From Lab to Market: The VC Perspective on Deep Tech | By Howard Morgan, Chairman of B Capital

Thank you. Thank you all. And it is a pleasure to me to be here in Singapore with you today. B Capital, which I chair, also has a headquarters in Singapore. My co-founding partner, Eduardo Saverin, is based here in Singapore and we have a very large group of people including today eight people who focus completely on using AI to help us in the venture capital world, as well as our investment partners and investment people for Singapore, for India and for Asia. And then of course we have a very large operation in the US with our other co-founder Raj Ganguly based in Los Angeles. I'm based in New York, San Francisco. My partner Daisy Cai is here who's based in San Francisco. So we are a global, global venture firm. But I want to talk a little bit about the journey and about the whole methodology that we use in evaluating ventures and how we got to being very interested in what's going on today in the deep tech world. So yes, as Gary mentioned, I taught at Cornell and then at Caltech and then went to the University of Pennsylvania in 1972 and in '74 we got machine number 50 on the ARPANET which is the precursor to the internet and it became clear to me at that point that the world was going to change in a very dramatic way and I was able to write papers and lead people into thinking about what we then called the office of the future. So throughout the '70s I was preaching around the world that the world is going to change. We're going to use email. We're going to communicate and in fact you'll use your own email. In fact, one day I wrote about the fact that managers and executives were going to type their own emails and have fewer secretaries. And I was picketed by the secretaries association at the university. But I got it right because as far as I know, almost everybody today either writes their own emails or speaks them because we've worked a lot in that whole area of text to speech. In 1982, after helping start a company that was funded by Jim Simons, Jim asked me and we started Renaissance Technology. Renaissance is legendary as a quant investor. It is pretty much still one of the greatest quant investors ever. Their internal fund called Medallion, which was external until about 2004, has continued to generate gross returns of 100% a year, net returns after fees of about 38% a year. And what happened was at Renaissance in the first few years we also had venture capital and the venture capital did better than the quant because the quant wasn't up to those levels yet so we were generating 25% returns with companies like Silink which had the equivalent of RSA for security. We encrypted the Swift network with the Diffie-Hellman stuff. It's still going today, although it's part of another company. And once the venture became less profitable than Quant, we took the venture out of Renaissance and left Renaissance as pure quant. And then I started doing venture from the same office with my money and Jim's money and investing in a whole bunch of different things including a company called Idealab. And Idealab was the first incubator. And Idealab has incubated several hundred companies. It's still going strong today. One of the brilliant inventions of its founder Bill Gross was paid search. We created a company that became Overture, still used as part of Microsoft's Bing. And we tried to buy Google when it came out. And unfortunately the venture capitalist behind Google, John Doerr, wouldn't let them sell to us even though we offered them an enormous amount of money at the time, two and a half billion dollars. And they were one-tenth our sales. But we knew that they were a great company. So we couldn't buy them. Eventually we sold them Google Photos, what is today Google Photos, called Picasa for many years and had an interesting journey in doing things and then starting to do things in the energy space in early 2000s in the solar space and trying to get that wave. And my blog when I was writing it was called waytooearly.com because that's how I've made my money investing. I invest either too early or way too early. And if you invest way too early, you have to have patience, real patience, but the returns can be enormous. So, just as an example, First Round Capital Fund Two invested in 70 companies and ended up with 68 of them returning five times the fund, which is a very good venture fund. And the 69th was Uber. It started in our office and it ended up returning 22 times the fund and it was about an entrepreneur that we had known and we said whatever you're going to do next. We had told this previous company to eBay just come in we'll give you a quarter million dollars and we'll give you a desk and you start what you want to start and he said I'm ready and we said here's your desk. What are you starting? He said well I'm calling it Uber Cab. We're going to call black cars in San Francisco and make it easier for people because now we have the iPhone and we have maps and we can show the people where the car is that it's actually getting closer to them because one of the problems in San Francisco in particular was you'd call a taxi and 10 minutes later they wouldn't show up and you wouldn't see them and then you'd call back again. They say, 'Oh, they're on their way.' And maybe they were, maybe they weren't. And the brilliance that Garrett had was to say, 'If I can actually give you the comfort to show you, you'll wait 10 minutes if you see he's actually making his way to you.' And of course, that became a phenomenon and returned a huge amount to us. We invested at $4 million valuation. We got out at 60 billion. But that wasn't the best investment in the fund. The best investment in the fund was one that we turned down when my partner Josh and I looked at it. It was a gaming company. And the partner who brought it to us, Chris, said about a month and a half later, I'm bringing my eight-year-old son, Max, to pitch you on Roblox. And so Max pitched us on Roblox. We did Roblox. It returned even more than Uber did. So, the area of sort of high stakes, long, and both of those took 12 years to pay off. So it's not a quick payoff in the venture space but