theCUBE + NYSE Wired: Physical AI & Robotics Leaders | Bill Tai & Sanjay Gupta, Auradine

Clips
More from theCUBE + NYSE Wired: Physical AI & Robotics Leaders

Bill Tai

Chairman

ACTAI Global

Sanjay Gupta

Chief Strategy Officer

Auradine

play_circle_outline Overview of the robotics and AI infrastructure event with key industry experts.

play_circle_outline Bill Tai discusses growth in Bitcoin mining infrastructure post-halving and innovations.

play_circle_outline The convergence of AI and Bitcoin mining leading to new business opportunities.

play_circle_outline Importance of energy supply and load balancing for AI and Bitcoin mining sectors.

play_circle_outline Driving Innovation: Sanjay Gupta Highlights AI Infrastructure and the Need for Energy-Efficient Semiconductor Technologies

play_circle_outline Emphasis on decentralized infrastructure and its role in global economic stability.

play_circle_outline Empowering Innovation: Entrepreneurs Navigating Energy, Networking, and Cooling Technologies Amidst Geopolitical Infrastructure Challenges

play_circle_outline Transformation of Hut 8 from Bitcoin mining to energy infrastructure company.

Info
Transcript

Bill Tai & Sanjay Gupta, Auradine

Bill Tai

Chairman ACTAI Global

Sanjay Gupta

Chief Strategy Officer Auradine

Bill Tai, venture capitalist and board chair at Hut8, and Sanjay Gupta, chief strategy officer at Auradine Inc., join theCUBE’s John Furrier during theCUBE + NYSE Wired: Robotics & AI Infrastructure Leaders 2025 event to discuss the infrastructure realities behind robotics and AI’s explosive growth. Their conversation unpacks the critical intersection of decentralization, energy innovation and data center evolution.

Gupta outlines shifts in Bitcoin mining and semiconductor design as key drivers of infrastructure demand, especially as robotics and AI ... Read more

explore Keep Exploring

What is the context of the event being covered in the text? add

What are the current macro trends in Bitcoin mining and AI infrastructure in data centers? add

What is the current state of demand for AI and robotics infrastructure, particularly regarding the energy needs for data centers? add

What is the expected energy demand for AI data centers over the next five years, and how does it compare to energy consumption in American homes? add

What are the key challenges and opportunities related to energy supply and demand, particularly in the context of AI and Bitcoin mining? add

What is the relationship between decentralization, the internet's origins, and the development of digital currencies? add

What changes are anticipated in the geopolitical landscape and how do technologies like Bitcoin and AI relate to economic value? add

What changes is Hut 8 making in its business structure and focus? add

bolt Powered by CUBE AI

Bill Tai & Sanjay Gupta, Auradine

search

>> Hello, I'm John Furrier here in our Palo Alto CUBE Studios, your host. We are here for a three-day wall-to-wall coverage of robotics and AI infrastructure. This is our second annual event, NYSE Wired plus theCUBE in collaboration, the networks grow, and the content's organic and the experts are all here in-house. But we have one coming in remote. Bill Tai, who's a venture capitalist, seasoned venture capitalist I should say, CUBE alumni of course, board chair of Hut 8, really instrumental in not only in the Bitcoin mining, but the internet infrastructure next generation. Bill's joining us here, Sanjay Gupta, chief strategist at Auradine's here. This is about large scale. Sanjay, great to see you again.

Sanjay Gupta

>> Thank you.

>> Bill, of course, coming in remote with the kite surfing in the background. By the way, thank you for your awesome ACTI event this past month. Appreciate it. Guys, thanks for coming in. Sanjay, kick us off. What's the large scale state of the market right now? Because you're seeing the geopolitical situation, you're seeing the large scale demand for infrastructure. What's your perspective?

