China Tech Still Strong in Spite of US Pressure

Bloomberg

2 days ago

00:00
While deep sea stunned the world and made headlines earlier this year when it released its air rivaling that of open air models at a fraction of the cost. Meanwhile, you've got robots from unitary out of Hangzhou running factories from AI to cutting edge robotics to EVs. It's very clear that China's startups are working overtime to really close the gap with the US and dominate the technologies of the future. We've got a great panel here and I want to chat, chat with them about all of these developments. Now, Yuan, let me start with you. Venture capital investments in two Chinese based companies last year, as well as in the last couple of years have been falling. You're one of the biggest Chinese angel investors there. What do you see as the most exciting development? I mean, has deep sea really sparked this frenzy in terms of AI and finding the next deep seek there? Yeah, I think while Deep Sea is a very special case, this is that they actually never took any venture funding. It'll be started as a side project of entrepreneur who did contracting himself. So. So it's it's not really it wasn't a commercial project to start with and we're obviously not investors in deep sea but I think we did notice from deep sea that they are starting to be very genuine innovations in both the like the very foundation of foundational model layer as well as application maker in the in the ventures. I and I think that it did boost the overall confidence in the entire entrepreneur community in China in building something that is a world class. That's great. And I want to talk to you, Will, about there be a capital. I mean, obviously you guys started as an investment arm of German media conglomerate Bertelsmann and of course, now have a portfolio of more than 200 tech companies there. When you look at the tech space over the next 18 months, what are the most exciting places where there's unlocked potential? What are you guys specifically looking at, especially as it comes into? I know I think the post moment and also the reasoning models give opportunities for especially applications. And we see developers in China are really I think they are really advantages in developing these apps and also ship it out to global market like what we have in China recently. We see a lot of like general agent products and also vertical Asian products going on. So I think in the next 16 to 18 months, we will see a very good group of new AI applications and also enterprises going. One example here is if you look at the video AI video generating apps right now in the Global App Store ranking, I think three or four out of five top ranked apps are from China, like Halo or like clean pics versus. So we already see like these players battling in the global battleground and taking shows from what used to be like opening and opening related applications in the market. Yeah, great. Now, Esther, you're kind of the anomaly of year, the battle in that you don't really invest in Chinese based companies, but obviously companies globally are very important. I wonder if you can talk about your strategy of why overseas and not specifically Chinese companies. Well, I mean, think about how long it takes a true deep tech company to be profitable, right? So Salesforce, it took Salesforce ten years. It took Amazon ten years. It took Uber over ten years. It took Tesla. I don't know how many of you know this, but it took Tesla 17 years to be profitable. Now, any of these companies, they might have a hard time getting listed in China because as all of you know, all of us here that the Chinese are local market realize that there is actually very strict listing rules. And what and behind that strict listing rules is very strict also investment rules. So when you actually take money from, let's say, a local entity, usually government related, they actually would have a lot of criterias that comes with it, such as, you know, you want to relocate your headquarter there, you want to put some, you know, research team there. You want to do Sidney Poitier here. And one of the mutual is the time to IPO. So that actually gives founders a little bit of, you know, mismatch incentive between how to grow a company versus how to make the company quickly commercialized. I mean, these these to fruition, nothing wrong with it. It's just that companies in China are more encouraged to be having a faster commercialization path than, you know, outside of China. So nothing about my investment where I think I can make the biggest investment impact, the investment impact and I respect that people are investing in the application size agents and all those robots are fantastic. But I think the biggest impact would actually come from A.I.. Infrastructure, i.e. how we currently create the process of AI compute. The air factory that we know today is actually dated, right? The reason that Nvidia went up three times in five years is because GPU happen to be the right tools to treat models. But now 2025, we're at this pivotal year for the first time in history, the compute used for inference, i.e. people that use this every day versus the computer used for training is actually higher. So that means that there must be a better way to produce AI to meet the surging demand, right? Jensen Huang from NVIDIA said, you know, the demand for for treat for inferencing is 1 billion times higher than training. So of course we need a new paradigm in how to do inference. So these company may not get profitable in five years, ten years, but I want to invest in these companies. So that's a very long answer to say why I want to put some money outside of China and outside of China where specifically or I mean, is it most of us that you guys are looking at? You know, you have U.S., you