Ramini Hasani, co-founder and Chief Executive Officer of Liquid AI, discusses groundbreaking advancements in data centers and artificial intelligence technology. This conversation is part of theCUBE's NYC Wired series, hosted by John Furrier, co-founder and co-CEO of SiliconANGLE Media. Liquid AI's pioneering work in neural networks and edge technologies highlights the evolving landscape of AI factories shaping future data centers.
In this episode, Ramini Hasani shares insights on Liquid AI's origins and mission, exploring their innovative liquid neural networks born from research at MIT's Computer Science and Artificial Intelligence Lab. Hasani and fellow co-founders, including notable figures from MIT, develop adaptable AI systems capable of delivering powerful intelligence in compact forms, aimed at revolutionizing edge devices and robotics. The conversation also touches on the work of theCUBE Research Analysts who delve into the company's contributions to AI advancement.
Key takeaways from the discussion include the importance of adaptability and efficiency in AI systems, as emphasized by Hasani. They highlight the capability of small foundation models called nanos, which can match the performance of larger models, offering low-latency and privacy-sensitive computing. The conversation also addresses the potential for future architectures where hybrid solutions will bridge on-device intelligence with cloud-based systems, in line with insights from industry experts and analysts.
Forgot Password
Almost there!
We just sent you a verification email. Please verify your account to gain access to
theCUBE + NYSE Wired: The AI Factory - Data Center of the Future. If you don’t think you received an email check your
spam folder.
Sign in to AI Factories - Data Centers of the Future.
In order to sign in, enter the email address you used to registered for the event. Once completed, you will receive an email with a verification link. Open the link to automatically sign into the site.
Register for AI Factories - Data Centers of the Future
Please fill out the information below. You will receive an email with a verification link confirming your registration. Click the link to automatically sign into the site.
You’re almost there!
We just sent you a verification email. Please click the verification button in the email. Once your email address is verified, you will have full access to all event content for AI Factories - Data Centers of the Future.
I want my badge and interests to be visible to all attendees.
Checking this box will display your presense on the attendees list, view your profile and allow other attendees to contact you via 1-1 chat. Read the Privacy Policy. At any time, you can choose to disable this preference.
Select your Interests!
add
Upload your photo
Uploading..
OR
Connect via Twitter
Connect via Linkedin
EDIT PASSWORD
Share
Forgot Password
Almost there!
We just sent you a verification email. Please verify your account to gain access to
theCUBE + NYSE Wired: The AI Factory - Data Center of the Future. If you don’t think you received an email check your
spam folder.
Sign in to AI Factories - Data Centers of the Future.
In order to sign in, enter the email address you used to registered for the event. Once completed, you will receive an email with a verification link. Open the link to automatically sign into the site.
Sign in to gain access to theCUBE + NYSE Wired: The AI Factory - Data Center of the Future
Please sign in with LinkedIn to continue to theCUBE + NYSE Wired: The AI Factory - Data Center of the Future. Signing in with LinkedIn ensures a professional environment.
Are you sure you want to remove access rights for this user?
Details
Manage Access
email address
Community Invitation
Andrew Sobko, Argentum AI
Ramini Hasani, co-founder and Chief Executive Officer of Liquid AI, discusses groundbreaking advancements in data centers and artificial intelligence technology. This conversation is part of theCUBE's NYC Wired series, hosted by John Furrier, co-founder and co-CEO of SiliconANGLE Media. Liquid AI's pioneering work in neural networks and edge technologies highlights the evolving landscape of AI factories shaping future data centers.
In this episode, Ramini Hasani shares insights on Liquid AI's origins and mission, exploring their innovative liquid neural networks born from research at MIT's Computer Science and Artificial Intelligence Lab. Hasani and fellow co-founders, including notable figures from MIT, develop adaptable AI systems capable of delivering powerful intelligence in compact forms, aimed at revolutionizing edge devices and robotics. The conversation also touches on the work of theCUBE Research Analysts who delve into the company's contributions to AI advancement.