Sanjay Gupta

>> Sure, John. So, I think there's probably two or three macro trends that are taking place. One is that in the world of Bitcoin and Bitcoin mining infrastructure, we're seeing significant ongoing growth post having that took place last year, and there's a tremendous focus upon latest innovations that we can talk about in the area of Bitcoin mining, absolutely cutting edge semiconductor and technology which is needed for data center scale mining. So that's one. At the same time, you also see a tremendous focus upon AI infrastructure for data centers, and it's also interesting to see the two coming together in different very interesting ways, for data center operators, which are operating both in AI and Bitcoin, with some of the same technological excitement and some of the same challenges and opportunities as well.

>> Bill, I want to get your thoughts, because the topic of this event is robotics and AI infrastructure leaders, of which you need energy to power the servers. You need robotics that's intelligent. You need to get that data, like electricity, into the AI physical plant, but the infrastructure is also a physical plant. So the large scale infrastructure is booming. What is your take right now as you look at the demand curve right now for infrastructure? I mean, it's not a day goes by where you hear about another 20 billion data center deployment. The demand on the token side is massive, and then the robotics engineer is getting stronger, and the decentralized infrastructure is continuing to thunder away. The AI tailwind is magical. What's your take on all that?

Bill Tai

>> I'll tell you, this is one of the biggest demand waves our whole society has ever seen. If you think about the demand load now presented by just the AI data centers, or data centers in general that have been announced over the next five years, some estimates would say that to fuel the data centers that have been announced, and funded for the next five years, it's the equivalent energy consumption of 40 million American homes, and there's around 100 million American homes that have been built over many, many decades. So that's all going to happen in five years. There's some estimates that are a little bit higher than that. Eric Schmidt was testifying in front of Congress saying that to meet the demand for AI, the US would need 27 gigawatts by the end of 2027, which is not that far away, literally just a couple of years. That's 27 nuclear power plants, and to meet the demand by 2030, it's another 63 gigawatts, or 63 nuclear power plants. We can't build them fast enough. So, I think we have stress coming out of this system that was totally unexpected, in a world where our infrastructure to provide energy is kind of 100 years old. So I think it's a big gating item, and I think there are a bunch of industries that are around the data center area, like Bitcoin mining, that can help provide a little bit of load balancing, some demand metrics to provide incentive for producers to put up the CapEx, to do things like very large hydro plants or nuclear power plants. You don't want to build those unless you have customers. Those customers will be there. AI demand goes up and down. You want something that's going to be a load balancer that's constantly taking energy and Bitcoin mining can be part of that.

>> Yeah, Bill, I love that reference to not... The unforecasted demand, and the uncertainty. Sanjay, you're in strategy. You got to play the chess game, think three moves ahead. Bill, I remember when you did Hut 8 investment, that was a long time ago by these standards. It just seems like yesterday. But I have to ask you guys both, from a vantage point of chip software and geography, what's changed in the past 18 months the most? Bill, I know you've had some options with your companies, you've been growing, there's been some corporate development work, there's clearly opportunities. Sanjay, in your business, you're at the front lines. What is the biggest change in the past 18 months? You've got the chips, you got the software, you get the geography challenges. What are some of the things you guys see?

Sanjay Gupta

>> Yeah, and John, I would say it's both in terms of challenges and opportunities. So, one is like Bill mentioned, energy is going to be critical, and energy is both on the supply side of energy, and the infrastructure for energy. Talked about the 27 gigawatts that need to come on stream just in the US alone, but there's also the demand side of energy. So one of the big things that is a key driver of both AI as well as Bitcoin mining is the need for ultra-energy-efficient machines. And that gets manifested in the form of even at the base level of semiconductors that we are looking for the latest process technology nodes. Three-nanometer process technology, getting down to two-nanometer process technology, which are the most energy efficient. So, that's one big dimension. The second big change is the whole notion about demand and load balancing. So the need to balance the supply and demand of the electrical grid is becoming even more paramount, and that's where some of the AI data centers are... Need to focus upon, and this combination of play between AI and Bitcoin data centers becomes also more relevant.

