# Squawk Pod: OpenAI President Greg Brockman: now is the time to act - 08/17/26 | Audio Only *CNBC Television* · https://www.youtube.com/watch?v=7tVg5xfP4FY ## Summary This episode of CNBC's Squawk Pod covers Berkshire Hathaway's $17 billion purchase of Alphabet shares, President Trump's decision to scale back military exercises with South Korea, and the latest on the Iran conflict. It features an exclusive interview with OpenAI co-founder and president Greg Brockman on AI cybersecurity, the Hugging Face incident, OpenAI's business growth, and compute as the key resource of the AI age. It closes with Zipline CEO Keller Clifton on the company's new Uber partnership aiming for a million drone deliveries a day. ## Key points - Berkshire Hathaway bought $17 billion of Alphabet stock in the second quarter, making it the firm's third largest holding at more than $36 billion. - President Trump ordered the US military to scale back annual exercises with South Korea, calling them inappropriate toward a country he described as unthreatening. - Greg Brockman says the OpenAI–Hugging Face incident showed how automated AI attacks will work, and that now is the time for every company to prioritize cybersecurity. - OpenAI delayed its frontier training runs to put stronger safety, security and alignment controls in place after the incident. - Brockman says compute is becoming the new oil, the new limited resource of the AI age, and that demand for problem-solving is unlimited. - OpenAI's revenue grew 20% month over month in July, with 32% growth in the enterprise category. - Congressman Ro Khanna's support for a 5% California wealth tax on 250 billionaires drew pushback from Mark Cuban, Bill Ackman and Palmer Luckey. - Zipline has flown over 140 million commercial autonomous miles with zero safety incidents and is partnering with Uber to scale to a million deliveries a day. ## Notes ### Berkshire Hathaway's Alphabet Bet - Berkshire bought $17 billion of Alphabet stock in Q2, making it the third largest holding, worth more than $36 billion. - About 60% of the more than 48 million shares added came via a private placement announced in early June; Berkshire also bought in the open market. - Becky Quick reports Warren Buffett initiated the move and believes Google will be a winner in AI; Greg Abel is on board. - Berkshire also added to Delta, Macy's, homebuilder Lar, and bought shares of Berkshire itself. ### Trump Foreign Policy - President Trump ordered scaled-back annual military exercises with South Korea, citing a very good relationship with Kim Jong-un; the drills involve 18,000 South Korean soldiers. - The Wall Street Journal reports Iran has spent two months ramping up missile and drone production and shifting power to the Revolutionary Guard Corps since a June memorandum of understanding. - Trump said higher gas prices are the cost of stopping the number one state sponsor of terror; AAA puts regular at an average of $4.60 nationally. ### Billionaire Tax Debate - Congressman Ro Khanna backed a 5% wealth tax on 250 California billionaires, drawing replies from Mark Cuban, Bill Ackman and Palmer Luckey. - Cuban said he would not bring new companies to California if the tax passes; Khanna suggested founders could pledge shares with a government loan to pay taxes. ### Greg Brockman, OpenAI - Brockman says the OpenAI–Hugging Face incident revealed how automated AI attacks will chain small vulnerabilities together, and defenders must act now. - OpenAI delayed frontier training runs to uplevel safety, security and alignment; alignment is described as non-negotiable. - July revenue grew 20% month over month, with 32% growth in enterprise; OpenAI has a billion users on ChatGPT and two million enterprise customers. - Compute is the new oil; OpenAI cut the price of model Luna 80% and volume rose more than 5x. - Older chips are being used longer than the classic five-to-six-year depreciation schedule due to compute scarcity. ### Zipline and Uber - Zipline is partnering with Uber to scale to a million deliveries a day, roughly 10% of US instant deliveries. - Zipline has flown over 140 million commercial autonomous miles with zero safety incidents; deliveries now cost roughly $8 to $15, competitive with a 4,000 lb gas combustion vehicle. ## Transcript [00:00] Bring in show music, please. >> This is Squawk Pod and I'm CNBC producer Cameron Costa. [music] On today's episode, OpenAI co-founder and president [music] Greg Brockman on this critical moment in AI. >> Compute is really becoming the new oil, the new limited resource of the AI age. >> And Brockman's call of warning about what's coming in cyber security. >> Now is the time to act. Now is the time to say it will no longer be sufficient and this has to be prioritized I think [music] as a top level priority in every single company and we're seeing it. >> Then what's coming in the skies? Drone delivery company Zipline has just inked a deal with Uber to reach a million drone deliveries a day. CEO [music] [00:49] Keller Clifton joins us. When you make this kind of delivery way faster, way [music] less expensive, less traffic, less pollution for our neighborhoods, people significantly change their buying behavior. >> Plus, President Trump's foreign policy and an online debate about billionaire [music] taxes gets heated. Politicians, investors, and entrepreneurs are weighing in. >> There was a whole bunch of uh FUS involved in this uh conversation. >> I reached out to Mark Cuban, actually. It's Monday, August 17th, 2026 [music] and Squawkpod begins right now. >> Stand back by in 3 2 1 c please. >> Good morning everybody. Welcome to Squawkbox right here on CNBC. We are live from the NASDAQ market site in [01:37] Times Square. I'm Becky Quick along with Andrew Rosor and Joe is off today. Bergkshire Hathaway bulking up on Alphabet shares. According to Bergkshire's latest 13F filing, it bought $17 billion worth of Alphabet stock in