Key takeaways from the discussion include the importance of adaptability and efficiency in AI systems, as emphasized by Hasani. They highlight the capability of small foundation models called nanos, which can match the performance of larger models, offering low-latency and privacy-sensitive computing. The conversation also addresses the potential for future architectures where hybrid solutions will bridge on-device intelligence with cloud-based systems, in line with insights from industry experts and analysts.
play_circle_outlineDemocratizing Access to GPU Resources: How Argentum AI is Transforming Compute for Developers and Companies
replyShare Clip
play_circle_outlineEmpowering GPU Entrepreneurs: Building a Marketplace Through Strategic Partnerships with Data Centers and Operators for Global Capacity Growth
replyShare Clip
play_circle_outlineTransforming Access: How Argentum AI's Marketplace Ensures Transparent Pricing and Standardized Compute Solutions
replyShare Clip
play_circle_outlineDemand for AI compute resources is high among enterprises and researchers.
In this AI Factories – Data Centers of the Future segment from theCUBE’s NYSE studio, Andrew Sobko, chief executive officer and founder of Argentum AI, joins theCUBE’s John Furrier to unpack how Argentum is building an asset-light marketplace that aggregates GPU and compute capacity across operators – from Bitcoin miners to institutional data centers – into a single, enterprise-ready service layer. Sobko shares why “GPU entrepreneurs,” second-life and third-life GPUs (the “four-year-old Lamborghini” analogy) and standardized, transparent pricing can expand af...Read more
exploreKeep Exploring
What is the approach being taken to democratize access to compute resources in the marketplace?add
What strategies can be employed to support the development of a global network for GPU entrepreneurs?add
What is the significance of creating a service layer with simple pricing in relation to the marketplace model and the emergence of GPU entrepreneurship in today's economy?add
What trends and challenges are enterprises currently facing in accessing compute resources?add
What focus does the company prioritize in its partnerships, and how does it address issues with underutilized resources from professional services companies?add
>> Hello, I'm John Furrier, host of theCUBE. We are here at theCUBE's New York Stock Exchange studio for our AI Factory series, where we feature the leaders who are making it happen in AI, as compute, XPUs and the entire large-scale infrastructure is completely building out to support the future of AI, generative AI, agentic, and soon, physical AI happening. The leaders of NVIDIA and many more, Dell Technologies, are continuing to invest in what is a computing paradigm shift. And we talk about all the leaders, we've got a great one here, Andrew Sobko, who's the CEO and founder of a company called Argentum AI, the innovative opportunity to democratize compute and scale it up. Andrew, thanks for coming on theCUBE. You're in the building. Peter Tuchman brought you in. Hat tip to Peter.>> Thank you for having me.>> Thanks for coming on.>> Thanks for having me.>> So, I put your deal together in a nanosecond because we've been covering it. You're doing something very, very innovative. You've done it before, another venture you started. You're essentially democratizing compute by aggregating GPU cycles and compute cycles, and offering it in the marketplace. So, if I'm just a developer or a company, I can just tap into the resource. This is not new for you. So, explain what you guys are doing and how you got here.>> Yeah, exactly. So, number one, I'm personally big believer in marketplace businesses and leveraging asset-heavy industries and building great technology on top of it that is enterprise-level ready. So, my playbook historically was always how can you consolidate tons of capacity globally? Let's talk about GPU entrepreneurs or think about Bitcoin miners globally. How can you convert all those Bitcoin miners and also leverage all of the new data centers that are getting built out? Put this all together, package us into one space, one spot where people believe and trust and have full confidence and can go and get access to compute seamlessly, accessibly, affordably, and number one, it's available today? So, that was always like my playbook and we've been deeply thinking around how can we incentivize more, I call it, GPU entrepreneurs. Trying to create this new economy of GPU entrepreneurs. What can we do? How can we help them to build this huge network globally? And the best way to think about it, we can partner up with hundreds, and we're already partnered up with tons of institutional-based data centers or I call it also GPU operators, and we'll keep partnering up with them globally. But then instead of us, again building hundreds of data centers globally, we can partner up with all of them, have one standard, one benchmark for pricing, one website where they can go get access to compute similarly to what Uber app is offering based on stability, latency, what kind of GPUs you want and get access to it.