have Japan, Singapore. And so so a lot of these innovation can come anywhere. Right. But U.S. is a very natural destination just because historically U.S. is a place with a lot of breadth and depth of different kind of VCs. And so, yeah, you always take on a good chunk of it. Great. And of course, Tim, I mean, you invest in public companies. How are public companies thinking about when it comes to unlocking the AI potential there? And the companies that you're investing, particularly in? Chips is obviously a big part of that. Yeah, chips are obviously a big part of that. I think in the past two or three years, if you look at public equities, a lot of these value creation is were coming from the infrastructure side, which goes up to natural gas, nuclear power powered turbine and goes all the way down to how do you build a data center, You have to have connectivity, you have to have cooling, and also you need chips at the end. So I think in the past two or three years, it's about debottlenecking the entire supply chain to build the factory to to make the models work. And I think and if you're looking to that value creation process, a lot of that, I mean, predominantly a lot of that value creation come from public equities, public companies. And it's I mean, it's it's I think it's a it's how it works in the previous technology revolution as well. But I think going forward, which which is very interesting, as Esther mentioned, is that, I mean, we're moving from training to inference. We're seeing token I mean, token consumption on the inference side, I mean, skyrocketed this year. So that creates an interesting angle to think about. If you look into the technology cycles, every every technology cycle is starts with technology in the lab, in the university. And and it's it's a physical and we invest a lot into the infrastructure, as was the PCs in the eighties. It was the Internet, it was the Cisco in the nineties and it was the tower is 3G in the 20 tens. And now we're investing a lot into the data centers. But the value creation and the next phase, if we're truly believer inference, I think it has to come from applications. We have to build real. I mean, I mean we need to build real applications. We need to we need to solve real problems and we need to create real use cases. I think that's from our perspective, going for the value creation. I mean, from my perspective, I look at public equities, I think the market is going to be tremendous. Sorry to use that word, but I think we look at education as we look at financials, we look at health care, we look at even gaming. We see a lot of public companies. They I mean, they've been serving their client, they're serving their users for for multiple years. They have a lot of data. And now with A.I., they started to be able to leverage AI, revamp their user experience and create something new. So I think that value creation in the public equities are going to be very different from what we've seen in the past two or three years, which is, I mean, very narrowly focused in to semiconductors and infrastructures. What do we what do you think in terms of apps? I mean, in what ways are the apps leveraging A.I. and and yeah, what's the most exciting stuff that you're seeing? I think people recently are talking about one man unicorn and, uh, and I think that's really, really possible. I like a couple of entrepreneurs are invested in what when do you think and where will that merged what may unicorn Instead? It's like one or two years in e-commerce, in sales and the social and maybe one one big K AOL on Twitter or on the Instagram can be that scale. Yesterday, I think last week I met. A funder, which is like a couple. They generate like 10 million revenue from selling stuff on Amazon this year and they target to do 100 million US dollar sales next year only it still was a couple and the lady and the wife is still taking care of the kids. So basically the leverage agents to do all kind of like execute development, sourcing care, reach outs, marketing and, and also supply chain management with all kind of agents. The founder has 20 years of ecommerce experience and he put all his knowhow into orchestrate trading. This the whole workflow. And his goal is to expand business without expanding hiring humans. So yeah, a couple of a couple running 100 million USD revenue, selling maybe 200 300 SKUs on Amazon and Tick Tock next year will see these kind of business. And OpEx is it's like marginally zero, right? So it's only just feeding the couple and the child. So so the margin will be high and we'll see that kind of business emerging not only in potentially like people easily to imagine the SaaS, but also I think in e-commerce, in cloud, in gaming, we see that kind of things emerging recently. That's very promising. Yeah, that's I mean that's that's amazing, right? A couple, Yeah. Running up. Yeah. I mean, I guess, you know, you could run it from your bedroom. Yeah. There are not even doing arbitrage. This is not near not selling the SKU that Chinese manufacturers are not selling in us. It's just using AI to develop the SKU and also try to use A.I. to capture some of the long tail demand from the market that usually a human operator will easily neglect. So that's very interesting. Hmm. That's great. Now, you guys invested Jen Fund recently invested in Manus or reinvested in Manus as well. And that all that most of that is is that work I saw it on on the YouTube was really cool in terms of what they were doing in terms of agents. I wonder if you could talk about that kind of growth and why Manus? Are there lots of other Manus is around that you guys are looking at? Well, yeah, sure. I think almost every Chinese venture capital investor is looking at, say, local team trying to build a second or third of Manus and