Transcript-AI-Investments

Pan-Pacific Investment Opportunities in AI - owner Canice Wu - Moderator  Jay Eum

 

Panel Members:

Evan Schnidman - Head of Fidelity Labs & Investor

David Gerster - Investment Partner, Fusion Fund

Ryan Wang - General Partner, Solaris Venture Partners & Outpost Capital

David Suh - Head of Korea Investment Partners

Cindy Chen - Brookfield Asset Management

++++++++++++++++++++++++++++++++++++++++++++++++++++

·  If the investor panel folks can step up to the podium, please take your seats.

·  All right, please. Do you want to speak in particular order? Maybe you just follow the, That's not us. I'm Jay, by the way. Kevin, good to meet you. Yeah, we've met, haven't we? We've definitely… Oh, yes, so please take a seat.

·  Hey, David, how are you? Thanks for joining us. Cindy, nice to meet you. Just, maybe you want to match up? Nice to meet you. Right here. Hey, Ryan. Nice to meet you, David. Ryan. Nice to meet you.

·  Yeah, I get that.

·  It's probably, yeah, you guys used to work together, right? Yeah, yeah. Yeah, please, just take a seat, we'll get started. Hi. Nice to meet you. Nice to meet you, please.

·  All right, everybody. Final panel before networking and drinks. Obviously, we've been chatting a lot about AI. I think every panel we've heard so far has addressed the investment issue, and we have actually a great panel of very diverse investors that have been actively investing in the AI stack. And so, what I'd like to do is to kick things off, if each of you could introduce yourselves, your firms, and actually, I'd like to add one more.

·  Because there's so much diversity here, if you could actually include a specific investment that you or your firm has made, just to get a better sense of what you're actually investing in AI would be super helpful. So why don't we get started? Evan, please. Yeah, sure.

·  All right, there we go. Well, first of all, thank you all. I'm happy to be here. So I'm Evan Schneeman. I'm the head of Fidelity Labs. So I'm wearing two different hats here today. One as a personal angel investor who's been fairly prolific in the AI space now for a little over a decade. And then separately, Fidelity Labs is Fidelity Investments' in -house innovation incubation division. So we actually  you can think of us like a venture studio. So our investment allocation is actually more around what are we building internally?

·  to then actually be able to have those either be the future businesses of Fidelity or have the potential to spin out and scale independently. And so right now we've got nine businesses in Fidelity Labs, only five of which I'm allowed to talk about externally. So apologies on that. But to your point, Jay, I think it's interesting that we've layered AI into every business we build.

·  It's just a question of, one, when were they started? Some go all the way back to 2019. And two, when essentially is AI the core product?

·  Or is it just an enhancer of what we're offering? And so, I can look at a compliance technology business like Safer, where we're actually using  the internal language, the internal data we have within Fidelity to train a compliance system that actually automates away a lot of the compliance headaches for our clients. That's actually one of the oldest businesses in labs, and we've put a lot of resources into building that business so that we could actually train older AI models. We've now had to go through the, I think, tough evolution that a lot of businesses have in the last couple of years, which is to retrain those systems to modernize in the current era, and that business is better for it, and the product is significantly better for it.

·  Great. Thank you, Evan.

·  I guess it's not perfectly in sequence, but that's okay. Ryan, please. Yes. Hi, everyone. My name is

·  My name is Ryan Wong. So I have been working as a venture capitalist in Silicon Valley for the past 11 years. Right now, I primarily manage a fund called Solaris Venture Partners, where we focus on AI infrastructure investment and also physical AI.

·  So one of our thesis is really, you know, AI is shifting everything, both in the digital world and the physical world  large language model is very powerful, and it's very capable, can do a lot of things in the digital world, right? In a way that we can already achieve AGI in the digital side of the applications.

·  But for what's AI's impact in the physical world, right? Colleagues can, regardless of robotics, autonomous driving, or, you know, edge AI, or some kind of AI hardware, sensory systems, this is, pretty much an uncharted territory, and such a blue ocean opportunity. And to put that in perspective with what we understand about, you know.