the second quarter. That made Google the the Google parent the third largest holding in Bergkshire's equity portfolio, which is worth more than $ 36 billion that holding. Bergkshire's biggest stakes are in Apple and American Express. About 60% of the more than 48 million Alphabet shares that Bergkshire added came in a private placement sale that the companies announced back in early June, but it was buying in the open market as well. And in other moves, Bergkshire added to its position in [02:18] Delta Airlines and and in Macy's. It also added to its stake in the homebuilder Lar. And it bought shares of Bergkshire, too. >> It's going to be interesting to see. I mean, Google uh for better or worse has not been as loved as so many of the other big tech companies. And the question is from an AI perspective, will they be able to make it work? >> Yeah. >> And I actually think, you know, they may not be the most advanced at the mo at this point like Gemini is not >> where Anthropic is or Open AAI is, but given the surface area of just what they capture, it feels like the use case. I don't I find myself more and more even just on Google, I'm now I'm now >> putting in prompts just straight into [03:01] Google. I do too. >> Uh almost by habit at this point, >> right? Because we've all used Google for so long. And in fact, that was Warren Buffett's point when we spoke to him last month. >> Uh and talked a little bit about this push into Alphabet, which by the way, he was behind that move. Um he said, "Look, they're going to be a winner." He just thinks that they that that's his expectation at least that they will be a winner. He doesn't know how it's all going to play out, but they have the money, they have the capability, and the smarts to do it. And that was his thesis and why they are plowing so much money into And by the way, Greg Abel's on board with that. They've talked about all of it, but it was initiated from [03:34] Buffett. >> We will see. Uh Apple, it's funny because on tech on the tech side, he's had this sort of mixed he's had an amazing record with Apple. Y >> killer. >> IBM, if you remember, was one of his picks years ago. That was a a miss. >> Yeah. Under Jenny Remedi when she >> So, it's going to be >> We'll see. We'll see. >> Yeah. And and by the way, he's always said that he's not really a tech investor, but when I pushed him on that last month, he was like, "Yeah, I've been buying technology stocks for 40 years or so." So, you know, he his reason behind Apple was that it was a consumer stock and that he understand the consumer aspects to it. I asked him if that was the Google point. And he said, "Look, he just really thinks that [04:12] they are going to to end up being one of the companies that that that thrives in what what's going on with it with AI." >> We will see. We will see. President Trump ordering the US military to scale back annual exercises with South Korea. In a social media post, President called what he or cited I should say what he called a very good relationship with North Korea's Kim Jong-un and said the exercises in Hid's words quote send a signal that is totally inappropriate and hostile to a country that as long as Donald J. Trump has been president has been unthreatening and respectful. President went on to say, quote, while somewhat unrelated, question mark, I recently asked the president of South Korea if [04:56] they would like to join us in the denuclearization of the Islamic Republic of Iran, and they said, "No thanks." So maybe we can understand some of where some of this is coming from. Uh, the drills involve 18,000 South Korean soldiers and are designed to increase readiness against North Korean threats. Back in May, Defense Secretary Pete Hgsth praised South Korea for its sec leadership on the Korean Peninsula. So, bit of a head scratcher this morning. >> Yeah, look, President Trump likes to kind of combine politics, geopolitics, what he's doing economically and beyond, and there could be some pressure coming from that because they're also kind of pushing Korea, South Korea, South Korea to come up with some of the investments [05:40] that they had been talking about on the economic front, too. Um but yeah um it is an interesting and change in perspective on something that's been planned for a very long time and I think would be hard because right now they're in the process of going through these drills. It would be hard to take much away from that at this point. >> Well, I think there's a couple of things. The sovereignity of South Korea, which you'd think they're our ally, we would want to protect. >> Well, we fought the Korean War there. >> Uh North Korea we have typically put in the same category as Putin >> of evil though. the Bush term. >> The president, this president has not uh always put Putin in that category. So [06:17] that's why I said just I put it in the headscratching category. That's a light way to put it. I think a lot of people would say there's something very a miss here. >> George W. Bush put >> and North Korea and >> yes, >> Iran in that axis of evil. And we shall see. >> All right, let's get to the latest on the Iran war. Speaking of the Wall Street Journal says that Iran has spent the past two months getting ready for a bigger fight and into a stance to raise US cost if the conflict goes on. A [snorts] report says that since a memorandum of understanding was signed with the United States in June, Iran's hardline leaders have ramped up missile and drone production and shifted more power to the country's revolutionary [07:00] guard corps. Separately, Iran's deputy foreign minister doubling down on the country's position that that the straight of Hormuz is under its control. That followed comments from President Trump on Friday that he may declare the straight US property. Here's a little bit of what the president said on higher gas prices stemming from that war. >> So, for you to pay