>> I love the smaller, faster, cheaper. Bill, when you guys started Hut 8, and where you are now, it reminds me of that comment about the unexpected. When Zoom was invested in, it was pre-pandemic, and then the pandemic was an unforecasted event. That created uncertainty. It was an unknown, and Zoom clearly took off, you were a big investor there, early stage. Now you have the same dynamic where there's a lot of unknown out there that's going to be in front of us soon. Your thoughts on how you're navigating this kind of news, I won't call it Moore's Law, but smaller, faster, cheaper software, chips. Now you've got decentralized infrastructure. What's your view on this?

Bill Tai

>> They all kind of dovetail. I think every wave lays a foundation for the next one. But Hut 8 was an outgrowth of a company that I funded in 2010, '11 era called Bitfury, and we had built data centers in Iceland, Finland, Norway, Republic of Georgia. I'd say some of the major changes that you were hinting at, one of the big, big changes is the change in the political environment, and the acceptance therefore of cryptocurrency related things that has kind of unleashed that industry to grow. I mean, it was growing fast anyway, but I think the gloves are kind of off, in terms of building out new infrastructure. That I think is a precursor, and a big, big positive for what's happening with AI, because the scale at which the Bitcoin mining operations had been working to lay the... I think a lot of the foundation for how data centers are constructed and built today that are appropriate for AI. In the '90s, back in the internet build out, I was also the founder of a company called iAsiaWorks that built data centers in 10 countries to lay out the internet foundation. At that time, we would build energy density of 15 kilowatts per six foot rack. The AI build outs today, some of them are 200 to 500 in extreme cases, kilowatts per rack. That's a huge step-up. If we hadn't had in between those two generations, the build out of the data centers for things like Bitcoin mining, I don't know if the technologies to deliver that energy would've been readily available. The energy density in Bitcoin is basically... It's one of these things where the more, the better. So if you can put huge, huge amounts to any six foot rack, with the appropriate silicon that Sanjay was talking about, your productivity is just a lot higher. So, I think it's just been amazing to see the utility of what happened in the Bitcoin mining sector lay the foundation for what is happening with AI today.

>> That's a great point. I'd love to get your thoughts, Bill, and then Sanjay, weigh in too, because Bitcoin represents money, programmable money, it's been talked about. But then you add AI to it. Now you have data centers as computers, and what came up yesterday in our kickoff Day Zero panel was this idea of sovereignty as not just competitiveness, and for run data in your country or in your geography, but wealth creation, and entrepreneurship. So, energy's bounded here. So bounded by energy, what is the entrepreneurial equation? How do companies combine the best of Bitcoin, the future of money, with the future of sovereignty together? Because now it's not just the Bitcoiners and all the energy, it's Bitcoin and the AI, which is the world, kind of coalescing into one. Bill, we'll start with you. What's your view on this?

Bill Tai

>> You're basically pointing out the fundamental theme of decentralization. So, while the internet by its very nature is somewhat decentralized, there's no single point of attack where you can kill it, it was primarily a US phenomena at the start. So you could almost look at it as sort of like a neural network, a system of neurons and nerves where there's a distributed set of nerves going around the world, but they all lead back to Silicon Valley in terms of the Googles, and Facebooks, and Amazon's not Silicon Valley, but basically a US centric thing, with little extensions and little pods outside in other countries. As we move towards the notion of what you talked about, sovereignty over your currency, over your energy production as a result, because currency and energy are the same thing today, in digital currency, you've got the spread now, where there's major nodes being built everywhere in the world. I think because of the US SEC suppression of cryptocurrency for 10 years, that forced a lot of the development to be outside. So that basically forced it to spread quickly. And then you've got the AI boom now, which again, looked a little bit like the internet boom, started as a US thing. It's starting to be decentralized, and I think as nation states think about what they need to do in AI, every credible, well-financed sovereign government, particularly look at what's happening in the Middle East, they are building massive capability in AI everywhere, at the center part, where you're doing the training, and things like that, but soon we're going to move into this world of inference. So as you talked about robots, and other self-driving cars and things that take AI at the edge, there's going to be another wave of edge network development that takes it yet a further step, in terms of decentralization.

>> Sanjay, weigh in on this, because it's a whole nother level, the network piece-

Sanjay Gupta

>> Yeah, no, no. No-

>> The geopolitical access is going to change.

