that vehicle gives us the upfront money for our capital, 0:06:51.040 --> 0:06:54.000 and we also have an annuity revenue coming out of service. 0:06:54.360 --> 0:06:56.800 So we have two streams of revenue if you want 0:06:56.839 --> 0:07:01.839 to think about it. Installed capital equipment that gets paid 0:07:01.880 --> 0:07:08.560 to us when we ship the equipment, a continuous service 0:07:08.640 --> 0:07:12.440 anuality over the tenure of the contract. That's how we 0:07:12.480 --> 0:07:16.040 make our money, irrespective of whether the customer actually buys 0:07:16.040 --> 0:07:18.760 the capital equipment or a third party buys the capital 0:07:18.760 --> 0:07:23.840 equipment and then charges the monthly rate from the customer. 0:07:23.960 --> 0:07:26.440 Heay, care way, I'm curious. I'm looking at one of 0:07:26.480 --> 0:07:29.040 my favorite functions on the BLOOMBERGTS and supply chain function, 0:07:29.080 --> 0:07:31.640 and I'm looking at your customers. Your biggest is I 0:07:31.640 --> 0:07:36.480 think American Electric Power Company, but I'm also looking at 0:07:36.520 --> 0:07:40.040 Alphabet Microsoft. You know, yep, sound familiar. These are some 0:07:40.080 --> 0:07:43.080 of the hyperscalers tell me how they are amping up 0:07:43.240 --> 0:07:45.840 their business, because they're the ones who are doing the massive, 0:07:46.000 --> 0:07:48.720 massive AI spend, But it looks like they're a very 0:07:48.760 --> 0:07:52.119 small part of your business. I'm just curious about what's 0:07:52.520 --> 0:07:55.520 the conversations you're having with them. Are they ramping up 0:07:55.520 --> 0:07:57.680 to do more with you? That would be very telling. 0:07:59.080 --> 0:08:03.440 Yes, the answer to your question is very simple. 0:08:03.640 --> 0:08:07.880 In the past, these hyperscalers would spend all their capital 0:08:07.920 --> 0:08:11.240 equipment just on building the data centers and they could 0:08:11.240 --> 0:08:13.800 simply pick up the phone and call the utility company 0:08:13.840 --> 0:08:17.880 and they would provide the power. No longer is that 0:08:18.000 --> 0:08:21.440 a reality because the utility does not have the amounts 0:08:21.440 --> 0:08:27.480 of power that these hyperscalers need. So now electricity power, 0:08:27.560 --> 0:08:31.640 which is what ultimately the data center needs to monetize 0:08:31.680 --> 0:08:36.120 the data center, is a supply chain that comes from 0:08:36.160 --> 0:08:38.800 the common good called the utility. So they have to 0:08:38.840 --> 0:08:41.079 own that supply chain or they are to procure from 0:08:41.120 --> 0:08:44.160 that supply chain. We are very much an integral part 0:08:44.200 --> 0:08:47.160 of that supply chain, and we are having conversations with. 0:08:47.160 --> 0:08:48.520 All the hyperscalers today. 0:08:49.040 --> 0:08:52.959 And you've seen the announcement aetorical expect to see similar 0:08:53.000 --> 0:08:58.120 things coming soon. Our American Electric Power contract was actually 0:08:58.200 --> 0:09:00.520 for AWS, which is another high for Scaler. 0:09:01.640 --> 0:09:04.440 I wonder if you could situate your company and this 0:09:04.520 --> 0:09:07.199 industry in the moment that we're in. So we're coming 0:09:07.240 --> 0:09:09.360 out of an administration that was doing a lot to 0:09:10.280 --> 0:09:13.800 support or encourage the adoption of alternative energy. It's way 0:09:13.840 --> 0:09:16.240 to say this administration that's not doing that, there's much 0:09:16.320 --> 0:09:20.240 less impulse to do that going forward. Here what do 0:09:20.320 --> 0:09:24.160 you see here as the ideal role of government in 0:09:24.280 --> 0:09:27.240 encouraging the adoption of technologies like yours? In other words, 0:09:27.280 --> 0:09:30.680 if you'd like to see fuel cell technology take off 0:09:30.720 --> 0:09:33.520 be more widely adopted, what is the role of government 0:09:33.720 --> 0:09:35.360 in making that happen if any. 0:09:36.720 --> 0:09:39.920 I think a simple role that they can play going 0:09:40.040 --> 0:09:44.680 forward is the US government, between the various government functions, 0:09:45.040 --> 0:09:49.000 the military, and the intelligence, is the largest procurer of 0:09:49.080 --> 0:09:54.880 electricity to be able to purchase electricity from technologies that 0:09:54.960 --> 0:09:58.280 are made in America. We are made in America technology. 0:09:58.600 --> 0:10:01.440 No other country in the world has the technology that 0:10:01.480 --> 0:10:04.840 we have. So they should embrace us, and they should 0:10:04.880 --> 0:10:08.000 give us an opportunity to serve the US government, the 0:10:08.080 --> 0:10:11.160 US military, and US intelligence and make us a really 0:10:11.200 --> 0:10:11.920 strong country. 