>> So, you basically create a service layer that's simple, easy, not cryptic pricing. You actually have software that makes that happen, so that people can just get the compute. I love the idea and, first of all, I love marketplaces, Amazon Marketplace... The cloud has shown the marketplace playbook, no doubt about it. And you've done this before in your other business in trucking. And the CapEx spend is also a tailwind for you because there's a race for those GPUs. So, I love that. But I love this also concept that you mentioned GPU entrepreneur. I think that's important. I want to do a drill-down on that because we've been talking about here on theCUBE, on AI Factories and other shows that there's an AI-native wave. It's not just entrepreneurs, it's entrepreneurial thinking. So, if I work for a big bank, I'm technically not starting a company, but I'm starting a new platform, system architecture in, say, the bank. The old way was IT. That's antiquated, maybe abstract over that, but I want to take advantage of that and I'm an entrepreneur. I don't have CapEx money, I don't have billions to compete with the hyperscalers or the neoclouds, so what am I going to do? So, this seems like a good fit. Did I get that right? Expand on that because I think this might open up a massive developer surge both on premises and native entrepreneurs building native apps.>> So, think about it on one side, when we started, when we kicked this off, we want to make sure that we have very stable service. We're focusing on stability and latency. Think about us partnered up with tens of CoreWeaves, a little bit smaller CoreWeaves of the world.>> Maybe it's CoreWeave.>> Maybe it's CoreWeaves. They might maybe convince NVIDIA to get some latest GPUs, might have GPUs coming available there soon, but they were not able to secure a customer quick enough. In many cases, we're also seeing, again, monetization of some idle GPUs. We'll talk about it later on. But my big picture and global scale, looking at this global scale, how can you build this marketplace and this AI infrastructure of this and you have this economy of these GPU entrepreneurs works at scale globally? And we deeply been thinking around the incentive. It's around incentives. How can you incentivize people? Remember when Bitcoin mining started back, I don't know, eight, 10 years ago, it was incredibly hot space. A lot of people put tons of money into it and you call it professional retail. Some people I saw had 10 GPUs in their house and they were doing Bitcoin mining, and some people built a tiny mini data center, maybe in their garage or maybe it's more institutional lakes. So, our job was how do we convince all of those Bitcoin miners that understand the space pretty well to turn into AI mining, or I call it compute mining. And instead of just paying them whatever hourly price GPU per power, we're thinking what else can we do for them? How can we incentivize them more maybe to get access to cheaper GPUs, right? We'll talk about it the second-life GPUs and the third-life GPUs, which is a major separate problem on its own. How can we give them access to those GPUs that ChatGPT was created on? I like to say it's like a four-year-old Lamborghini. It still goes fast. It's still pretty good car, but it's not the newest Lamborghini. So, how can we give->> It goes from point A to point B.>> Exactly. How can we give them access to those GPUs? And then, immediately in those particular countries and like to say emerging markets, emerging countries, they will have access to compute at scale, and that's what we're doing.>> Andrew, I really like what you're doing because one, I'm watching the mainstream narrative on AI bubble, and their entire thesis of the naysayers is, "Well, they can spend all this money, but unlike the fiber..." They always go to the fiber example, which I think is a shitty example. They laid fiber down their internet, but the experience wasn't in demand. So, yeah, dark fiber maintained and then it came back. I think it's a flawed analogy to compare the build-out of GPUs, the fiber, because the reuse which you're talking about shows that no matter what the CapEx is the architecture and use cases of reuse becomes a factor. Because you mentioned latency, you can actually take the old gear, GPUs, and configure them in a way that a DeepSeek-like software layer could take advantage of it. So, that kills the idea that the bubble will burst just on latest GPUs. Now, Jensen at NVIDIA wants to sell more GPUs, but that'll go to the premium market. So, you've got a premium, then you've got the tiered markets. What's your reaction to that? Do you agree and what's your thoughts on that? Because I think what you're doing points to the fact that no one really cares. How much am I getting? Does it serve my purpose? What's the price?