Sanjay Gupta

>> I think this is going to be a... It's a very exciting and different times, and I've seen that cycle evolve over the past 35 years that I've been in the industry. But let me just start with one point earlier. You talked about Bitcoin converting energy into economic value. AI is actually doing the same. So in different forms, it's actually taking large amounts of energy and converting it in different forms to economic value. Now, during the days of the '90s when I was in technology, there was a whole theme around globalization, right? The whole notion was-

>> Grid computing was around then, too. Remember that?

Sanjay Gupta

>> Exactly. I remember those early days, you said, and then you said, "Let's ensure that we can have global supply chains, globalization of different parts." There's a notion that now, the output represents economic... Real economic value, and we've seen the world of ongoing geopoliticals, and geopolitical areas, that the need for different geographies, or sovereigns to ensure that they have the appropriate technological base within their own region has become even more relevant. So therefore, you're seeing a lot of drive, for instance, within the US to ensure that critical technologies in the US, you're seeing data centers, a lot of push on data centers. The same thing in like what Bill is talking about, you see different countries around the world saying, "We need to have sovereign data centers for both AI as well as potentially Bitcoin mining, or anything else that creates energy into economic value." And that is going to be, I think, an ongoing trend that we're going to see.

>> And the themes that came out of Bill's ACTI Maui retreat was also similar. You heard democratization, transparency, geopolitical, but when you look at how network computing is, there are nodes in the network, Bill. So you have a local node in a country, you don't know what the entrepreneurship is going to look like there. So you say, "Okay, assume value creation happens." That's going to spawn entrepreneurship. So while the debate is all the jobs going away, Sanjay and Bill, it's like there's all another side of the coin, which is what will be created. So, share your thoughts on the value creation, value extraction piece, because if you guys continue to do your job with the energy, and the compute, and building out the systems, that's going to enable a new class of software, a new class of methodologies, tackle the unknown, create new opportunities. Bill, what's your vision on this? Because you have a great eye around the corner. I know it's unknown, it's a little foggy right now, but entrepreneurship will thrive. What do you see evolve? You make big bets early.

Bill Tai

>> Yeah, well, we are in front of some incredibly societally transformational things at the moment. And when you go down to the level that you discussed, at the very foundational level, what is currency? Currency is basically stored excess productivity. If you think about the old feudal days, where you had farmers growing grain, and you had extra, you stored that, that became tradable, and then you would represent those bushels of grain, or whatever it might be, or excess cows or whatever with another form of currency in either paper, or a rock, or pebbles, shells, feathers. If you read Niall Ferguson's Ascent of Money, which is I think an incredible read for people that want to understand what is money, what is currency? So, if you think about what is the exchange mechanism of stored excess productivity, and how do you unlock productivity? We are in a world now, we've been in a world now for about 10 or 15 years, 15 for cryptocurrency and maybe longer in AI, but the methods in terms of tokens, that's more recent. We are, as a technology society today, generating currency by creating these digital assets, whether they're Bitcoin or Ethereum or Solana or meme coins. We're also generating, in AI, the unlocking of productivity in tokens. And if you look at the... Jensen is always talking about the growth rate of token generation. That's the key to unlocking because you use those as currency, almost like you'd put quarters into a pinball machine. So you've got to basically generate tokens, and save them up and hold them, or have access to the production to unlock the usage of the tokens, expressing efficient productivity by using AI. So we're redefining currencies. We're redefining productivity, we're redefining stored value. We're redefining our economies as we speak.

>> Yeah, I mean like to highlight the entrepreneurship, Sanjay, because what Bill pointed out is, as the table stakes get done as the utility of value can happen, and then extracted, tokens, you got tokens on the crypto side, you got tokens on the software side. So it's a token economy. The infrastructure-

Bill Tai

>> Let me chime in, John, one more before Sanjay.

>> Yeah.