·  AI for the digital world. I think we are still in the pre-transformer days, right? Where people are, you know, nowadays tinkering around different VLA, word action models, and different kind of word models. It's almost similar to, you know, back in the days 10, 12 years ago, when people are looking through different kind of architecture, like CN, RN, long short-term memory, right?

·  So we haven't reached the transformer moment equivalent for the physical AI. And from a timing perspective, if we are able to make the right bet at the right time, then we can still capture another like 100, 1000 growth in that industry. So that's why we are super interested in physical AI. Before that, before Solaris, I managed another fund called Outpost Capital. focused on early-stage frontier technology investment in U.S. and also in Asia. And before that, I was an investment banker in New York City, so a little bit East Coast, West Coast experience as well. Thank you. Specific company you've invested in? Right, so recently, I invested in a company called Atta Robotics. It's a company in the vertical robotics application site. They are developing a system that can convert traditional forklift into autonomous forklift, and through that, they have a management system to do, you know, automatic warehousing and logistics as well.

·  Great, thank you. David S, please.

·  Hi everyone, David here from Korea Investment Partners, in short, KIP. We're not a sovereign wealth fund, but a lot of people confuse between us and many other KI, etc. We are the largest Korean VC platform from Korea, headquartered in Korea, of course. run the operation here, also for European investments.

·  Straight to your question, we invested in Lots of robotics companies, actually, and life sciences. These two are the main focus right now for us, including data center. For robotics, electronic, we're looking at some of the component-level companies that are more focused on, dexterity.

·  I would say, advanced materials and IP companies. Life sciences, we invested in a company called Precision Neuroscience. So basically where compute meets bio life sciences, that's an area that we're heavily focused on right now.

·  Okay.

·  Thank you, David S. And, Sydney, Cindy? Cindy, please.

·  Hi, everyone. I'm Cindy Chun. I help lead Brookfield's global strategy group. For those of you less familiar with Brookfield, we are one of the world's largest global alternative asset managers. We have over a trillion of assets under management. Our investments primarily span across infrastructure, energy, real estate, credit, and private equity.

·  And within the Global Strategy Group, we're really focused on partnering with emerging technology companies that we believe will reshape the global economy over the coming decade.

·  And in terms of investments within the AI category, we've partnered with businesses across the entire AI stack. So a lot of lab businesses, application businesses, energy businesses, robotics as well, anything from inception stage businesses to pre-IPO.

·  Great, thank you, Sydney, and then finally, David G, please.

·  Thanks, Jay.

·  I'm David Gerster. I'm the Enterprise AI Partner at Fusion Fund. We are a seed stage firm headquartered in Palo Alto. We invest across three broad sectors. There's Enterprise AI, which is my main focus. We also invest in healthcare and

·  Deep tech, which includes everything from space tech to semiconductors to robotics.

·  We've been investing in deep tech since our inception 11 years ago. A investment you may have heard of is a company called SpaceX, which we invested in back in 2017.

·  Great. Thank you so much, guys. I realized that I felt so comfortable that I knew everybody here that I forgot my own introduction, so… My name's Jay, I'm one of the co-founding managing directors at GFT Ventures, Global Frontier Technology Ventures. Those that were actually on time today.

·  I heard my partner give his keynote fireside chat with Canice. So I'm the guy that actually pulled my partner, Jeff, out of NVIDIA, which he still, gives me a hard time for. But, It turned out that our thesis around building a frontier tech firm focused around AI and physical AI made a lot of sense. In fact, timing was immaculate in a way since we launched in 2021.

·  And that helped us with our fundraising as we were fundraising, as we were wrapping up fundraising.

·  ChatGPT was announced, and that boosted our fundraising, so we were oversubscribed for our fund.

·  And some of the results that we've seen in these hypergrowth AI companies, we've been able to benefit from. So one example that I'll give, a recent investment in the AI space, is a company called Go.ai. Go.ai is one of the leading companies in on-prem enterprise AI.

·  I think it's become common knowledge that within the enterprise, a lot of customers have now realized that their data, which is supposedly secure.