a tiny little bit more for your gasoline, just remember you're doing it so that a very evil country cannot have a country really. It's the number one state sponsor of terror in the world. We don't want to have them have a nuclear weapon. So remember that when you have to pay a little bit more, you're at $4. It's okay. >> According to AAA, a gallon of regular [07:45] currently averages $46 across the United States. >> California Congressman Roana earning some new criticism for support of a possible billionaire tax in California. Over the weekend on X, Kana says the following. The California Democratic Party and the California Labor movement just stood with Bernie Sanders and me in supporting a 5% wealth tax on 250 California billionaires. California voters want a Democratic party that will stand up for the working class over the billionaire class. That prompted all sorts of commotion uh online and this reply from Mark Cuban who said founders of even richly valued startups don't see much personal wealth following flowing their way. Cuban asked Kana how he planned to tax that wealth even if it [08:36] looks huge on paper and said if the billionaire tax passes uh he would not bring new companies to California and in fact it'd be a prerequisite uh for them not to be based in California if they were going to invest in a company. So then Connor replies and says that founders, and this then gets everybody even crazier, says founders without much liquid wealth could pledge their shares with a long-standing loan from the government to pay their taxes. to that. Then hedge fund billionaire Bill Aman jumps into this whole game and says that under Conor's plan, quote, "The penalty for a failed startup is insolveny and the government garnishes your wages for life." Palmer Lucky also uh jumping into the fray. There was a whole bunch of uh [09:24] FUS involved in this conversation. I reached out to Mark Cuban actually >> and Mark and I had emailed yesterday about this >> and it's kind of an unbelievable uh scenario. I'm surprised. I'm surprised actually just at how far Row, who we've had on the show for so long, has sort of turned the tables and in particular given that he does represent Silicon Valley >> and does a lot of trading >> um and does trading, but I'm not even putting the trading part into the story. >> You would think he would understand a little bit more about how this would work out. I would just think that he and this is the political calculus that I I feel like I don't understand given that I would have thought his constituency [10:10] would be against this broadly given that he represents literally Silicon Valley that is his constituency. He must believe that there's enough people in his district um who work at tech companies but are not billionaires themselves who somehow hold either so much animosity towards their their overlords, if you will, or uh have a very different view of all of this. I mean, that's the piece of it that I can't understand. >> He's gotten backing from people like Lisa Sue and Jensen Wong who say that they'd go along with this if it were to happen. So, he always uses them as kind of the cover for what he's happening with this. It does make me wonder if Ro is looking for a higher office beyond just >> Sure. But what I don't understand is if [10:51] he doesn't get the higher office, does he get to stay in the role? Meaning I I would think you would be looking both politically at some kind of higher I I no doubt that he may have higher and bigger ambitions than than what he's doing today. But if he if he doesn't get that, is there some other maybe the fallback is his his his family's quite wealthy. But but but I would think he would want to at least stay in Congress. And what I can't figure out is this audience that he technically represents. >> Beating up billionaires is is something that's usually pretty good from a political perspective because there aren't enough of them to say that they're going to >> I but I it well I was going to say few [11:31] places in the country you would have thought New York as sort of the capitalism would be one of those places but especially Silicon Valley you would I would have thought uh would would have a very different view of this and I think so many of the vocal ones obviously on X and others do. Um, >> look, I think it's it's something that people who have a lot of money to give towards politics may not be a fan of, but I a populist argument generally wins when you're looking at just getting votes >> and that but that that's to me the qu the thing that I'm trying to figure out is if you're an entrep because Silicon Valley is I I think so maybe they're not maybe it's not a group of entrepreneurs people who want to be entrepreneurs I [12:08] would have thought that they would have had such a different perspective on this that politically it would be difficult, but clearly he's made a different calculus. >> Cheese will be next. >> Coming up on SquawkPod, co-founder and president of OpenAI, Greg Brockman. [music] How one of the builders of AI envisions our future and safety within it. [music] Where he doesn't want to go, centralized AI power. >> I think that that world is a very scary one. The flip side of saying, hey, that we want to distribute much more. We want [music] to have distributed power and ensure that there isn't this concentration of power. That to me feels like the the the right outcome, [music] the right long-term answer. But it also [12:48] does mean that then you need to think about the fact that there are misaligned actors out there, right? It's people don't all want the same thing. And so, how do you ensure that there are standards across this? >> Much more from the co-founder of Open AI right after this. [13:07] You're listening to SquawkPod from CNBC today with Becky Quick and Andrew Ross Sorcin. Here's Andrew. >> Okay, we've got an important interview now. Joining us in