they literally sometimes even do look the same. But my guess is if there is a second Manus, it wouldn't look like Iran. This is actually a very interesting and interesting story in the sense that so so the funding him arranging it here from China by the product not just madness but also the preceding product. Ammonia was actually not even available in the China market. So from day one for a team that has never been to the US or I'm not even sure if it's being out of country at the time, we're able to build a product that is used globally outside of China and because they work with a lot of models like say, the US companies openly and so and they do not provide a service to the local market like the US model so that the products I should only available also to China. So I think that's a very interesting phenomena in the sense that this season entrepreneurs who went through the very ferocious competition during the mobile era in China that they are probably born and raised up in China but were kind of self educated with the US narrative like they they grew up on Twitter in a sense. Since high school they've been scanning Twitter and Reddit. They not just men as founders, but also like founders that we have back to in China, that they actually there are startup manuals here and the work and Jesus books, right and program is like their God basically. So they were able to experience the mobile error war, the competition control competition in China and at the same time be able to stay curious, exploring everything happening in the Silicon Valley. And I think as a result, the former experience like the like the battlefield experience in competing in the local training ground in China provide them with the capability to iterate really fast. And at the same time, they remain very educated and curious. All the new playbooks like how to hire, how to scale, how to kind of explore new paradigms from from the from the valley world, right? So and they hear the pitches from there are kind of this kind of mentors in the in the valley. So like I always joke about how you know Peter to an American dream and have no idea like how many mentees they have in China. So I think it's it's very encouraging for us to see that such a local team can build a such a international product, not just from the user base standpoint, but from the product design, the aesthetics of the product. And and so, yeah, I think we would see more and more of such teams be able to build a like a product on a global audience. But at the same time they do have to tackle a lot of geopolitical issues and try to avoid being the victim. Of the bigger debate they have little control over. So, yeah, we definitely want to talk about the geopolitical scene as well. I want to remind everyone we've got a QR code up. You know, we'd love to hear questions from you as well, so please use those QR QR codes and send us questions. I guess it I mean that in some ways that's a perfect segue to you. I mean, you were talking about allocation of resources. We're talking about talent. I mean, you're looking obviously overseas, but a lot of these founders and when we were talking just just behind behind there in the prep room, a lot of these founders that that you're focused on, I mean, they are Chinese, you know, origin as well, right? Yes, I think manners is a perfect example of the best marriage of both worlds. Right. Because, I mean, what is it? What does it take to create itself in its four elements? You need some in these data, right? So that data is a publicly available. Same in China, the same in us. You need algorithm. Algorithm is talent. So, I mean, China is uniquely advantage in this point because even look at China, even U.S. scientists, half of them has Chinese origin. So we know that we are good with, you know, texts of. So that's the machine. The mystery is the hot weather in which there is a lot of debate on that part happening. And the fourth one, and I think very important one is the feedback loop, like how can you iterate a product quickly? And that in that sense, China has a very unique advantage because because of the population was such a vast country and people all have the propensity to try something new. So the testing feedback loop is really fast. The, the refining and we either rate the products really fast and if we can bring that along, the talent and the product market fit that, let's say refine in China to promote overseas is such a very good marriage isn't to be fair is not an area that I invested in. I just feel that this is a very good example of how, you know, two wars can work together. All right. We've talked about bottlenecks a little bit, too. But I want to come back to you. I mean, what do you obviously ships and data centers, You know, these are the structural stuff is very important. What do you see as kind of the biggest bottlenecks for for for and what are you looking at in terms of companies that can address that? I think we look at that, I mean, which basically we're looking at two categories of of companies. One is the companies that are building all these data centers that need the infrastructure guys. And we've been getting that question a lot about I mean, China is under-investing. And I because if you put I mean, put all these numbers together, Alibaba, I mean, recently increased their CapEx. But if you look at that, I mean, they're spending about 20 billion a year in CapEx. Tencent is slightly below that by tens is about 20 billion as well. And then you have some of the smaller guys and if you add everything up, you get to you get to about 60 to 70 billion annual CapEx. And if you look at us the last time we checked, I mean, so 350 billion annual CapEx, so not is a huge gap. It's like a one, two five. But if you look into the data, I mean, two things. One is very