>> Yeah. So, we've seen people care a lot about accessibility, of course. Number two is affordability and then security layer. Security is probably one of the most important things, I think when addressing all of it. About AI bubble, I cannot agree with that statement as we are seeing since we launched Argentum AI->> You agree with it?>> I do not.>> Okay.>> We received thousands and thousands of applications->> You agree with my statement or the bubble will burst?>> I don't agree The bubble will burst. I don't think there's a bubble yet. We've seen ridiculous amount of demand coming in right now from very sophisticated also enterprises globally. They're not maybe Googles of the world or OpenAIs of the world, but they're still businesses at scale, have strong balance sheets and they just cannot get access to compute. And in many cases. We've seen also a lot of law firms. We saw what happened with, unfortunately, with AWS, and we started seeing a lot of law firms trying to think about, "Okay, what else can I do? How can I maybe build my own data center? Maybe I can find somebody to partner up with, maybe somebody else who has," I call it, decentralized type of place, which is historically always more secure. So, I don't agree with there's AI bubble. Of course NVIDIA is absolutely incentivized to continue selling new GPUs and that's where the issue comes in with second-life GPUs. Again, we can talk about electricity grid issues. We're predicting that there will be hundreds of thousands of GPUs that needs to get rotated. I call it second-life GPUs, third-life GPUs.>> I like the second-life GPU angle.>> And again, even back to the latency issue, some companies we're getting from Eastern Europe, or India, or Middle East, they would love to get access to H100s today and they just cannot. And if we can get those second-life GPUs and, again, at scale, distribute them, partner up with a lot of really good operators globally, everyone committing back to our marketplace and people can just log into those marketplace and they can make their own decision based on... Maybe they care about stability, maybe they don't. Maybe they care about latency, maybe they don't. But our job is to give them tons of capacity, affordably, accessibly, at scale.>> It's like electricity. It's like electricity. So, let me ask you this because one of the things that's happening, obviously, we see the CapEx rush, the hyperscalers and neoclouds, they're buying whatever they can. As soon as it comes off the line, they're buying it. But you mentioned getting a customer. They have the big customers, I mean most of the neoclouds have six customers.>> Yeah.>> You're talking about a whole nother level of customer base. So, your customer profile isn't those big guys per se, it's who?>> So, again, our focus on that mid-level enterprises. Even though we started focusing just on that, we received tons of demand from even hyperscalers that will basically would like to get any capacity they can get their hands on because they have so much demand coming in because they're such incredible brands, they would like to partner up with us. But our core focus, again, mid-level enterprises. But again, back to->> Your platform's agnostic?>> Exactly. But now back to the neoclouds. The neoclouds, in many cases again, they might be successful convincing some professional services company to commit to a three-year contract on B100 GPUs. In some cases, we're seeing that those professional services companies using those GPUs only six months of the year. So, you're dealing with a problem where very credible large company with strong balance sheet committed to a three-year contract and they're using that only six months of the year. What are you going to do with it? So, our job is, I like to call it bringing in liquidity into idle GPUs or retired GPUs on both sides. If we can help them monetize on those idle GPUs, even though it's committed to Microsoft or one of the hyperscalers and it sits in the Microsoft Azure system, we can help them monetize on it. Everyone wins.>> It's like bandwidth. Remember the old over-provision? We always have connectivity. "Oh, you got a one gig, you got 100 meg connection." Well, I'm sharing that with everybody else. Now, that's over-provisioning. But what you're saying is that the CapEx spend, the usage is not there. You're taking advantage of that with your software. So, that's like a classic, "Hey, it's idle or whatever your idle time is, I'll take those." And by the way, the CapEx is forcing the neoclouds to actually have excess capacity that they've already bought. So, for them, why wouldn't I sign a contract with you?