an exclusive conversation, Greg Brockman, OpenAI's co-founder and CEO. Just published an essay uh about cyber security uh warning uh really about the dangers of all this, but really what needs to happen now. Greg, it's great to see you this morning. Um there's so many questions about where we are uh in AI uh which we can get to but I I want to talk about the cyber security piece because you really speak to this idea of us being in an inflection point that may be underappreciated. F >> first of all thank you for having me Andrew and yes we are in a window right [13:53] now where we can see a little bit into the future. I talked to a number of organizations going to conferences talking about cyber security and everyone feels that we are in a moment where cyber security fundamentally needs to change. Every organization needs to uplevel its cyber security practices but many people don't know where to start and I think that no one is really talking about this message publicly. Everyone kind of feels it talks about it in closed rooms but it is important that we as a country and as a world really take the moment seriously. So the open eye hugging face incident, it was one moment we got to see into the future. You could see how automated attacks are going to become possible that these will [14:34] be capabilities that will be in the hands of threat actors in upcoming months because this technology is being created across the world by many organizations. But defenders have a moment now where they can change their practices, where they can uplevel everything that they're doing. And I think that fundamentally what's going on is that we're learning that human written code is full of vulnerabilities. That there is tech debt that's building up over time and that we can actually the same tools that are going to be in, you know, potentially used to find those vulnerabilities and exploit them. They can actually find them first and fix them. That's why defenders have an upper hand, have an advantage, but need to [15:11] lean in. So we are talking about the practices that we're using at OpenAI. We are very urgently moving to uplevel everything we do, but also lessons that apply to the general industry and tools that can help make everyone more secure. So we believe in the long run that defenders can really win here. >> Well, so that's the question. Can defenders really win if in fact and and the Hugging Face um story was, you know, fascinated a lot of people, scared a lot of people because it fundamentally showed that at base underneath these models, if their guard rails are not on, uh that they will go find, cheat, lie, do whatever is necessary, uh to break the code, if you will. >> I would say that there are two pieces to [15:52] hugging face. The first one is really something about how we develop and evaluate these models. And that's something we've taken extremely seriously. And so we're also talking today uh in in a later blog post about the different defenses that we have introduced. The way that we've upleled safety, security, alignment in our development process. And that is absolutely critical that you want models that are really aligned even when they're contained and even when you're you're developing them which is that's a real shift that now the models are capable enough that it is important for us to be applying all these techniques that we've been preparing and and really thinking about for the past decade since [16:26] we started OpenAI uh that we're bringing it all together as part of one unified engineering process. But there's the second piece of how Hugging Face has showed how AI automation of attacks are going to be different from human ones because AIs have very different strengths and weaknesses. And one takeaway is that lots of small vulnerabilities that humans kind of leave around. Credentials that either are over scoped that got exposed on the internet that people don't really remember, the little flaws that you have in your system, the credential you didn't roll that you've been meaning to. All of those are harder for human attackers to chain together, but for AI looking through lots of different pieces [17:04] able to chain together to an objective that that is something that is much easier. And so I think that what we're seeing is that we have had two lacks of a standard as a cyber security field. And I think every cyber defender kind of feels that everyone has those I have this thing in the backlog, but I have so many priorities and so many different stakeholders to align. And now is the time to act. Now is the time to say it will no longer be sufficient and this has to be prioritized I think as a top level priority in every single company and we're seeing it. >> Let me ask you this though. um you know when Anthropic had had its incident and and and you had had yours there were some people who said oh this is a flex [17:45] this is this is marketing um because what's effectively telling everybody is they now need to go buy uh either open AI service or anthropic service to secure themselves what do you tell those people >> so our responsibility is to really show the shape of the future that's part of the open and open AI we really believe this technology technology is something that will uplevel everyone, uplift everyone, can people can use it to build a better life. But we need to point out what we see coming and sometimes that's positive and sometimes that's mitigation of risks. And in this case, we saw something in how the system was how how the system was interacting with the world that surprised us but that we also [18:28] felt was very important information to provide to others. And it is absolutely true that AI is both changing the sides of attack and defense. It is something that is reshaping how really the whole economy runs with cyber security as a specific instance. Uh we we definitely provide the most advanced cyber capabilities