importantly, all the hyperscalers in the US, they're actually investing globally. So I mean, they are I mean their business is actually global, so it's not very fair to compare most of these Chinese companies versus their U.S. peers. And that's the first thing. And the second thing is a lot of these CapEx in the US are going into training. And and in China, one of the bottleneck was I mean, we don't have access to the state of our chips, which I think gets a little bit tricky for us to train the state of our model until I think deep sea came came along. And so the point I was trying to get is if you can find a way to efficiently train the state of our model or if you have a business model that focused solely on inference, not training, because someone else has already trained the model and it's a good model, it's an open source model, actually saves a lot of the the prerequisite investment. So so my argument is that from an investment perspective, I think the people will have the willingness to invest. People actually have the balance sheet to invest. The number actually are I think from from our perspective, if you're investing 70, 70 to 80 billion every year, focusing on inference, I think that's actually a good number. So the bottleneck is and the other bottleneck in the US we observe is power. Power is very power is it's a it's a very hot issue in the US. So and it's not such a big issue in China. So that's very interesting. So so from our perspective, if you look at infrastructure, I think the the biggest bottleneck is on chips. But the silver lining is again, back to deep sec. If you have someone acting as a air lab train, most of the models and an open source to everyone. And then the question for everybody else is just to do inference and the beauty of them, for instance, that you don't need state of our chips, you can train, you can you can you can use some of the outdated chips to do inference. I mean, one of my friend working in this in his companies are I mean he was joking if you if you're brave up everything I use I mean, you can you can you can use everything. As for inference, you can use the CPU as an inference. It's just a matter of cost. So so we can we see I mean, the the domestic alternatives like the Huawei or can we call some of these these companies and they are actually offering very good alternative if you're using their chips for inference. But if you're building a large class of, let's say high hundreds of thousands of chips to do training, I my understanding is and there are still bottlenecks. So so that's the infrastructure category. On the applications, I think there is no bottleneck. I would argue that. I mean, from Chinese perspective, from with a mature technology, if you want to build an application leverage attack, not not technology to serve your user, that's actually something that China has been doing very good at. We have much more sophisticated engineers, I mean, by sheer population engineers, product managers. I mean, those people were, I think, well-trained in the mobile Internet age. So I think on that category, I'm actually very, very confident. Now, just to remind everybody, Chad, Djibouti was launched in November 2022. And for the first year, you don't actually have very good product to say. And I mean, all these cursor, all these leading all these other other things, they I mean, started to ramp up last year. But we in China has only and turning to the tragic beating moment, if you consider deep sea as the tragic beauty moment of China, we've only been here for like three or four months. So I would argue people need to be a little bit patient on that, probably a little bit longer than exactly. But for foremost, yeah, four or five. Or our one. Yeah. Is there. And I guess I guess there will. I mean, in terms of the bottlenecks that you see and and you know, what Tim is talking about in terms of ships, I mean, if if China still cannot get access to, like, you know, the highest basis in video chips and obviously the U.S. is saying that's not on the table. You know, that's obviously something that they really protect. How do you see then, you know, what advantages then does China and do Chinese companies have in terms of developing? Um, you know, we see a lot of innovations from China that making those unaffordable service costs down. And like the example I all this the video generation. So I think right now there are certain app picks first that you can generate 5 seconds of video just using just 5 seconds. So and also it's free and the company still has like 70, 70, 80% gross margin because they have a very strong cost sensitive like tactics to make everything affordable. And also for applications. I think right now in China, we see about double all kind of like search based chat apps, air chat apps. So already, like the menu aggregation is already 50% of the Baidu, Baidu's volume. And also in terms of quality, I think it's run like 10 to 20% penetration. So I search in China in terms of menu penetration is already higher than that of us open now versus Google. And so and also it's also free to everyone. And I don't think there is a burden on Bytedance right now in terms of offering this totally free to everyone. And also I think in terms of application innovation, Bytedance, I think this year the revenue and profit and also growth rate will outpace matter and the valuation of its stock is only like 20 30% of the matter valuation. So in terms of investments, we see a lot of a very affordable assets in China, very competitive, demonstrated, demonstrating very good PNL and also growth potential. And for us it's really good time to invest. And what they are doing is shipping product at affordable price to mass adoption, which is a huge topic for AI in the future. So that's what I'm seeing right now. I