Bill Tai

>> So if you think about when you redefine currencies, you redefine everything about handling them, and payments infrastructure. So look at the boom that's happening right now in FinTech, and cryptocurrencies generally. Look at the recent IPOs. What is Circle valued at today? I don't know if it's $30 billion or $50 billion, and you've got Gemini coming public. So you've got this entire world that did not exist 10 years ago in decentralized finance, ways to handle all of that value, ways to unlock that value, so the entrepreneurial sector, there's a lot of it happening in that segment, as well.

>> Yeah.

Bill Tai

>> .

>> Commerce is online, and you have to see the opportunities and capture them. That's why I'm so bullish on the entrepreneurship equation right now. But if you don't move the bits around, you got to get the job done.

Sanjay Gupta

>> No, sure. So John, I just come back to the theme around opportunity for entrepreneurship within this new economy, so to say. So I'll just build on what Bill just mentioned. I said there's been a... If I look at even just in the past 10 years, the types of opportunities, I'll give you probably four key illustrations of that. So, one area is the whole area around ultra-high-speed networking that you need to connect in the AI data centers. The same thing is true in other data centers. That did not exist until a few years ago. So the need to actually move at terabit speed, hundreds of terabit speed, that is one area there. Optics technology is another big boom area. We were at the Global Semiconductor Alliance award ceremony last year, a number of companies and teams in that. The whole area around energy efficiency is a new area that has spawned up in the past few years, and the fourth is cooling technologies, where for instance, just in the past few years, we have been really at the forefront of some cooling technologies. Like Bill said in a rack, earlier, in a data center rack, you just needed to cool 10 to 15 kilowatts in the entire rack. We launched a direct liquid cooling product just in the past few months. We are cooling, within one rack unit, 10 kilowatts. That itself corresponds to 300, 400 kilowatts of cooling in one rack itself. These are all opportunities that did not... The need was not there, and these are spawning new entrepreneurial ecosystems that did not exist even 10 years ago.

>> And this highlights the theme of robotics and the AI leaders, Bill, because the physical world is coming together with digital, and if things are digital, it could be hacked. So you have to have some really tight security. You got to have high performance, you got to have the cooling and the power, all... And the robots too, whether it's manufacturing, or a robot that's helping me out at the house, but you got to power it, too. So you need the network, you need the components. Bill, this is a shift. Just... You are the first employee at TSMC. What's your view on the hardware side? Because now you've got first party integration of digital physical now coming in, you're going to need the computer vision, you're going to need the power and cooling. You need the bits, you're going to need that, maybe even vertically integrated robot. I mean, so heterogeneous, homogeneous? I mean, are we going to be living... Cats and dogs living together kind of in one moment here?

Bill Tai

>> Well, yeah. So I think if you think about the progression of what networks have been able to handle over the decades, so of course the electronics moving the electrons, which meant expressing them in bits, in the very, very early days, if you're old enough to remember those first browsers you used and how slow they were, there was a big step using things like word processors, right? So you're expressing text, and then data, and voice. So obviously text, voice, video, and the more data you can handle, because all those are forms of data, the more value you can transmit in the same period of time. And now that's reaching the humanoid robots, and self-driving cars. So the extensions that are basically using those delivered bits to deliver productivity, time savings, advancement of our economies, it's getting better and better because of all the things that Sanjay talked about. Because without the photonics, and the high-speed SerDes and networking technologies, the intense low-power silicon in those data centers, moving bigger chunks of value in the same amount of time, we wouldn't have the world coming that we have coming today. And of course, if you think about the stability of economies, and sometimes I feel like the world is becoming a more unstable place. I don't know if that's true or not, but it's really about providing economic security for people and their nations. And to the extent that you can grow your GDP, it's all about increasing productivity per person, and it's all going to happen because of the movement of bits.