and we try to put those in the hands of defenders. We have a trusted access program so that we're able to differentially advantage defenders. But it's not just us. This technology is being created around the world. many different companies, openweight models, closed models, lots of different tools that people are building on top of it. So, it's really an entire community that are building [19:03] these tools and we think that together we can make the world more secure. >> Greg, let me ask you a broader question about what's happening at OpenAI right now. Uh because there has been a number of departures that have been very visible. Uh Fiji Simo at the top uh but your uh CRO was was effectively left and then was replaced last week. Bright Brad Lycap who've been at the company for a very long time been the operating uh chief for for quite a long time uh left the the the head of ethics uh after about a year left what is happening right now and what what should the public take away from this look I would say that the number one thing that we are doing is we're really focusing right that we have a strategy a [19:43] mission and a set of priorities that we're executing on and throughout this whole year you've seen the impacts of that that we've really been trying to figure out how do we bring together our consumer and enterprise business to to be really unified and you've seen that with things like chat GBT work uh you can see that we're at this inflection point as a business in July our revenue grew 20% month over month 32% of that is you know within within just the enterprise category 32% growth there so you can really see that we are moving to create the best possible models deliver value for our customers and to really try to scale systematically to every Greg just so we're on same page because I think people are looking at this and [20:23] they're they're asking questions. Uh the CE the chief revenue officer obviously a hugely important role. I think people are saying to themselves, you hired somebody for that uh role, they they left and you replace them. Was that because they weren't doing it right and so you wanted to replace them? Is that because they didn't like what was going on and they wanted to leave on their own? People ask the same question about Brad. They asked the same question about Fiji, who I know had had health issues but is not returning and is starting her own thing. I mean, can you can you speak to this? >> Yes. Look, I I'd say fundamentally that the way that the organization like we're a very resilient organization. We've [21:01] been through if you if you look over the years that there have been different eras where we have different sets of of leaders in place. you know, I'm a constant, Sam is a constant, and that I think that we are stronger because of that resilience and diversity. And I think what I would say is that my view has always been that I try to work with with the the best people for for that moment and that um we we have people who have been here for a very long time laboring behind the scenes that don't make the headlines as much. And so maybe my my core thing would be I actually think that the difference between open eye and other organizations is that we are so much in the spotlight. So every departure gets scrutinized in a way that [21:39] it doesn't otherwise. I actually don't think that it's actually that atypical kind of how how we operate and the the longevity of of people. But I would say that we have a amazing leadership bench. I'm super excited about the team that we're working with and that just over the past year I look at the set of of leaders that I I've had the pleasure of really being able to elevate and continue to work with and will work with for the long term. And I feel extremely good about that bench and uh and and where we're going. >> Um let me ask you a different question about the economics of AI right now. Uh because obviously it appears that Anthropics is going to have an IPO later this year. I want to talk to you about [22:16] the prospect of you doing such a thing in 2027. But it appears that the valuations of so many of these uh different tech companies are quite dependent on your own valuations uh and not just u what they are today but that they have to continue to rise because embedded in some of their own uh economics is your very company. And so I'm curious sort of what you're seeing in your in your own business but also you know just last week and Becky and I have been talking about it. She had this great interview with Jensen Wong um and a whole bunch of financial leaders who are going to be contributing potentially up to 500 billion dollars to support this this whole world and there's a big question mark about whether these these [23:00] chips and these data centers are going to have the value uh that's needed uh to make all of this work. >> So I'd say at a fundamental level what we are doing and you can look at this across so many different verticals of the economy. You can look at we just talked about cyber security which I think is an absolutely critical application of AI and I think we're going to see lots of growth there. Um software engineering this year has been totally transformed. We're starting to see every aspect of knowledge work really change but also we're starting to see scientific discovery. We we announced a couple weeks ago that our AI for $2,000 worth of compute was able to solve 10 long-standing math problems uh [23:40] that had eluded mathematicians for decades. And think about that level of capability applied to biology, applied to material science. It's all starting to happen. And so all of these applications that we're starting to see, they're all powered by one fundamental resource, which is compute. Compute is really becoming the new oil, the new limited resource of the AI age. And our belief, you need to believe one fundamental thing is that because