that's what we've been looking at and I've been focusing on. I try to keep my eyes on the far horizon sort of 10 years in the future and wave one was the internet and I was fortunate to be a researcher doing research on the internet as Gary mentioned we were doing optimizations in the network we were also working on very large database systems. Now remember a very large database in 1977 when I chaired the very large database conference was something far smaller than what you have in your phone today. But being able to search very large databases became extremely important. And then wave two which we began to see in the '90s as machine learning and it continued to grow. And then finally the breakthrough in 2012 with AlexNet in image recognition was AI really AI taking over and between 2011-2012. So let me go back for a second in 1970 I published the first paper on spelling correction because at Cornell we were very big on making sure that if you were writing a computer program in those days you put it on punch cards. You went to the computer center, you gave them your punch cards, and three or four hours later, you'd get back your output, and if there were problems or mistakes, then you had to go repunch the cards and submit it again. And so, the turnaround was very long. And Dick Conway, one of my partners and professor advisers said, well, you know, computers shouldn't be stupid. If I type RP, the computer should know just like any human knowing that I meant to type print. So why don't we just correct it and keep going and we built error correcting compilers and spelling correction was done in the way that I had laid it out for many decades. Then we got to Google and Google figured out there's a much better way to do spelling correction because they had millions and millions of examples of people typing something in and then three steps later figuring out the right way to type it and getting what they wanted. And Google took all that data and turned it into spelling correction. And that's the way spelling correction is done today. It's much more a statistical thing than an algorithmic thing. And so machine learning was moving fast. The same thing with image recognition where AlexNet and convolutional neural networks figured out what's a cat from looking at millions of images of cats. Now it's interesting because our two-year-olds figure it out without millions of images. They figure it out with a lot fewer and there's still a lot of learning left. And then in 2017 Google put out the transformers paper and that led to four years later the whole generative AI boom which is a gigantic boom. So we've done a lot of investing in the artificial intelligence space but the next frontier is where atoms are involved. And Nicholas Negroponte, an MIT professor in the '70s and '80s and '90s said back then, 'Look, what we're doing with atoms, we're going to do with bits. And what we're going to do with bits, we're going to do with atoms.' He said, 'What we're doing with wires, is going to become wireless.' And that all happened pretty well. So, okay. So that we've seen that kind of transformation happen and that's what has happened to my own thinking about where to invest. It was usually let's invest in capital light software companies that don't need a lot of capital but starting as we were with Idealab in the energy space which did take more capital and then moving into space where I invested at First Round in Planet Labs in satellites we saw that you sometimes do need capital investment but you can do it in areas that are going to have very high returns so infrastructure, rockets, quantum computing, all those things become important to this whole area. Now, having worked at universities for so long and understanding how they work, we have to think through the other piece of it, which is that translating technology insights and science insights into companies is not easy. Great science does not automatically give you a venture company. And if you're an academic and I was one for 14 years, you get rewarded for publications. You get rewarded for novelty. But as a VC, what we want to reward is something that's scalable, something that can be commercially executed and really move you further along. So what you need for a company to take a scientific discovery and turn it into a very successful company is a pain point. You really need, technology push doesn't work. That is to say here's a great technology don't you want to use it and people don't understand the technology usually but if you show them well here's a pain point you have, you got a pain point of poor internet connectivity around the world in rural areas, okay, now Starlink can come in and solve that problem and become a very highly fundable company or gee getting taxis in San Francisco is a very difficult problem and now Uber can come and solve that. But just saying, 'Oh, I have GPS and I have maps and now how do I push that in great products?' doesn't work. You need to see how can I get the market to pull. How can I find an urgent problem that customers have because if you do that, you'll be able to pull customers in, get customers to want to buy what you've got as particularly if it's an innovative solution. It's something they haven't seen before, but they see the application. They see how it goes. So when you're thinking about the technologies that you're creating and you're thinking about how do I turn this into a company, think about first what is the pain point? What is it that I'm going to solve? You know the old necessity is the mother of invention. And so what's the necessity that's not there today that I need as an invention? And some of that happened in the SaaS revolution. So Salesforce recognized that people needed to keep good lists of their customers. They needed to, and by the way, we'll make these slides available to people later on so you don't have to take pictures of everyone. But it did try to solve that pain point of I have a salesman. I have six salesmen. Each of them are having customer contacts. I'd like to know about that. And here comes a company that says here's how we can do that easily now that we have online access to things now that we have very easy