>> I think you're right on. I've been saying to my young adult children, and others, if you get involved in tech right now, you could shape the future, because data centers are as powerful as... With the sovereign equation, they could reconfigure the global landscape from a peace standpoint. I mean, a lot of people don't have the power. If you have the power, that might be a sharing opportunity, an alliance. So I think there's going to be some real changes around how we see this, because a robot is a connected network device, but it's also a data center. So you have the same challenges with robotics as you do these large data centers, which are acting as supercomputing. So, the question is, as we look at this, what is the formula for success? Because back in the old days, Bill, in the processor days, it was no one really cared if it was proprietary software under the processor, it worked. So, it was heterogeneous for Intel, or others. So the question that comes out is heterogeneous and homogeneous devices, security, sovereignty. I want my car to be locked and loaded, ready to go, but also open. So, you've got heterogeneous networks, maybe, open source is heterogeneous. What's the striking balance between heterogeneous and homogeneous in this new future? Because you can argue that you can see them working together because of the requirements. If I'm going to build a car, I might want to have that vertically integrated. Guys, thoughts on this? It's kind of a open question. It's not really talked about publicly, but I want to throw it out there.

Sanjay Gupta

>> Yeah, so I'll actually talk about maybe on this, John, the evolution of what we've seen even within Auradine itself over the past few years. So we started off with two major focus areas. One is blockchain and the other was AI. Both secular, disruptive trends, huge opportunities in the right trend. Within these two areas, we also had a big focus on security. So that's actually another interesting area that we focused upon, to ensure that security is going to be paramount. We actually created a company called Aurascape.ai, and we had that at the intersection of AI and security. Within the world of AI, we've been seeing an interesting development. It's almost like, drives you back again, to the world of 30, 35 years ago, when you had the days of IBM, and you had IBM have a proprietary stack.

>> Yeah, that's the name of it. Proprietary operating system.

Sanjay Gupta

>> Right from the semiconductor, operating system, the databases, the entire unit hardware system. And then eventually it moved into open standards, and you had open standards really drive the tremendous growth that we saw through the '90s, early 2000s, and 2010s. What we think is that in the world of AI, we'll see something similar. So today you take a world of NVIDIA. NVIDIA has got a great solution in the marketplace, but it's proprietary. You've got-

>> It works....

Sanjay Gupta

>> their GPU, right? It works great, but it's GPUs. You have the networking, which is an NVLink, networking switch-based. You have the scale-out networking, and then you have the software layer. It's all proprietary. What we're seeing in the more recent past is the rise of open standards, even within the world of AI. So you see the focus upon an area called Ultra Accelerator Link open standards. You're seeing scale-out networking called Ultra Ethernet Consortium open standards. You're seeing the software layer on the top, which is now being based upon more sonic-based open standards. So it's still early days, but if you take the knowledge you could actually see open on that.

>> Yeah. Well, I mean, everyone knows I'm biased on open standards. Bill, you know me, I love open standards. I think that's the way to go. However, the argument could be, if I'm a builder, I will develop something. I might want to have, I mean, I guess proprietary, but maybe that's the wrong word. I might want to create value, and just have it contained, but it's running in an open environment. So this is, again, I'm throwing it out there, Bill.

Bill Tai

>> I think that... John, I think that always happens at the beginning of new industries, because typically you have a bunch of humans that are experts in their field, and they're so far ahead of everybody else, or far enough ahead of everybody else, that they have to solve every problem by themselves. Think about what Tesla did with electric cars. There were other electric cars before, of course, but they solved a serial set of problems. And like Sanjay was talking about with IBM or digital equipment, those were giants that did their own silicon, their own assembly, their own operating systems, their application sales service support, because they had a unique architecture at the time. But we live in a world of entropy, and there's always pressure, and if you've got a multivariate equation that's stacked, if any of those layers has entrepreneurs in them that do something a little bit better, there's economic pressure to horizontalize. And in the case of IBM, it became Wintel, right? So you had the processor layer open up, you had the operating system layer open up, a bunch of applications developers that Microsoft ended up taking over. But eventually things become more economically efficient. And look at the silicon industry. Intel was the king of the proprietary stack. Architecture, manufacturing, design, silicon, all the way up and down in that area. Now, TSMC has carved out... They're the king of the horizontal manufacturing piece, and there's a whole other world of design economy sitting on top of that. So, I think you're right, John, that it's going to start with companies, say like a Figure.ai that are doing everything in robots, but eventually you'll have enough maturity in the industry that parts will evolve. I don't know if it's a 10-year thing or a 30-year thing, but eventually there will be some standard kind of robotic operating systems, and then the interconnectedness, that presents a whole other set of issues. When you talk about sovereignty and security and defense, think about what's happening to the defense industry now. Certainly you will still need human soldiers, but how much of that is being replaced by drones now? And like the computer business went from $10 million mainframes to $1,000 PCs, and many of them, what's going to happen to these big jets at 30, or 40, or 50 million or more in the face of $1,000 dollars drones?