·  actually may not be. And from a compliance perspective, it's very difficult to prove otherwise. And so, having your own infrastructure, training on your own data, where you manage that infrastructure, is a huge benefit. And secondly, to be able to predict  cost of AI. So CFOs love the company because it's basically your own infrastructure and all you can eat. So instead of worrying about token maxing and managing employee AI usage patterns.

·  You can literally install the, the infrastructure upfront and all you can eat. And so this company has grown over 35 x in the last twelve months and doing really well.

·  So again, we have a broad range of investors here across the table from a corporate innovation lab to a cross-border frontier fund, not a sovereign wealth fund out of Korea, but the largest venture platform.

·  And then a mega infrastructure fund and one of the leading deep tech funds out there. I'd like to start by asking you, there's a broad range or spectrum within the AI stack. to focus on, and I'm just really curious, kind of, among the various areas from semiconductor chips to infrastructure to models to application layer and beyond, what are the areas within the AI stack that excites you most, and why?

·  Oh, boy. Okay, let me… let me preface with, Canice told me I need to explain a little bit about what it is that we do in Fidelity Labs, because

·  Fidelity, I assume everybody here is vaguely familiar with Fidelity Investments. Is that fair? Maybe. Okay. Maybe heard of it. 35 % of all Americans' retirement accounts. So it's a little company. We have 20 trillion in AUA. It's fine. So one of the tricky things about doing this inside of a massive asset manager that also has a huge brokerage business, wealth business.

·  We are not the only arm of the business that is making venture investments, right? So we're tasked with building businesses internally within the existing organization that have the potential to spin out externally or stay internal.

·  So that is an interesting limiting factor for us because there are parts of the business that invest in data centers. We have, in fact, a whole data center investment vehicle. Our late stage venture group has large stakes in the hyperscalers as well as large stakes in all the AI labs.

·  So, from a diversification standpoint, it's our job as investors across Fidelity's universe to make sure we are safeguarding the wealth of all of essentially every American who has a retirement account with us, a wealth account with us, a brokerage account with us. So the first and foremost directive inside of Fidelity is never violate that trust, which means that anything we are working on internally.  within Fidelity, we actually have to be quite careful about making sure that there's no data leakage, that there's no infosec issues, that we're following a very strict legal risk and compliance framework.

·  Now, with that said, within labs, we have way more freedom than anybody else. And that's by definition. our job is to move faster than the rest of the organization. So what I get most excited about is just having a background in the data industry is how can we leverage a novel data pipeline to build something that is totally unique?

·  We're totally differentiated. And so one of the areas that I've looked at very closely in the last few years is what's going on in private markets and how is that marketplace underserved by the existing data infrastructure and how can AI actually be implemented to frankly make it a lot easier for all of us. I'd venture a guess that everybody on this panel has no idea what their net worth is.

·  Anybody.

·  Could you mark to market your whole portfolio right now?

·  Can you? Absolutely not. That's the problem I'm most interested in solving. Okay.

·  Great. Thank you. Ryan, any examples, especially since you focus on physical AI, which area that you know, even within the physical AI stack, gets your interest. Right. Yeah. So previously I mentioned a little bit on why I'm so bullish on physical AI, right? But within the physical AI space, we actually look into pretty much everything from software to hardware to, you know, different world models and VLA, VLM, robot brains, right? Different components, dexterity. hands to actuators, right? More on the supply chain side of things as well. And also we look into different vertical applications. Within all those kind of ranges, I think two things really stand out for us, right? One is the intelligence side.  how to build the next… the future of robot brain. Why? Because, like I mentioned, we are at very, very early days for, you know, AI for the physical world. We haven't reached to the transformer. equivalent for physical AI, where it's an intelligent model that is, you know, powerful enough, and also have enough general purpose capability to be able to cover most of the use cases, most of the skills, and to be able to kind of switch from one task to another, switch from one environment to another, and switch from one set of hardware to another set of hardware, right? So I think there is still a long road ahead to solve the intelligence side of the problem.

·  And that's why, you know, it's very, very interesting. And, you know, we want to, we don't, we don't, we, we, we no longer need all this, you know, stupid robots anymore, right? Other than just doing the dancing and all these crazy backflips and doing kung fu. They're entertaining. Right. It is entertaining, but then, you know, we need to figure out where's the value come from, right? So I think the value clicks when they kick somebody.