internet things it solves that point but that was pure software. Now when we get into deep tech we have a much longer development cycle. I'll talk about some of our companies in a little bit but a company like Precision Neuroscience which is putting basically the thin film array on top of your brain underneath your scalp and getting it through the FDA approvals and the government approvals and getting it tested and working. That is a much longer cycle time. That's five years, eight years. It's not, you know, gee, I can build this by vibe coding it this weekend. You have real and you have to bend metal as they say. In that case, we're creating things that will have to fit inside your scalp, inside your brain, and inside other parts of your body to hold electronics and to hold the batteries. Much longer sales cycle. Building the things that we're doing at Starcatcher in the satellite world means we're putting together satellites. Now, putting together satellites in 2000 was a hundred million dollar proposition. The reason we invested in Planet Labs in 2008 was that they figured out a way to build a satellite for a million dollars and it was a 100 to 1 difference. And the reason they said, 'Look, if we put up our satellites and we only want them to stay up for a year and they don't have humans in them, we can use all sorts of commercial off-the-shelf products in them. Cameras, the cell phone cameras, all sorts of things that have a consumer use. So they're cheap and we can put those into a satellite and make that work in a much better way. So the capital intensity is there but it doesn't have to be insane capital intensity. So a company like Starcatcher we can probably get it done for 50 to 100 million. Now in my earlier days in venture that was an enormous number. In today's venture world where we see companies routinely raising a billion dollars, it's a little different. The other thing about deep tech is there are regulatory challenges. In the case of Precision, you got the regulatory challenges of FDA approval. In the case of space, you have all sorts of regulations you have to deal with, plus all sorts of authorizations you need to be able to put something in space. Make sure everyone else in space knows where it is and where it's going and is it going to fit and make sure that because you don't really get to try space, putting something in space is not the same as a software program. You can't move fast and break things. It's just too expensive. There's a very good blog post recently about how NASA codes versus how Facebook codes when the coders at NASA are coding for no failures for zero. And that means that they test everything that they check every input. They make sure when a function gets an input, the first thing the function does is to see if the inputs are in range. At Facebook, they never do that. They just move very quickly. So you really have larger capital, but the offset of that is you've got dramatically larger opportunities. You've got really big opportunities. So what do I think about as a VC? What's the real lens that I use in looking at things? And I call it the six Ps. Now it is traditional for venture capitalists to tell you that the most important thing is the entrepreneur, but that's a lie. The entrepreneurs are very important but the most important thing is a giant market. If you have a really giant market and a great entrepreneur then you can have a big win. If you have a really niche market no matter how good the entrepreneur it may not be a venture scale company. If the market for this company is $20 million a year and you can build a product for it that will make you 15 million of profit. That's a great company for you to have. But it's not a great company for me to back as a venture capitalist. I need something that has really large markets. And the other thing is if you start with a really large market and the first product you produce isn't exactly right, you can pivot and tweak it around because the market is still big enough. That said, I have to be traditional. So in the six Ps, the first one is people. It is the entrepreneur and I want to make sure the entrepreneur has intelligence and integrity. Those are the two most critical things. Intelligence because life's too short to work with stupid people. It's just not fun. So I like to work with intelligent people and I know people at universities like NUS typically fall into that category. And then you need integrity. I do want people to tell me the bad news. I do want them to be honest. And we've had situations in our investing at First Round and at B Capital where we've had frauds. We've had companies where the founders literally have created frauds and they're not telling us the truth and they're afraid of that. So they try to cover it up and it's really a problem. So number one is the people. Number two is the product-market. You know how big is it? Now in the early stage whatever product you show me on day one is probably not exactly the product that you're going to be shipping a year from now. So that's why I want the market to be big enough. But I also want you to be customer centric. You have to listen to your customers. You have to know what is it that the customer wants and how are they reacting to what you think. You can put out something you think is great and the customers may want to tweak it in some way and you have to listen to that. The third thing is plans. Can you build a spreadsheet and model of what you're doing? And that allows for budget planning, allows for planning of building, particularly in the deep tech world, building the hardware. Is your hardware scheduling done with the right amount of cushion in it for slippages. Things go wrong in all kinds of ways. Supply chain problems that you have. You know, if you have something that involves a China supply chain and you are a US defense tech company, you have problems today because the US government is concerned about that piece of it. Hopefully, relations are a little better. We are getting magnets now. We are getting motors. But planning is important. Now, if I could take every financial plan that was shown to me and turn it into reality, I'd be much much wealthier. But