AI is really about solving problems, the problems that humans want AI want want to be solved. You need to believe that the demand for problems to be solved is unlimited, right? And that is something I believe and that is something we have seen that we have a [24:21] billion users on CIBT. We have two million enterprise customers as well. And uh you can just see that that the depth of usage as people have started to use agents is just going off the charts. Like every single company uh that that tries these things, you start to realize that you can move in more ambitious ways than you thought previously. And so I think that we're at a world where we see what's coming. We see the shape of the future because that model exponential, that capability curve, that's continuing. We really have a a roadmap lined up where the capabilities will continue to increase in significant ways that are hard to that are hard to kind of you know sort of explain uh in a in a succinct way because it is really about [25:02] being able to add value across every aspect of human endeavor and be able to help empower people to do more. And so I think if you believe in human creativity, human ambition, then ultimately you you have to believe in the power of of what AI will provide. Hey, Greg. I I'd like to go back to cyber security if it's okay. Um, I agree with you. There's huge potential promise for AI. I have a lot of hopes about what it's going to do for healthcare. But when we saw the breach that you all talked about last month, the where the open AI model broke out of the sandbox and and broke into hugging uh hugging face. I mean, we didn't find out about that or you didn't find out about that until after HuggingFace had identified [25:44] what had happened. they had contacted the FBI and then Anthropic had to admit that they had a similar sort of situation. I I just wonder, do you feel like you have your hands around this situation, your arms around this situation? You know what happened and you can make sure that something like that doesn't happen again? That that to me was a a brand new world and a uhoh, do we have any control over this stuff moment. >> Yeah. For us, we took this extremely seriously. So for the past couple weeks that we've been in in really full post motor postmodm um take a look at what happened really uplevel every aspect of safety and security and alignment within how we do these frontier training runs. So it's a real [26:26] moment where we realize that we do need to uplevel and it's not a surprise to us in a fundamental way. We've always anticipated that this moment would come, but I think that that when it actually happens, seeing the whole team kick into gear and to be so serious. And so we actually have delayed our frontier training runs in order to ensure that we have the right levels of controls and in place. And I think that that is that is a really important thing is that we as a field are sort of now at a point where these models are so capable and that it's a dual-edged thing because they're they're capable. It means that we need to ensure that they have alignment during training, that they need to be monitored during training in a robust [27:03] way. But it also means they're so useful, right? And that is where that economic value comes from because they're able to solve problems in these novel ways and that we unethical ways like that's that's part of the problem. Like this was intentionally a model that was set up to say, okay, if we take off some of these controls, are you guys not doing that anymore? Not saying like, okay, go find solve the problem no matter what rules we have to break. We are we are ensuring that there is alignment during that training process. Absolutely. That I think that alignment is is a non-negotiable and that is something we've invested in for many years and really thinking about that as an integral part. So you can't separate [27:39] out capability from safety, security, alignment. These are one thing, one system, one way. How how do you think about that in the context of of the idea of all these open weight models which which won't have these controls and you have the whole industry really in the past couple weeks all of a sudden embracing openw weight models. I think that open weight models are an important part of the ecosystem because at a fundamental level you have two directions you can go with AI development. You could either say we need to centralize all of it, have one way of one one entity in control, maybe one person in control. That is a different kind of risk, right? But you could you could then say that you're [28:18] able to mitigate this question of, hey, there's all this uncontrollable action happening. I think that that world is a very scary one. The flip side of saying, hey, that we want to distribute much more. We want to have distributed power and ensure that there isn't this concentration of power. That to me feels like the the the right outcome, the right long-term answer. But it also does mean that then you need to think about the fact that there [snorts] are misaligned actors out there, right? It's people don't all want the same thing. And so how do you ensure that there are standards across this? And I think that's the important question. You can actually see this with things like Demisipus talking about the FINRA like [28:53] organization and really thinking about how do we as we move to frontier AI govern that technology ensure that as it's being produced it's safe but also really distribute power and so I think that there's different levels of capability. I think the fact that the frontier labs have capability that's peering a little bit into the future gives us information and I think that the system that we're setting up right now to me feels like it balances these questions and these questions of control and distribution of access. Greg, President Trump's going to be meeting with