don't see bottleneck. I think like everybody is using brains smart enough to do everything based on aesthetics and also cost reasonable cost structure, not overspending into new and all those data centers that may not in aggregate generate that revenue in the future for investors. So, yeah, I mean, certainly lots of huge valuations. What do you think is a runway in terms of of actually capitalizing on that? And for for companies like the AI, investment firms, like the AI, what's the runway for cashing out? Yeah. Oh, yeah. We recently had a new form kicking off and it's also kind of like very good fundraising programs. And also our existing ones do have two years up. Not, not really two years. We started 2021. So yeah, as a 2031. So we still have lost to invest. And also this year I tell my colleagues, like it's it feels like going back to 2013 when I was talking to a sole half million you a 70 million valuation a super affordable and micro data change like the due rate from 7% to to 15%. So I joined the team bringing all the algorithms. I feel like the same vibe right now, seeing all the applications with revenue, air and also traffic ramping up. And we see a lot of like good entrepreneurs. We're super ambitious and we see deals like under 100 million valuation and again just told me like average valuation they recently invested, you know, in terms and doing investment is super affordable, less than 10 million. And we do have some very early stage 10 million valuation deals, very, very ambitious and also legit founders. So yeah, I think our bullet was prepared for more, more growth deals. But recently we invest a lot in early stage founders and that's that's something similar to 2013, not 2021, although some of the deals right now is feeling like 2021, the consumer sector in China, which is super expensive, but the overall vibe, average vibe is to support early stage and to produce right now. Yeah. And in terms of what what the what the prospects look like right now, is it expensive? Is it a good time to come in? I mean, what's what's your outlook in terms of the next 18 months? And and yeah, I, I was sharing with the well over lunch about the valuation number which she did not disclose but it was surprisingly low even for us. But I think part of the reason was that I think there was a bifurcation of the type of deals that investors are interested in nowadays, and there are superstar founders who are very cool in the past. And I think when they come to a startup, it's very it's very pricey. And then there are the other sets of, say, very young founders, usually first time founders, if they're ever in school. And it takes some courage for the investor to make a bet on those people. And I think in the mobile era when there was a more prosperous market, there are more investors investing in such, more like, say, I would say audacious bets tend to more green founders, but I think less so these days because investors, even venture capital investors, are supposed to take a risk. They have more kind of like a security mindset and they're more betting towards superstar proven founders. And we do those type of ideas, too. But but then we continue the bets at the very, very, very young founders who start building companies very early. And I think we face much less competition than before. And as a result, for this group of founders, it's a much lower valuation we're entering. So we think it's great to take off with the path and make contrarian bets. And I think hopefully that the courage those great founders will pay off. Great. Esther, I think let me see if I can ask you this. We've got a question from the crowd here. With Chinese tech becoming more prevalent on the world stage, what further role can Hong Kong play aside from the standard benefits of the city, such as low taxes, connectivity, etc.? You leave me the hot one. Okay, So, yeah, I think the previous speaker see me saying every time when he comes back to Hong Kong, he feels like he's going back to a museum. There is some elements of truth in it, right? Because Hong Kong, despite the fact that we have a very, well, structural legal framework and a very, very free market, the adaptation of tech is pretty darn slow. Let's face it, right? I mean, we we still we still paid our taxes mostly by cash. Right. I mean, we we we don't we don't do a lot of things that, you know, is taken for granted in the greater part of China. But, you know, Hong Kong really does have I think I'm quite, um, pressures and close to my heart, as some of you probably know, I used to I'm one of the earliest members of Sensetime Sensetime is a Hong Kong first a unicorn and is founded by one of the professors at Chinese University of Hong Kong. So among the world's top universities and research is to study A.I.. In fact, Hong Kong, all five universities in there. So what Hong Kong has is a lot of AI and related talents. What Hong Kong does not have is after you nurture those talents, how do you incubate them and how to help them to go to the next level? Because I used to I used to be the judge for Hong Kong lab, which is that which is an organization between the government, the science park, and then Sensetime and then Alibaba. So every time when I look at these applicants from from from Hong Kong universities, I feel like, yes, this idea is fantastic, but there's no testing ground for these ideas. So at the end of the day, I see that these ideas die because it's a student project or for the very, very few to survive. A lot of them actually went to Shenzhen because the market is significantly bigger. And remember why I say to explore things, to create AI. The last and I think the most important thing is the feedback loop in Hong Kong. If there is so many regulations, the implementation is difficult. The feedback loop becomes