normally what we do is we take your financial plan, we move the revenues out by six months and we keep the expenses the same because we know that people are optimistic. You can't be an entrepreneur without being optimistic. And the problem is not being too optimistic. And training our entrepreneurs to under-promise and over-deliver is very hard. You've got to train them to rein back in their optimism, build a budget based on a pessimistic forecast, and then make their optimistic thinking work and then everybody's happy. If you go the other way, everybody's unhappy. The fourth P is profits because that's what we're in business for. We're in business to make profits. That's what investors pay for and it is a crucial piece of it. So that means what are the unit economics of what I'm building? And I will say that Josh Kopelman at First Round and Eduardo Saverin at B Capital are both fanatically focused on the unit economics of whatever product we're looking at. How much margin is there at a unit level of each one? If you can make a little bit on each particular widget that you're selling, then and multiply it by a very big number, then you can cover the other costs, the indirect costs. But if you're not making money on the individual widgets, you're never going to make money. So the unit cost economics is very critical and that leads you to what get profits. But then there are two more Ps that are critical for me in assessing entrepreneurship. One is what I call passion. And passion is basically can you show me that you're so passionate about this idea that you're going to run through walls to make it happen that also can you show me why I need to love it and that is something that entrepreneurs have to do because to be a great entrepreneur is to be a great storyteller. You have to know how do I tell this story? How do I make this story work? Not just to tell it to the customers, not just to tell it to the VCs, but to tell it to the people I'm trying to hire, you know, and that is the real superpower of most great entrepreneurs because all you have to give out in many cases is worthless stock or at least worthless at the time you give it out. Hopefully later on it's worth a lot and you have to get the best entrepreneurs to join you. And we've seen recent evidence of this. We funded a company recently called Axiom Math which released its first product today which wants to prove theorems using language models. And Karina Holland, the woman who started it out at Stanford, was able to attract people from Meta and from OpenAI and from Anthropic, people who were getting far more, three, four, five times as much salary as she could offer them because she showed them a mission that she had passion about and that she could instill that passion in them. She also instilled it in myself and Yann LeCun and the other people on our team so that we would do the deal. And now after releasing their first product today, they're going to start to do that in customers. So passion is really critical and we can only judge that with kind of in-person meetings. That's something that we're not ready to let AI judge for us. Although AI can do a lot of the other stuff. And the final one is persistence, which is to say along the journey, along the entrepreneurial journey, you're going to have speed bumps. Things are not going to go right all the time. And the question is, you know, you'll be working 24/7. Your spouse will tell you, why are you doing this? This is crazy. You're making half of what you used to make. Do you have the persistence to overcome that to stick with it and to have some people call it grit, but that is the final critical element for what we're looking at. So that's the key for founder DNA, the ability to tell that story, the ability to really recruit beyond the lab. When you have technical founders, you know, having them able to recruit great marketing people, great salespeople, which is not an area that they've necessarily known, it becomes very important. When you look at a company like Facebook, the real success was bringing in Sheryl Sandberg to run the marketing and sales and all that operation because as smart as Sergey and Larry are, that wasn't an area that they had expertise in. And the problem is that very smart founders think, well, I can learn anything. And they probably can if they have five years, but if you're in a competitive space, you don't have five years. So bring in someone who's done it before, but that means having the ability to bring them in. And that again is part of that passion that we spoke about. Then having operational discipline. Understand that in a startup cash is king. You do not want to run out of money. When I talk to a founder and I say, 'How much do we have left? How many months of runway do we have left?' If they can't give me the answer off the top of their heads, I'm really worried. They ought to know that we've got, you know, five million in the bank and we're burning 400,000 a month and that means 12 months. And if they say, I'll ask my CFO, I get really worried. But you need to have that operational discipline of knowing how you're spending, what you're spending, and where you're spending and when you're going to run out so you know when you have to raise and that becomes very critical. Bad news first. When I was first starting with Jim Simons and we would look at companies and then I would be on the board and I'd come back from a board meeting, he'd say, 'I know it's a wonderful company. What's the bad news? I want the bad news first. Then we'll get the good news, but the bad news is something that we can work on to figure out how to fix. Good news, okay, we can smile about it, but we need to know how we can help the company.' So make sure you find founders who will give you the bad news and give it to you early so you can fix problems before they snowball and become too big for anyone. Okay, so now let's talk about deep tech a little more. What makes it fundable to us and to me is number one an order of magnitude improvement over the alternatives. If you're just 2x better that's not good enough. You really have to have something seriously better in what you're doing. So for example, Precision Neuro