·  That's part of the value, but, you know, you can't justify a billion-dollar valuation just with that kind of value, right? So, I think, you know, that's why we keep on looking for the R&D side in the robotic brain and intelligence side of things to say, hey, are there any new architectures? Are there any new format, and are there any new R&D breakthroughs that we can potentially adopt to the industry so that, you know, it can take the robot intelligence to the next level. The second thing I want to highlight is on the vertical applications. Why? Because

·  You know, it's very, very hard to deploy it robotics in the real world.

·  And also, it's even harder to deploy robotics in a very reliable and in a scalable way, right? So I think there's an analogy between what it takes between, you know, software to get adopted in real world versus hardware. So in the software, if you are building a SaaS, we're talking about, hey, if you are able to deliver a POC to a client.

·  Pretty much, you've done 80% of the work.

·  But translate that into physical AI, into robotics. A good-looking POC is only the beginning point, right? It's only the initial 10%.  The remaining percent, you have different kind of edge cases, you have, you know, a thousand ways on how things can break, and also, you need to worry about how you can integrate your solution with their workflow, with their facility, and with their, you know, typical way of doing business as well. So, I think, you know, a POC in the physical AI, in the robotics  kind of scenario, it's only, like, 10-20% of the work, and the majority of the work, the majority of the heavy lifting is in the integration, is in the deployment, is in the go-to-market, right? So that's why we feel like, hey, let's.  you know, there are a lot of money, and there are a lot of excitement in this sector, but we haven't seen large-scale deployment, and I think this year and the next two, three years is a time when we need to see the large-scale deployment start to happen, and that's why we also focus on vertical applications as well.

·  Okay, makes sense, David. Perhaps.

·  I'd love to hear more about, especially on the healthcare medical side, what you're seeing as opportunities up and down the AI stack since you're investing actively in that area.

·  Right. So just FYI, I actually joined KIP from Samsung Ventures. So Samsung Ventures, I was there for 12 years running semiconductor fund. So I'm not an expert in life sciences, yet KIP has traditionally been a very big powerhouse for life sciences.

·  So when we look at this life sciences space from KIP point of view.

·  We, you know, in general, I think we serve as a gateway to Korean partnerships, right? Right now, it's a very booming, you know, timing for us, because you look at semiconductor, life sciences, and energy, these are the three biggest industries that Korea is involved in.

·  And so, as a gateway partner for these partners, we're looking at investing in early stages, because we believe now there's a lot of money, not just life sciences, but semiconductor companies are just stockpiling lots of reserves. And we predict that within the next year or two, there'll be a lot of acquisitions, more so than IPOs. This is our thesis. When we look at our fund strategy, we start from a macro point of view. We believe that valuations, of course, we talk about the bubbles.

·  But, you know, I think there will be, of course, a correction, but, you know, in defense of the industry now. anything that deals with AI is a little bit different than what we saw in dot-com bubble. It's no longer a bandwidth issue that goes into application, right? This is more of, you know, where is the money being created in the stack? So, going back to your question.

·  On the stack, on the AI stack, we believe that, when you talk about physical AI for us, it really means on a chip level.on atom level, on energy level. So, again, going back to the focus on industry, these three areas really represent the focus areas that we're looking at. Specifically for life sciences.

·  We're very much looking at advanced materials, AI for life sciences. So companies such as AI for semiconductors and advanced materials, so CosmAI in Europe, for example, Lila Sciences. I think there will be more of these kind of companies coming out. in that sector. Longevity is a big, topic for us. For Korea, as you guys know, cosmetics industry is one of the hottest trends in Korea, and so, we continue to look at companies that we can also deliver as a manufacturing partnership, because we have SK Biosciences.

·  Samsung as well, very, very much involved in that area.

·  So if you have good IP, we try to bring that to Korea so we can strike a manufacturing partnership. Same thing in semiconductor. Within that, we're looking at, of course, advanced packaging tools. memory hierarchy, orchestration layer, so very much on down to deep tech level, IP level, for acquisition opportunities. I'll stop there. Great, thank you, David. Cindy, I'm personally also very curious, kind of.  massive infrastructure property management company, what are the areas within the AI stack you're focused on? Yeah, well, we think there's a tremendous amount of opportunities across the AI stack, and I would say I probably look at this in two lenses.