>> Yeah.

Bill Tai

>> Swarms of drones I think are going to win in the end, if they're handled right. So I think we are in a world of fundamental entropy, enabled by technology, towards a path of decentralization, but it's all going to be networked. So, back to your point again about internetworking. Now we've got a world where you can almost look at the digital networks as sort of like a human nervous system in a body, where you've got this connected... You've got little operating system for your hand, one for your feet, one for your legs. They're all tied together, and they're all kind of connected to a central decentralized set of neurons in your head. That presents a set of security issues that we haven't seen yet, because you could have viruses enter somewhere and spread across that network and do crazy things like that movie the Obamas produced with Tesla cars all...

>> Yeah. Taking over.

Bill Tai

>> I think the fear of having Chinese cars here, that suddenly in a war, someone flips a switch, and all the cars just drive themselves, and crash on freeways, and disrupt our transportation system. And if you were running a country and you really wanted to do war today, you see the war in the Middle East where they're trying to bomb nuclear power plants? You wouldn't do that if you were attacking the USA. You would just find every major data center of Amazon, and drop bombs on that, and America would stop. Right? So we have to-

>> They've got satellites, too.

Bill Tai

>> Yeah. .

>> Guys, great insight. Great insight on the open versus... I think industries do go vertical. Great insight, Bill. Just to wrap up, I want to just find out what you guys are working on. You guys are experts and leaders in the industry. Share what you're working on right now, because the robotics, AI infrastructure, decentralization, Bitcoin for money, the confluence of all these things coming together presents great opportunities and challenges, but the future's unwritten right now. Sanjay, start with you. What are you working on? What are some of the things your company's doing? Can you share some notable breakthroughs?

Sanjay Gupta

>> Sure. So John, so our company was set up to really take care of two major disruptive forces, like I mentioned. One is around the world of blockchain, and the other is the world of AI, and the two are somewhat interconnected. In the world of blockchain, we focused upon Bitcoin, and really driving the forefront of Bitcoin on a few key areas. So there we're focused upon absolute latest stage energy efficiency, we're focused upon the ability to actually move your energy up or down in response to the energy ecosystem as well. We're involved in new types of cooling. We just launched a new hydro direct liquid cooling system that I mentioned can cool 10 kilowatts in one rack unit itself. So that's where a lot of innovation that we're actually focused upon. In the world of AI, we're focusing upon infrastructure and the need for critical infrastructure in the form of ultra high-speed networking. So, we're actually focusing upon ensuring that you can allow multiple GPUs to interconnect at ultra-high speed, form what is called a coherent memory stack within that, and the ability to actually manage that in form of open software on top of that as well. So, it's a very exciting time between both them. And then we're also focusing upon the interplay of taking cooling technologies from blockchain to AI, energy load balancing from one to the other, is what they're focusing upon. So again, we think that both are tremendous opportunities areas, but both in the area of disruptive infrastructure technology for both .

>> That's interesting, Sanjay, because all the hot software conversations I love to have, what's the future operating system? Bill was kind of hinting about the openness of it. It's all networking. The software action's in networking because you got to connect all these large-scale systems together. So the software in networking is super hot because it's controlling and linking things, connecting the data, loading the data. It's like... Sounds like an operating system to me.

Sanjay Gupta

>> One thought just on that is simply the analogy you often give is that if you have the basic core processing, taking the case of AI GPU, if that's like the car and you're trying to develop faster and faster cars, you need the highways and the interconnects between that. If you don't have the highways and you have potholes on the roads, you're not going to be driving your ultra-fast car very, very fast. So that's what you're focusing upon to create that highway ecosystem, that city ecosystem to allow these cars to interconnect and move around .