really long and inefficient. And I think that's one unique benefit that China has. And Hong Kong should take more advantage of the assets to China and to more access to the Greater Bay Area resources so that we don't waste these talents that's being incubated in Hong Kong. That's and I think talent is obviously such a huge issue as well. And I know in the prep room you guys were duking it out because fighting about. Yes, fighting. You have to tell us somebody stole somebody. She accused us of skinning an intern. You guys did. You have not confirmed that. Yes, you guys did. We'll look into it. So now because, you know, because of geopolitics and I'm sure you guys are well to with this. But I mean, would you see a trend of a lot of these super smart Chinese students that from U.S. universities that come back to, say, Hong Kong, China, back to Asia in general? So we have this amazing kids from Carnegie Mellon. She in turn that matter and Internet and video and she's supposed to be our summer intern. She's written for her, so she's fine with me. And then after she saw me say, I'm so sorry, I got I got approached by another nameless V.C. that they want to go and then say, why is that? Because, you know, because interestingly, last question. You know why? Because she said there's not much going on in Hong Kong, more going on in Asia. In her case is Shanghai. More going in Shanghai. Again, going back within because you offered way more money to your government. I Is that true, though? I don't think that was the case. But we have a big intern class, so I don't know what attracted her in terms, but it could be the commander between the interns. I don't know. But but to be fair, that there is that there we do see a lot more people like the Chinese talents in the US incubated there and then coming back to to China, to Hong Kong, you know in recent months especially. So that's that's been a trend. Now do you guys think that might be accelerated, accelerated now at this moment, because obviously the Trump administration has been talking about limiting student visas, targeting Chinese students in STEM as well. I mean, will we see have we already started to see reverse migration? Are people calling you up for jobs? Yes, I'm really in gratitude to what the chaos he created. The market that's really, really pushing a lot of Chinese talent is finally clarifying their mind. That's the U.S. option. And really going back to, I think, post deep sea moment with we had our AGM was or our LPs. It's the first time in the last three years that really all of most of our foreign LPs come to Shanghai, come to Shanghai and try to understand what's deep sea and going on and understand the portfolio there. And they just like looking at the new fuel. It feel like they are looking at the new land, like familiar, not familiar new land, that all the fundamentals are really good and all of the talents are really, really impressive. And so yeah, so we have to transfer one. China is Chinese, a Chinese funders going abroad, sending up entities in Singapore and also trying to go to U.S. We have several portfolios, Founder SpaceX himself into Silicon Valley, talked to the Silicon Valley funds and trying to make it a global company or pretend to be a U.S. company. But also we have a lot of talents coming back from us, especially US university graduates, and they're are not going to UK for sure and they are going back. So that's a that's a good sign and we are really happy to invest in those talents right now. Right. Well, we only have less than a minute left, but I do want to throw to all of you one question, since I'm also interested, what in your mind is the most exciting use of AI that you're dreaming about, that you're looking to invest in any kind of crazy, wild ideas out there? You wanna start? Start with the venture guys. Oh, but where? Well, we never have a, like, a premeditated thesis on thematic, like, say, products or a sector. So we're more founder focused. So. So I guess we'll take it from the founder perspective. So what we're saying is that founders, where we're looking at are younger or younger these days. So today, actually there was a great company. I'm we're now investor of Insta360 that just went public this amazing IPO. And I think two years ago the founder was like in the famous batch a post nineties founders like founders who are born after 1990. So we're like seeing a lot of founders these days that are born after 2000 so the post 2000 founders and sometimes we're like, Oh, this is not enough. We're looking for if I had founders at like say, 17 year old, right, when they're just about to enter college because, like, that's so because all the technology is so new, every employee is kind of forced to the same level as even starting ground. And I think the founders are they start to get their, say, computer science education, younger, younger. So where you're saying they start to be able to build something when there are, say, like 1230 or something. All right. Just like the classic generation of founders like job. So it's going to work in the U.S. So we're starting to see this kids being able to have the capacity and the resource to explore something on a purely out of curiosity and to start building. And so we're just really interested to see that, you know, what is the future? Who is the future, say, putting a founder there where we can bet So that and that's our perspective is that of a product that we can envision. Yeah. And U.S. So for me, I'm a little bit biased because I don't invest in applications, so I don't back in one man billion dollar company. I do a lot of infra and I do believe that the most exciting thing is happening on how to reimagine how air is being created. Most of the there are many ways a can be created significantly more efficiently. So