>> Exactly.>> Because I'm going to pay you because you're bringing me business. It's kind of like AWS. The more the EC2 runs, the more the bill goes up.>> Exactly. But again, we're seeing right is AWS did a phenomenal job owning, I would like to call it the cloud economy. They have like 30%, 40% market share. But we do see that people are getting very, again, thoughtful around allocating and AWSs of the world controlling 40% of AI infrastructure economy. That's a pretty big risk for them. So, we've seen a lot of shift going towards some sort of decentralized solution and that's where we come->> All right. So, I'm sold on the market opportunity. It's great opportunity, I love it. Business model, how does it work? SLAs? Take me through some of the business issues that you're dealing with. What's the consumption level look like? Is it easy to execute?>> So, at the beginning, as I mentioned on the supply side, we partner up with the most stable operators. Think about neoclouds, but not at scale yet. In our average deal size, let others focus on those $1 billion transactions or $200 million transactions or $20 billion transactions. We'll sit somewhere in between. Let's say we'll do $10 million, $20, $50 million transactions. That's number one. Our business model is always again helping them monetize on their sitting idle GPUs, create like a spot market. And then, I think the most innovative thing that we're building, we build this called living AI benchmark. So, think about whenever AI workload gets submitted, typically people are struggling understanding what kind of GPUs they need, for how long they need. Of course, people at scale. Again, I don't want to keep mentioning hyperscalers, we'll convince them to sign very long contract for as many GPUs as possible. But our job is how can we be more efficient, bring efficiency. And within a second, whenever workload gets submitted, how can we price it and match it with the right GPUs for the right period of time and give customer that right to make their own decision around this?>> And so, your secret sauce is that matching engine with the workload?>> And the price upfront.>> And price?>> The price upfront within a second. Think about your Uber app. Using your Uber app where exactly where you're going->> On a black car->> The pricing, you have four different options. We're doing the same. You have H100 option, you have H200 option, you have option with excellent latency->> And I can see all the performance you're going to guarantee.>> Exactly. And we're going to guarantee it.>> All right. So, take me through where you guys are at. The company formation, stage. Where are you? Obviously, you're up and running. What's the progress? Give us some stats.>> So, we've been doing a lot of R&D at the beginning, trying to come up what can we deliver that is super unique and enterprises are comfortable using? So, at the beginning I've been always very, very focused on how can I convince these enterprises? So, my focus was security. Security was a big issue. So, we picked this new security layer called zero-knowledge and staking-based trust. Eric Smith from Google talks a lot about, Ben Horowitz from Andreessen. There were a of conversations. So, we picked the zero-knowledge security layer on top it, which is let's say on top of the current data center infrastructure, we're adding additional security layer, which is phenomenal. Then, we were forming team and I've interviewed probably hundreds and hundreds of people and trying to understand who really understands from zero-knowledge security layer, who has experience building this? So, we hired approximately, I would say in April-May, we hired about 10 engineers full-time. Company was not even formed yet. Just to dive in, figure this out to help us, again, with how the marketplace will operate. Officially, we launched the company in August. We kicked off our pre-seed round. We were oversubscribed in less than 24 hours. That was incredible. That was very interesting to me.>> Phone was blowing up.>> Yeah.>> Who'd you pick?>> We picked one of our core investors, was a company called Banyan Ventures, Kraken exchange, which is a very successful exchange.>> Yeah, I know those guys. Yeah.>> And one of our partners became Victor Morgenstern. Victor was original, was former chairman of Valor Equity Partners. And Victor, one of their original investors, again, in and all of Elon-associated companies, so SpaceX or Tesla and others. Victor, great partner, Kraken exchange->> So, you have some strong people behind it?