0:10:13.120 --> 0:10:15.679 Well there, I was also curious about. And I'm looking 0:10:15.720 --> 0:10:19.560 at your market cap it's almost about nine point six 0:10:20.080 --> 0:10:23.920 billion dollars, and it is this excitement over AI. You know, 0:10:24.040 --> 0:10:27.400 there's all the money that's being spent care on the buildout, 0:10:27.520 --> 0:10:29.520 and I get that, but I'm just curious. And then 0:10:29.559 --> 0:10:31.280 there's going to be the maintenance and we might not 0:10:31.360 --> 0:10:33.600 need as many workers. And you know, we're all trying 0:10:33.600 --> 0:10:35.760 to get our head about around whether you know, the 0:10:35.760 --> 0:10:39.720 boom the bust part of maybe AI after the build out, 0:10:40.160 --> 0:10:42.720 what is the business for you? Is it maintenance? Is 0:10:42.720 --> 0:10:44.920 it replacement values? I'm just curious. 0:10:45.720 --> 0:10:50.800 Oh, thank you for asking that question, Carol. The buildout 0:10:51.160 --> 0:10:54.199 is just a tip of the iceberg. Why would these 0:10:54.400 --> 0:10:58.760 large hyperscalers spend close to a trillion dollars worth of 0:10:58.840 --> 0:11:03.280 capex just this year alone. If all that they're going 0:11:03.360 --> 0:11:06.840 to do is that build out, that's the training data centers. 0:11:07.200 --> 0:11:10.840 The inference data centers will be at least an order 0:11:10.880 --> 0:11:14.240 of magnitude larger than these training data centers, which are 0:11:14.280 --> 0:11:20.559 the foundation building blocks every factory, every bank, every law firm. 0:11:20.840 --> 0:11:23.480 They're going to have inference data centers close to the 0:11:23.640 --> 0:11:26.320 edge and these edge data centers are going to be 0:11:26.400 --> 0:11:30.280 located where you all live, where people are, where equipment is, 0:11:30.720 --> 0:11:36.080 and in those congested cities, getting the additional power needed 0:11:36.400 --> 0:11:40.360 ten to twenty megawats in every edge data center, which 0:11:40.400 --> 0:11:41.840 is an inference data center. 0:11:42.480 --> 0:11:44.679 We are ideally suited not just for the. 0:11:44.679 --> 0:11:47.800 Large data centers, but for these edge data centers because 0:11:47.800 --> 0:11:49.319 we are non polluting. 0:11:48.880 --> 0:11:50.840 We're quiet, we're reliable, and. 0:11:50.760 --> 0:11:54.200 We can provide it on site very quickly. So we're 0:11:54.280 --> 0:11:58.120 super excited about what we think is a really secular 0:11:59.480 --> 0:12:02.400 phenomenon and that's happening here that's going to last for 0:12:02.440 --> 0:12:04.959 a very long time to come. This is not just 0:12:05.000 --> 0:12:09.000 a build and bust cycle at all as far. 0:12:08.840 --> 0:12:09.360 As we see. 0:12:09.480 --> 0:12:11.080 So when you say a long time, are you saying 0:12:11.120 --> 0:12:12.200 decades or. 0:12:12.640 --> 0:12:14.400 I'm just going at least a couple of decades. At 0:12:14.480 --> 0:12:15.520 least a couple of decades. 0:12:16.120 --> 0:12:18.640 All right, Good to know. We'll stay in Touch'd love 0:12:18.720 --> 0:12:22.080 to hear because you certainly have a front row seat 0:12:22.120 --> 0:12:24.320 in terms of this spend and what is going on. 0:12:24.360 --> 0:12:27.160 Certainly all the power that's needed, it's just kind of astounding. 0:12:27.720 --> 0:12:30.600 Doctor Kerr Schweder. He is founder, chairman and chief executive 0:12:30.640 --> 0:12:33.240 officer of bloom Energy. The stock David up another nine 0:12:33.280 --> 0:12:35.559 percent in today's session, and it's really been on a 0:12:35.600 --> 0:12:38.800 tear this year, up about eighty four percent. 0:12:38.920 --> 0:12:41.120 So so interesting to talk about these companies because we 0:12:41.240 --> 0:12:43.679 focus so much on the AI providers, So as we're 0:12:43.720 --> 0:12:46.360 making these chatbots and all the developments there, but as 0:12:46.400 --> 0:12:49.360 we were just discussing the apparatus that you need to 0:12:49.400 --> 0:12:53.080 make all of that work so huge, so sprawling, and 0:12:53.120 --> 0:12:54.880 I'm just fascinating every time I kind of learn what 0:12:55.000 --> 0:12:57.520 goes into that making all of this stuff run right. 0:12:57.600 --> 0:12:59.600 Right, We just hear about that kind of the big hyperskeleton, 0:13:00.040 --> 0:13:02.000 and there's a lot of stuff that needs to be done. 0:13:05.240 --> 0:13:05.680 M hmm.