·  From a broader Brookfield perspective, of course, we would naturally point to the infrastructure layer. We think that there are, you know, probably over $7 trillion worth of opportunities over the coming decade as it relates to AI infrastructure build-out to continue to support the growth in the category. And so, naturally, there is a ton of opportunities there. It's actually so interesting to me, personally, that a lot of the, call it, constraints and bottleneck to scaling this

·  Technology is no longer purely technological. It really is physical. But within the global strategy group, we also spend a tremendous amount of time with emerging technology companies.

·  And within that, I would particularly point to, I would say, the application layer as being super, super interesting. You know, we've seen so much progress in the intelligence and in the model companies. And I do think that the next chapter is going to be about how do you translate this intelligence into real economic value. And within that, we are very excited about Agentic solutions. We're very excited  about physical AI broadly, just given kind of the sheer breadth of application and use cases and the opportunity to create enterprise value and productivity gain, not just in call it the traditional digital world, but in real physical industries.

·  That makes sense, absolutely. And then, David, please, any… what are the areas that you guys are focused on?

·  So I would certainly echo what Cindy just said about the opportunities in the app layer. If you look at Agentic AI, I mean  you know, in 2025, Agentic was kind of a buzzword that you saw in every PowerPoint deck. In 2026, it actually happened. You know, starting with co-work in January, and now there's been, you know, something like 100x price performance improvement.

·  AI has basically solved math, as far as I can tell, and on and on. And so, yes, actually harnessing that new intelligence to actually do things.

·  I think that that's very powerful. To pick just one example from our portfolio, we invested in a company called Reality Defender that does deep fake detection, real-time deep fake detection for audio and video. If anyone remembers that story  from a Hong Kong bank that wired $25,000,000 to a fraudster based on a completely fake conference call with multiple fake execs participating via Zoom.

·  Right? I think that's a great example of the type of  Vertical application opportunity that's out there.

·  That's great. I'll just add my two cents into this. You know, we're… relatively full stack as an AI investor. We've invested in semiconductors from our backgrounds at Samsung and NVIDIA, all the way up to various infrastructure and application.

·  But it's interesting because while there's plenty of massive opportunities to build out the infrastructure, we believe that the application layer is still relatively early. There's probably maybe a couple dozen  AI vertical applications, horizontal applications that are hitting or getting close to that billion dollar revenue mark.

·  And, if you look at the fact that over time, once the infrastructure is built out, and we can also include LLMs as part of the infrastructure, specific vertical applications or horizontal applications, there are so many opportunities to build customer-specific  applications that are trusted more than a generic application, the same way that each of us would see a specialist for, let's say, a cardiac issue  rather than just see your family doctor and physician. And so we're focused on those areas. We still think that there's multiple billion dollar revenue opportunities there, and that's the area that we're focused on.

I'd like to go back to the panel and ask  Kind of the last question, since we're running tight on time. Would love to get your perspectives on what everybody's talking about as soon as they talk about AI investments, which is valuation.

·  And I think most of us, maybe not Cindy, but have lived through these crazy bubbles from the telecom infrastructure, dot com. mobile, etc, where there's been a hyper run-up, early days of these technology disruptions, and then many of those companies have fallen back down to earth. So we'd love to hear your perspectives on how you guys think specifically about the current valuations out in the market.

·  I'll take this from two perspectives. One, I think Nate alluded to this on the earlier panel.

·  When I think about the world of how the AI models have developed and essentially what we've seen from valuations from the large labs, if you take the analogy of AI is an engine and data is fuel.

·  The very… you can look at Cummins, they're a great engine maker, they're a $50 billion company. The largest fuel provider in the world, largest oil company is Aramco, they're about a $2 trillion company.

·  So we've all been looking at the AI labs as, hey, they're the core engine of growth here.

·  But they're not actually the fuel provider that is going to continue to sustain that growth. And when we've looked at historical cycles, you can go all the way back to the Industrial Revolution, the steam engine. It's not just the dot-com bubble, or the app bubble, or, you know, we can look at the more recent instances.