>> Exactly, yeah.>> All right. So, what round are you in now? Are you raising money now?>> So, we are kicking off the next round. It's a $20 to $25 million round. We let market determined us. We're also in the middle of very large transactions with some very large companies right now. So, on the one side we already secured... I like to say we went around the globe secured over 10,000 latest NVIDIA GPUs and additional 10,000 of older version GPUs to the marketplace. Fully committed an additional 150 megawatts plus of size of capacity committed to us under full exclusivity. We're in the middle of signing a very, very large contract with very large well-known company in the hyperscaler space. And again, after that we predict, we'll continue->> Yah, it's interesting, there's an old expression in Silicon Valley, "Build it, they will come." And it's been a pejorative term, meaning, hey, any venture that has, "Build it, they will come," doesn't have any customers. But that doesn't work when you actually have a demand curve. So, here you built it first, knowing they will come because you knew the market demand was there. At least in the segment you knew you could knock down, but now you've got the bigger guys coming in. So, it's the term sheet's going to be oversubscribed again, I'm guaranteeing that. So, congratulations.>> Thank you.>> What's next for the customer? Because the impact here is significant. I mean just riffing out loud here, you're seeing a developer community with all the open source, even NVIDIA is going open. Look at Hugging Face leaderboards massive. The Linux Foundation is getting more heavily involved. So, all these developers are literally one click away from getting resource. So, if they have a data moat, you don't have to have a CapEx investment. I mean that's, to me, the real key. What's your vision on what will happen next? Because my thing is I would see this enabling huge access to compute for people who are innovating, who might not have big bank yet.>> Exactly. So, we already have a lot of partnerships that we've been a announcing recently from entire quant community researchers, universities. We just announced a big partnership with one of the major universities. We just partnered up with one scientific quant research lab in Europe. We just announced a partnership, I think it was yesterday. So, that side, we're getting tremendous demand, again, because of our security layer because of us being fully decentralized and global player, that's how creating this. And giving people accessible and affordable compute, of course that makes sense. People don't really want to go to, again, to some of the largest companies. Get on the wait list, number one. Sit on the wait list for potentially six months plus to get access to compute. And then, of course, overpay. We build this, we already offering, on average, 40% cheaper compute .and even some cases for some of the latest newest GPUs that are not currently available with some of the hyperscalers.>> So, on the risk, if I put my risk management hat on, I'd say, "Okay. Andrew, so if you're going to pre-buy all this compute, you better have buyers." Is that risky for you or is there an out clause or how do you look at... Because you're committing. So, do you pre-buy or you make a soft commitment?>> So, our big business model, and again, through my personal background, I'm a huge believer in asset-light marketplaces. Think about->> Yeah, it's a win>> Basically, you assemble it, you add security layer and you, again, help people, again, I call it bring liquidity to those illiquid assets. So, our commitments, mostly we're going around and there's some new data centers are getting, build some new GPU operators at scale that were already operating and they are maybe getting additional allocation they did not expect from NVIDIA. So, they need to quickly source a client for it. On our side, because we have such a strong demand on the customer base, that's where we come in.>> You're arbitraging payment?>> We are transparent. It's on our website. We're taking, on average, somewhere around 10% fee that our marketplace fee in between. And in some cases for very strong contracts, we take that fee down significantly.>> I had a conversation with Bill Tai about this because we were talking about energy options with GPUs, kind of similar. You're smiling. I can see where your head's thinking right now. Because if you think about it, the price is probably going to go up. So, in a way, it is a call. So, that's also a risk management upside for you that your bet is there's still going to be demand.>> Exactly.>> So, your only real focus on the risk side is to make sure that you're making the right calls on commitments, right?