one of the ways the thinking is how to make the data center itself more efficient because data center takes 12 months approval from the government. If the by the line you have to build a building and wants to do the building, 50% of your tech out of your electricity bill went to cooling. That is absurd. Okay. So we are actually backing a company that build data center in space because in space, solar energy is free. Of course, 99% efficiency versus 13%. And you don't have to pay the government, you don't have to worry about building. And guess what? Space is -250 degrees. So you don't have to worry about radiation cooling super effectively than the global cooling either. So so that is happening is not science fiction. We're going to have a first launch in August with Space X. I'm super excited. So that's one of them. And an example is how to make compute itself more efficient, because all the compute that we know right now is based on transform based token calculation. Okay. However, I'm a mom, okay? My daughter can look at dog twice and know that the third dog is is a dog. However, the same computer would take a look at literally 1000 pictures of a dog To realize 1001 picture is a dog. Why is that? Your brain is a lot more intuitive than the GPU. So the future how to make this compute more efficient is you should be bio computing. The data center will have two silicon silicon based chips along with bio computing chips. So I'm also very happy to say that we are one of the earliest backer of the world's first neural computer. So you use your brain neurons to make a CPU. Now, this is not science fiction. I'm very happy to say that the company has actually commercialized, so they've been getting significant interest. We've been getting calls non-stop, literally nonstop from drug companies, from robotic companies and actually from universities around the world. So I think that we have to rethink how to create AI, and this is what my company is focusing on. So we're incubating these companies to rethink AI production. But does truly sound like science fiction? Yeah. Tim, can you beat that? Yeah, well, I can. I'm a public equity guy. I, I'm, I'm very short sighted. So dream case of a I maybe I become very good invest now. So. So these guys don't kill each other. Try to fight for interns just welcome to enter here she will regret it and joke a side I think I mean I think I don't have a specific case of I mean like dream. I use case because if you if you listen to everybody talk about the big pictures of AI, it's about a AGI. And AGI essentially is do everything for you. And and eventually you don't you don't you don't need human. In the loop. So what worries about Mueller? I mean, what intrigues me is if you listen to all the air shows, I think 99% of the air pitches in public equities or in private equities in venture people were talking about saving time and saving money. It's all about efficiency. It's all about getting all the fat out of the system. So if you look at an M7, you look at I mean, I was I was I was checking the data this morning and my analyst sent me if you look at the recruiting data for, for undergrad student in computer science, they started to find it and they started to have a very difficult time to find a job in the US. So, I mean, something that I actually wanted to see is that, I mean, again, if you're looking to the technology revolutions, I mean, every cycle it's it starts as a efficiency tool. I mean, personal computer, any archiving, stuff like that. Everything starts as a efficiency, efficient productivity tool. And what really happened in the in the process is that we created a machine or created a technology that actually help us not only save time, but also help up, help us somehow make kill time, help us somehow keep us occupied. Imagine if one coder with cursor can code like 50 coder 50 developers a previously. You know what happened to the other 49 developers? I mean that they have to do something. So I think that's a yeah well may have somebody that's a that's as I mean I don't have an answer yet but I that's something that I keep in mind. That's great. Well, what about you? What's the most exciting? You know, we are more consumer facing. So maybe next month you will see one of our portfolio company launching an air glass that can record all the 10 hours your day and also help you to record voice, audio, and also understand the contacts, help you summarize and help you track your wellbeing by understanding what your scene and or your audience is saying. Although it's very privacy sensitive, but actually it's really going to. Yeah, okay. This is like going to record your life and help you memorize everything. So that's the glass. And another example is we see a company that beauty. So wherever you have an idea, want to be something you just typing whatever you want to build and they will generate a 3D model for you and ship it to you in a physical way. And so that's a very, very personalized ecommerce platform that it can ship everything you design from it. And lastly, I think the personalization will be huge in the future in I think for Tom Kenny, we are seeing the case that they will generates cartoon and also video by just scrolling a screen and it's not tick tock like one one minute cruise something and millions look at it. It's like everything. Everybody's looking at everything differently just by jumping to the the model and generates the content that you're looking at in a very instant way when the cost of inferencing is really going down and everybody will see different cartoon, different image, different news and different videos on the same app generated by. Yeah, that's amazing. So from personalization to air and space, we have lots to to, to look forward to. Thank you so much for your time. Thank you.