·  But where the value has been created is not just the raw fuel providers. It's the refiners. It's the pipelines. It's all of the tooling. It's that application layer. where we're going to see many, many, many more multi -billion or multi -trillion dollar companies pop up. And I do think that there, in my personal view, that there's an interesting question of…

·  Why are we valuing engine makers and undervaluing fuel providers and fuel transporters and all of the applications along the way? Now, the other piece of that, from my perspective, is we have two businesses inside of Fidelity Labs that we've incubated both Fidelity Private Shares and Fidelity Stock Transfer, where we live and breathe on working with these private market companies, in Fidelity Private Shares' case, providing cap table management, 409A evaluation, data room document storage, and helping these businesses make sure they're benchmarking correctly  in this process from the earliest stages. And then with Fidelity stock transfer, we have the first fully modern transfer agent in over 30 years. We processed the SpaceX IPO this year. Again, T.

·  Congrats, by the way. And To both of you. But it's, you know, we launched that business commercially about a year ago, and we've benefited enormously from the increased valuations in the market. So I'm not complaining about them by any means, but I think it's an interesting moment in time to look at who's able to go public and who can stand up to the scrutiny of what public markets are actually going to look at. What is going to be in that S1? How are we going to evaluate them as a public company versus how are private companies being valued?

·  So are you buying into the upcoming Anthropic IPO or not?

·  No comment. Fair enough. Ryan?

·  To answer your question, I think I have a very simple answer to this question. The valuation is very high right now, and it's… is so high that it feels a little bit unreasonable, but with a certain level of justification, right? So, I think, you know, on the.on the private side, so, a lot of these companies, they are raising their seed round with hundreds of millions, sometimes even, like, billion dollar valuation. They are really, you know, a lot of future potential has already priced in into the value of the company. And if we look into after they've raised so much money, where are they spending the money? Basically, three things, right? Talents. compute. and also data, right? And all these three things are essentially different kind of essential building blocks that is very, very capital-intensive for them to achieve what they wanted to do. So, although valuation is high. But it's, you know, the path on how we can potentially get there. So that's one way of justification. The other way, I think, is really on looking into the future. I mean, we are all venture capitalists here and we all look into not only what's happening in today's world, but what's going to happen in next three years, five years, eight years down the road, right? One thing we all know AI has increased a lot of productivity and the productivity gain can be a justification for the high valuation. And the other thing I want to mention that I think

·  You know, I haven't heard too much Of this kind of theory is that, Our current understanding about value creation might get so much disrupted 5, 10 years down the road.

·  So, you know, today there are only a handful of, you know, trillion-dollar companies in the world, right? But if with AI. We can get into an era where, super productivity and super growth and super value creation can be, you know, achieved. then maybe 5, 10 years down the road, there… we could live in a world where there are, you know, hundreds of trillion-dollar companies, and a trillion-dollar company 10 years down the road can feel like a billion-dollar company as of today, right? So I think, you know, on that perspective, if we are entering into a hyper-value creation era, then

·  You know, there's another way on how everything can be justified. And we are still at the very, very beginning. And look, maybe 10 years down the road, looking back, everything become very cheap. So Ryan, real quick, you mentioned three factors that they're spending their money on. None of them were distribution.

·  How are they going to actually make sales? I think distribution is one of the most critical ones, right? So, nowadays, especially for software, it's extremely easy for creation. But the distribution is always the most critical and always the most challenging one. That's why, you know, I think for us and for a lot of… I think Cindy also mentioned about real-world adoption, right? We not only need to think about what this technology is capable of, but what's the adoption and what's the, you know, value creation in the real world look like.

·  Please.

·  In terms of distribution, I don't think that's an issue these days, because if you look at SaaS, go-to-market added value to the workflow. These days, you would have headphones, Android platforms, Apple. I think a lot of the content or services are getting used without much of the value added from go-to-market. That's why you see, from year zero, ARR going up to $100 million.

·  So you know, I think, more importantly, when you look at the valuation question.

·  It's really specific to certain industry. I don't think it's applies to life sciences, not at least not yet.