>> Yeah, we need to make sure that at the beginning of year one, we have tons of stable capacity available now. So, if people want to come into us right now, they can get access to that. Look, we're in the middle already of some of the most important AI conversations today on the AI infrastructure side, where we have a very well-known company came to us and they wanted to source start with 2,000 GPUs and we were able to source, through our marketplace, again, over 2,000 latest NVIDIA GPUs from B200s to B300s, which are available today on our marketplace. And again, go back to some larger companies, you will get a huge wait list. You'll have to wait. And we're being, again, that flexible partner, then giving you flexible compute.>> Well, I think one value that jumps out at me is the time. I mean if you had to source this yourself, imagine I'm a tech guy, I'm like, "Hey, I need some GPUs." Procuring them or even if I wanted to rent them, I'd have to go out and figure out each one... You're right there, simplify the process, so it's faster time to value, instantly.>> Exactly. And we'll let experts to deal with... Again, we're not experts on building data centers, operating data centers. We'll let other experts do the job for us and we will focus on->> Well, your asset-light. They're the experts. You're providing two main benefits, availability and affordability.>> Exactly.>> All right. So, put a plug in for what you're working on. If people are watching, if you're an investor, tell them John sent you, you get a 20% discount on the next round. He's like, "No, I don't where you got that." Only kidding. What are you looking to do? Put a plug in. You're hiring. You got deals. Obviously, there's demand. What are you optimizing for? What's the focus?>> So, the focus right now, continue building up exceptional A team. We will be making an announcement in the next, I would say probably a week or two. We'll bring in very well-known executive from the well-known cloud space or let's call it neocloud space. So, it'll be major announcement. And again, the continued focusing on building great technology, continue building our benchmark and continue, again, focusing on consolidating demand and supply globally. So, global expansion, it's on our map today. We already partnered up with some partners in UK, Brazil, Germany, and we're trying to expand, again, globally as fast as possible.>> Where's the office here? Is it here in New York? Where's the location, headquarters?>> So, our core office is based in Chicago and we have our second office is Los Angeles.>> And any international yet? No international?>> We have some developers that are globally->> But they're working?>> They're working for the company, but we don't have office->> You don't have office?>> I think our next spot will be UK. We receive tons of demand from the FinTech side of clients in Northern London, for a million different reasons. So->> But no international operations, other than employees, remote workers?>> Exactly.>> Yeah, that's going to be more expansion. Well, Andrew, thanks for coming on theCUBE. Again, AI factories is our focus. We've been covering the AI infrastructure leaders, the trailblazers. And again, the demand is high. And I think the developer angle is huge because whether you're in an enterprise or starting a business, I remember when I used Amazon for the first time in 2008, no one even never heard of it. I didn't want to buy a server, I just put my credit card down. Similar vibe here.>> Exactly.>> And you're just doing all that aggregation.>> And number two, we like thinking deeply, "How can we make those cross-border transactions seamless?" So, instead of waiting for people to, again, send in the wire transfers, we're trying to make sure how can we have full verification of job was completed, completely verified within a second?>> All right. Well, good luck on the next round. I'm sure it's going to be oversubscribed again. Again, this is a part of the innovation we're seeing. Second-life GPUs, the GPU entrepreneur, AI-native software. Whole new generation coming in. Thanks for coming in. Appreciate your time.>> Thanks for having me.>> Thanks for coming on theCUBE. All right. I'm John Furrier, host of theCUBE here at our NYSE Studio, where we're breaking down the leaders in AI and the innovators are making it happen as Wall Street and Silicon Valley connect, tech and money. That's what's happening and it's happening all sectors. See cryptocurrency, that's bounded by energy. And of course AI, bounded by energy. Two hot areas where the infrastructure is under rebuild. It's disruptively enabling new things and we're seeing it every day. It's another great use case. We're doing our part to bring that to you. Thanks for watching.