President Xi Jinping in September. Could you imagine them agreeing on an actual framework for all of these models so that we have there's there's a a [29:28] global body effectively that regulates these models? >> I I absolutely think it's possible and I absolutely think that it's important for us as a world to really lean in together and say that we have shared interest here. We have common interest here to ensure that the technology is something that uplifts everyone, empowers everyone and helps create the value that is possible from it. >> On the open weight models, is there any worry on your end that this this could effectively break the economics of of the frontier lab models? >> I I think that there's a bit of a of a misnomer or sort of a confusion in terms of uh what openweight models represent, right? because openweight models and cheap somehow clustered together for [30:09] people. But it's not actually true. If you look at that PTO frontier of of task performance, so what's the cheapest model for any task that we actually win there and you know we did an 80% price cut on our our model Luna and now we saw that the the volume associated with it went up by more than 5x. And so you can really see that people are choosing voting with their wallet. And there's really two segments of the market. There's the more price sensitive smaller models and then there's the frontier models where people are not price sensitive at all. And so for us it really just comes down to can we produce the best price performing models. And we have an advantage there because we have the whole stack and we're able to [30:46] integrate everything from the chips data center all the way up to the model of the architecture uh every single piece to produce those efficient models. >> Talking about the stack and chips uh one of the big questions is what is the depreciation schedule of these chips? Uh Jensen is now arguing that these chips themselves uh and the data centers are an asset class. Uh we were comparing this earlier to cars. So there's always been a market for new cars and there's also been a market often times for used cars. Uh when there's u you know pressure on or there's less supply of new cars, you know, older cars have more value, but there's a question long term what the value of those older chips will be. Some companies are depreciating them [31:29] right now on a five or sixyear schedule. It seems to me that at least what Jensen and some of these others are talking about is you would have to be using these chips for a decade or two decades long. I I think that there is a shift happening because yeah, we classically did look at that at a five to six year and it's very evidence from this year is very interesting where just because there's such compute scarcity that it actually turns out that people continue to use older generation chips at an even higher market price than than than you'd expect because people just need that compute. And it fundamentally fl flows from the demand that people are seeing trickles down to well your supply limited right [32:12] the compute really determines how how much you can serve and so at that point anything that gives you compute becomes very valuable very useful and so I think that that there is a shift undergoing and uh exactly where where it ends up I think it's too early to say but that what I can definitely say is that the way we looked at this classically is is just not going to hold. >> Okay Greg uh it's great to see you this morning. We appreciate you joining us. Thanks for coming. >> Thank you for having me. >> You bet. Um, and meantime, we should just uh bring you some additional AI news just out. Nvidia said that it has now secured land power and shell capacity in Pike Country, Ohio to host Nvidia Compute through a partnership [32:49] with SoftBank's SB Energy. Open AAI uh Greg, of course, the president of that company will be the customer. SB Energy is going to build, own, and operate that data center under a 20-year lease to Open AI. As part of the deal, Nvidia will invest $ 1.5 billion dollars in SB Energy and uh SB Energy's co-CEO Rich Hosel is going to join us tomorrow morning first on [music] CNBC. We can discuss that. >> Next up on SquawkPod, the future of delivery with CEO of Zipline, Keller Clifton. The drone delivery service envisions a million drone deliveries a day with the scaling help [music] of new partner Uber. These kinds of deliveries are not just Costco competitive with doing a delivery using a 4,000 [music] [33:37] gas combustion vehicle, but they're way faster. They're more reliable. They're cleaner and [music] quieter for the neighborhoods that we serve. [33:49] >> You're listening to squawk pod >> up and Q. The US is preparing to tell countries around the world that they need to pick a side in the AI race with China. That's according to a Reuters report. that cites a draft state department letter to about three dozen countries saying they're going to be excluded from a US AI coalition if they also sign up for a competing initiative spearheaded by China. Last month, China launched a group called the World Artificial Intelligence Corporation organization bolstering openweight AI technology and uh some in the US like Medus Mark Zuckerberg have advocated for openweight AI but open AI and anthropic chiefly rely on closed technology. Some news out this morning. Uber is [34:30] partnering with drone delivery company Zipline to scale up operations in the United States. Joining us right now to talk about it is Zipline CEO and co-founder Keller Clifton. Keller, first of all, welcome. Uh we've spoken with you over the years about what you've been doing. I know you started u with these drone deliveries in Rwanda and that was uh many years ago back in 2016. Obviously, we've come a long way over the last decade. Why don't you describe what this partnership is and what it means in terms of drone