·  I don't think it applies to component level semiconductor companies as you go lower down the stack. So I think it really depends. I think if you're talking about AI application in general, yes.

·  AI frontier models? Holistically, yes. So, it's very industry-specific, but generally speaking. you know, if you look at all the bubbles, I mean, Jay, you've been through a lot more cycles than I have, but…

·  But you look very young, I don't mean that. But, in general, the alpha seekers will always be there, right? I think that, you know, through the dot-com bubbles, you know, living crisis, you know, it's all temporary, you know, boom and bust. In general, I think the value… the venture industry is all about value creation, whether it's from You know, 10x to 15x, or 0 to 3.

·  Right? That's why we, you know, KIP has taken a very strategic approach, because we're adding the early-stage fund, which I've never done before, because I think that's essentially defensible against any cycle anyway. And I think more so important now, because you're aiming for, as I said before.

·  Macro point of view, I think there'll be a lot more M and A plan to detect space after perhaps the cycle is done with the IPO market. So I think we're positioning our fund to actually ride a certain cycle in that perspective.

·  The other thing is that I think, you know, before valuation, when you talk about valuation, you know, energy sector you mentioned before.

·  Of course, we have to reinstate what is the sensible valuation metric. you know, one can argue that the paradigm has shifted forever, right, with this AI. It's a revolutionary platform. And so.

·  You know, to say 10x, 20x your multiple does that really matter anymore? I think it's the expectation evaluation, the growth expectation that is different than before. So if you know, Anthropic obviously is still losing money.

·  Right? But people were making money on it. We're also an investor in Anthropic, SpaceX as well.

·  I think it really depends at the end of the day. you know, if you accept the fact that the whole world will be changing. from now on. It's maybe even SpaceX. It's just the beginning. If they're trying to be an industry leader in that space where no one has ever led before, perhaps it still makes sense to invest in SpaceX.

·  At least Cathie Wood did, right? After IPO.

·  So leave it there.

·  Please, Cindy. Yeah, I think valuation certainly reflects optimism that investors have about what this technology could potentially do, and I think that certain part of that is warranted. Now, again, I want to echo David's point that, valuation

·  I think it differs dramatically by sub-vertical, and by geography as well, and so that's something to kind of caveat. But, overall, if you look at AI, it's… It's fascinating. It has transcended beyond just software and digital knowledge work. We're going into real industrial industries, physical world, right? Manufacturing, logistics, industrial automation. And I think the potential of that is so dramatic, and we're not even starting to see the beginning of that.

·  And so, not to say that I don't think some of this, these robotics companies' valuation are absolutely insane, but maybe on the other side of it, the opportunity, you know, can help justify some of this.

·  Okay, and then David?

·  So, since our inception, Fusion Fund has pursued a multi-sector approach. We have our three sectors that we invest in, and so I find that's pretty helpful when one sector gets kind of out of whack, right? You can stay disciplined on price and not FOMO yourself into a round. And so… I think there are good seed stage opportunities out there, at, you know. what we would consider to be normal seed stage valuations. One that comes to mind is a company called Starpath, You know, we invested in their seed round. Again, it was not at some, you know, crazy nine-figure valuation. They make, space-hardened solar panels.

·  And so you know, the multi-sector approach, and just kind of not being too anchored into any one category, right? I think, again, that helps with pricing discipline. I also think there's some element of uncorrelated return across those sectors.

·  Well, thank you. I'm just gonna share a couple of thoughts on valuation since I get asked that question a fair amount. So there used to be a time where people got super excited that Google hit the billion dollar revenue mark, and it took them about seven years.

·  Now what's going on is some of the fastest growing AI companies are counting months.

·  So the companies that my partner Jeff talked about, Higgsfield AI, one of the leading video gen companies, hit the billion dollar run rate mark in 18 months since they launched. Right? And there's plenty of other companies right behind them that are trying to beat that mark. So, one, we're looking at unprecedented revenue velocity. Right? So that's number one. Number two.

·  There used to be a time where a billion-dollar value and a billion-dollar exit used to be a good deal, right? They used to be called unicorns. Now we're talking about a startup. And let's be clear, Anthropic is still a startup. It's about six years old.