>> We know they're going to definitely have guardrails, because that's the number one conversation in AI. Bill, I want to get your thoughts on what you're working on. Hut 8 is transforming. You have multiple business opportunities ahead of you. You're obviously doing some great investing. What's the cool breakthroughs you're working on right now? What's cool and relevant in your world?

Bill Tai

>> Yeah, I'll mention three things. One is as representing Hut 8 as chairman here, we are migrating what started as a Bitcoin mining company into an energy infrastructure company. So we've basically taken our Bitcoin mining assets, dropped them down into a subsidiary. That subsidiary is called American Bitcoin. The Trump family owns 20% of that. We own 80, and we're going to do a set of raises, of course, to turn that into a separate publicly listed company, preserve the energy infrastructure. We've moved from kind of a gigawatt of energy under management to now 12 gigawatts plus at Hut 8, at the parent level. And we've taken our GPU data centers, and dropped them down into another subsidiary called Highrise.ai. So I think I want to move to where the puck is going. And as John, as you know, I think this energy shortage is a big deal. So I think we want to be in that game first. I think the other really amazing companies, Canva, of course, John, you were in the room three weeks ago where Melanie pitched Canva to get her seed round done.

>> Yeah.

Bill Tai

>> And that company has just been on fire, eight years of profitability, growing at... Well over $3 billion run right now, profitable for eight years. And it's a company that really encapsulates a new economy. If you think about the decentralization of work, and people all over the globe, in a world where your storefront is online, whether it's a real store, or your own persona, everything about you that's represented to the public world is done digitally, and Canva is that interface now that makes it really easy for you to explain, what is your business, what do you do, and now they're driving also into productivity tools. The new suite of spreadsheets, no code, other products that they have, just incredible. So I think not only will you be able to express who you are, you're going to run your business on Canva. And then lastly, I'd say the other area, I'm always looking for massive markets that are going through disruption. The biggest market that I'm addressing at the moment is US healthcare. UH healthcare is 19% of GDP. So one out of $5 spent in the United States is spent on healthcare. You have organizations like CMS.gov that Dr. Oz just took over, $2 trillion a year of payments. And a lot of that is not done efficiently, to the point that leaders in the payment space like UnitedHealthcare reject 32% of all claims, 100 million Americans are in medical debt, 60% of all bankruptcies happen because of medical debt, and CEOs are getting murdered, because the system is so messed up. If that isn't an area that speaks giant problem to solve, I don't know what is, where there's people getting murdered. So we have a little company called Dyad.net that's growing 10 to 15% per month every month, that does a semantic AI healthcare solution, that provides a visual graphical analysis based on a diagnosis of all the possible things that could be tied to the things a doctor sees, ties it to the insurance coding, and make sure that the things that are recommended are generally correct and less likely to be rejected. So we want to make that whole system more efficient, prevent all these bankruptcies, and hopefully no more CEOs get murdered.

>> And you think about the efficiencies, you've got the infrastructure, critical infrastructure down, and the compute storage, and networking, and database, and decentralize it, and then all that productivity gain with AI coming in, Bill. I mean, it's a huge... If you're an entrepreneur, you want to look... Your eyes are wide open, you're running at that problem. So, great, great call out there. Bill, thank you so much for coming in remotely. Sanjay, great to have you in person again, great to see you, and we love what you do, make things go faster on the networking side as well, and the blockchain work, good spot you're in. Thanks for coming on.

Sanjay Gupta

>> Thank you very much. Thanks, John.

>> Bill, thank you so much.

Bill Tai

>> Thank you, John. Thanks, Sanjay.

Sanjay Gupta

>> Thanks a lot, Bill.

>> I'm John Furrier here at theCUBE for the robotics and AI leaders, sharing their thoughts, opinions and commentary, and analyzing the market and presenting opportunities, folks out there, companies, society, countries, entrepreneurs. That's what we do here at theCUBE. I'm John Furrier. Thanks for watching.