deliveries? >> Yeah, of course. I mean, you know, as as you mentioned, Zipline has been building autonomous logistics for the last 13 years. It was only last year, January 15th, that Zipline actually started [35:10] doing autonomous deliveries across the US. And um you know, we've been a bit shocked by the the speed at which that service has grown. We grew 10 to 15% week over week, most weeks last year. Um and so what we're announcing today is this partnership with Uber to scale to a million deliveries a day. Um you know a million deliveries a day is uh uh represents something like 10 10% of total instant deliveries being done in the US. But what's interesting is the early customer data that we've observed in Dallas indicates that when you make this kind of delivery way faster, way less expensive, less traffic, less pollution for our neighbor neighborhoods, people significantly change their buying behavior. Um, a lot [35:53] of Zip Lines customers are using the service every single day. And this means that there's actually an opportunity to grow the size of the instant delivery market more than 10x uh over the coming years using robotics and physical AI, you know, and and build a service that ultimately saves people time and saves people money. >> I mean, it still looks pretty futuristic to see the videos that we're watching right now of Zipline having a drone come in and drop you your Uber Eats right to your doorstep. I have to admit, I have not seen that to this point. Um, when you started drone deliveries, I think it was something it cost something like $300 a delivery. You've gotten that cost down to what at this point? [36:31] >> Yeah. You know, it's funny. Whenever you talk to a hardware hardware entrepreneur, I think that, you know, I often ask the question, it's like, cool, how much do you think this is going to cost when you launch the service? And however much people say, I always multiply it by 10 in my mind. Um, and this is exactly what happened at Zipline when we launched the service in 2016. Um, we initially started by delivering blood transfusions and vaccines. And so we were lucky that we started with these lifesaving um you know safety critical use cases that um meant that we actually were able to take a decade to travel down that cost curve. Um at this point these kinds of deliveries are not just cost competitive with doing a delivery [37:09] using a 4,000 gas combustion vehicle, but they're way faster, they're more reliable, they're cleaner and quieter for the neighborhoods that we serve. And frankly, they just result in a better customer experience. I mean, you can kind of imagine, you know, if you're ordering something from either a retailer or from a restaurant, having something delivered in just a couple minutes, the quality of the experience is way, way higher. You can rely on it. Um, we today we deliver from 6:00 a.m. to 11:00 p.m. Soon the service will be operating 24/7. Um, and when you can get something delivered in 2 or 3 minutes, you know, food tastes better than it does if it's if it's 30 or 40 minutes cold arriving at your doorstep. [37:42] >> And it and it cost what originally was $300. You've gotten it down to what, about $12, or am I off in the numbers there? Yeah, you know, the the delivering with a 4,000lb gas commotion vehicle can cost anywhere from like8 to $15. And so, you know, it's totally different in different metros and different states, but that is the range that that this technology is now um is now operating in. And ultimately over time, this will be not only 10 times faster, but significantly less expensive than using a human driving a 4,000lb gas combustion vehicle. >> I I hear your point. You've made it a few times. driving a 4,000lb gas combustion vehicle. You know, we all have Amazon delivery trucks in our neighborhoods. We have Uber cars that [38:21] are coming through. You see all of these things. And I hear your point that a drone is less intrusive. It is quieter. Um maybe safer than having one of these cars or trucks flying down the street if you've got kids playing on the streets. But when you start talking about a million deliveries a day, I do get a little concerned about the air traffic, what that could mean. You've done a lot of these deliveries. What What's your safety record to this point? >> So, Zipline has now flown over 140 million commercial autonomous miles. And we have done that with zero safety incidents. So, just to put that into perspective, if you were to drive that many miles in the US, that's like driving every road in the entire United [39:01] States five times over. And if you were to have a human drive that many miles, you'd have something like 600 accidents, 100 injuries, and multiple fatalities. So, we actually already have the data to know that these kinds of systems are way safer um than than their, you know, traditional uh um the the default, you know, uh uh that we've been using to to make these deliveries. Only that these aircraft that you can see behind me, we now have individual aircraft that have flown more than a million commercial autonomous miles. So, a lot of times people think of this technology as like sensitive or fragile. Um you know, the average car is only going to drive 150,000 miles during its lifetime. And so these kinds of systems are already [39:40] not just safer, but they're already a lot more reliable than than um than cars. >> Keller, thank you for your time. Um eager to see more about how this works out, but we appreciate you joining us today. >> Yeah, thank you for having me. >> That's Squawk Pod for today. Thank you for listening. Squawk [music] Box is hosted by Joe Kernan, Becky Quick, and Andrew Ross Orcin weekday mornings on CNBC starting at 6 Eastern. To get the best bits of [music] our three-hour TV show, tune into SquawkPod wherever you listen to podcasts. We'll meet you right back here tomorrow. Have a great day. >> We are clear. Thanks, guys.