·  They've already hit a trillion dollar value.

·  And if they… and if and when they do go public, there's a debate, are they worth 1 or 2 trillion?

·  So we're talking about multiple orders of magnitude more to expect. And so if we talk about the potential at an earlier stage when these companies are private.

·  Is it completely unreasonable to assume some of these valuations are out of whack?

·  Maybe or maybe not.

·  So we may be living, and I really like what Ryan and David said about we may be actually playing a different game now as investors, and we may need to adapt our game accordingly. So anyway, with that, I'd love to for everybody to thank our panelists. Joining us today and sharing their insights. Thank you so much.

·  I was like, wait, you're not… you just listed three things, and none of them were actually how you sell stuff. I know, man. So… Yeah. Well, you have the most interesting job after that. I'm so jealous.

·  Thank you so much to the panelists, and thank you so much to all of you for joining us this afternoon. Before I pass it on to Canice to wrap it up.

·  There was a discussion a couple of weeks back about how do you slow down AI. I don't know if you guys remember it or not, right? And, I was on a board of a university, and the discussion was, bring it back to university, that will slow it down. So, I guess, that has not happened.

·  Anyway, thank you so much for all of you, and I want to thank all the sponsors. Again, the Hong Kong Trade and Economic Organization, thank you so much. Thank you so much for Cooley for providing such a beautiful  venue. Beautiful view here. We need to move our office here. Sure, yeah. We do that. And again, thank you.

·  Thank you so much to all the organizers and speakers. It was fantastic. I learned a lot, and we'll make sure I will continue looking at the AI or SIs. So thank you so much. Pass it to you.

·  Thank you, Nader. So, thank you for staying all day, appreciate it, and before we all go adjourn and have some, I guess we have some… a little food, a little appetizer, and some wine as well, so continue the networking. I think, first of all, thank you for all the speakers, thank you very, very much.

·  It was a lovely… anyway, anyway, I'm not gonna go into that for her. And also, what I, just a couple of things here before I wrap up. One is, I love to continue the conversation. I think there's a lot of amazing on concepts.challenges, strategy here that we discussed. Like, we need to figure a way in AAMA how to do that digitally, right? But maybe you can start by just commenting on my post on LinkedIn. If we're not connected LinkedIn to me, we can at least maybe start it there, and then we'll figure some way to kind of continue this dialogue, I think would be very helpful. And number two, I'm gonna… supposed, I think, stay connected with AAMA. We're trying to bring quality programming and quality speakers, and the next event is… We're going to have one in January, which is going to be the.

·  With the JP Morgan Health and Life Sciences. Yeah, so we actually have a very strong life science. Actually, talk about our communities. We have a strong life sciences and health community. We actually still have now hot again semiconductor.

·  You know, believe it or not, they were like been along the whole 47 year journey. They were now they're hot again. So it's a very, very active community. Please join that community. No, seriously, very, very interesting people here. Right. And of course, tech and software, etc. But our next event will be the healthcare conference.  time around the JP Morgan conference in January, so please stay, I guess, stay connected to the AAMA community, and, and I'm gonna let… we're gonna keep going back and forth. No, I'm gonna let you wrap up now.

·  So regarding the health care, that's actually the business that I am in. And two things. Number one, our friends at FDA, starting this year, they're using an Agentic AI called Elsa.

·  And that actually has gone, you know, really created a really interesting phenomena, because companies who are building medical devices, let's say 3 years ago, you submitted something which was a hazard, and guess what? AI doesn't forget, right? So, this is no longer humans looking at  submission. This is AI looking at people's submissions, so this is really changing the market. But the second thing I heard, which was really interesting, is longevity, and Dr. Ring, where is she? She was here.

·  And somebody says, you know, the business, longevity business is going down, and I said, why? And they said, because AI is going to kill everybody in the next 6 years anyway, right? So, who… 10 years ago, who needs… so who needs longevity? Anyway, thank you so much, and we're going to go to have some… some appetizers and some drinks. Thank you again.

·  It's a beautiful idea.

·  Quick question. You mentioned about the life science. He's the man to ask.