is spun out of people who left Neuralink because they didn't like sticking needles in people's brains and they felt that by doing it the way they were doing it, they could put 10 times the number of electrical pads, so they could sense a lot more detail about what's going on in the brain than Neuralink could. And they are doing that. You'd like defensible IP. It's not 100% because sometimes there's some areas it's very hard to patent. Some areas, particularly in the defense tech, governments may step in and classify the patents and do funny things. But you do want to have something that's differentiated enough from the competitors that they can't immediately rush in and do the same thing. You want a roadmap for a defense tech company and for a deep tech company. What are we going to be able to do in year one and year two and year three, what are our milestones so that you can fund based on those milestones if you meet this. So with our Starcatcher we said okay you want to beam power from space from orbiting satellites to satellites below them what makes you think you can beam laser power for a couple hundred kilometers. What's the milestone? So the milestone in that case was we beamed power on the earth over a kilometer and over a kilowatt which was a new record and we beamed it at a moving target that was moving at 18,000 miles an hour because that's one of the issues you have in space is that things are moving at 18,000 miles an hour. Now your relative speed difference is not that high because you're also moving in your own satellite but depending on what altitude you're at you're moving at slightly different rates. So what are the milestones and how can we when we evaluate those milestones we say okay now you're ready for the next round of funding if not maybe it's time to hang it up and then what is the pathway to commercialization so we require you to talk to customers, potential customers, they can be in government, they can be in defense, and try to get those customers to give you POCs, proof of concept contracts, try to give you contingent orders, we'll buy this if. So Starcatcher has $50 million worth of power purchase agreements. So if they can beam the power onto these satellites and hopefully that will be shown in our space test next spring, they already have customers for it. So those are what we need to have happen. So let me give you a few examples. About, I guess it's almost 24 months now, we saw a humanoid robotics company. So there are a couple of humanoid robotics companies out there. And we had looked at several of them. There's one that's inside of Tesla, so you can't look at that one. They're building the Tesla robot, the Optimus robot. So we were shown this company in Austin, Texas, called Apptronik. And Apptronik had been around for eight years, mostly doing work for the defense department, for NASA, building robots for them. They were also contracted by Tesla to help build the first Optimus robot. And we really liked what they were doing and so I was able to invest at a $250 million valuation. And their milestones were then trying to get a customer to work with them on the machine learning and the world model as we call it in robotics, the world model. And sure enough, three months after we invested, Google DeepMind decided to become their partner and to put $150 million in. And that led to us now raising over the last year a billion dollars, $950 million to move forward with building hundreds of robots. We have a manufacturing facility now in Mexico where we're building a robot a day and we're training on dozens of robots now every day teaching them to do tasks and getting them to become smarter and smarter so that we can have them learn their own tasks and that's what's going to happen in this humanoid robots will be a trillion dollar business. Elon has said that, he's right. There will be a number of players just like the auto industry has a number of players. Humanoids will have a number of players. China has some very strong competitors. Unitree, you've all seen the dancing robots and so on. They're particularly great in locomotion. Not as good in some of the other things in robotics. But the US companies particularly those that have any connection to defense are not going to be able to use Chinese robots. There's just too much paranoia and fear. So this market will be just like the auto industry will be separated into probably half a dozen major players to split up that initially one trillion and then probably even more and we will see robots in our homes in the next 10 years. I don't think they'll be in the homes next year. But they will be in factories. They will be in stores in logistics and data centers. Right now, we have Mercedes as a client who wants to put it on their factory floors. Figure AI has BMW as a client. The auto manufacturers want robots. And not that they haven't been using robots for decades, but they've been using robots to do one job, one task. What this does is replace the type of worker who could move from task to task, and the robot can learn a new task very quickly. So, it's one that we're very excited about. And as I say, we went from the $250 million valuation when I started to a $1.4 billion valuation last spring, I guess, and then to a $5 billion valuation in last month and only growing. Then we were brought a company that's doing unmanned surface vehicles. It's called Havoc AI. There are two or three companies doing them. The biggest and most well-known is a company called Saronic. And the difference is two differences. One, Saronic controls a boat, an unmanned boat and sells it to the Navy and the Army. By the way, the US Army has more boats than the Navy. Sells it to the US Army for $2.3 million. Havoc sells each of its 10-meter boats for $100,000. But Havoc's secret sauce is in the AI to run a swarm of them. So we did a demonstration for a bunch of people showing where we had 10 boats off of the coast of Portugal. We had 10 boats off of the coast of San Diego, which is where we were, and 10 off of the coast of Hawaii. And we could say to the swarm, okay, take this piece of ocean, figure out, divide it up amongst yourselves and search for this profile. So each boat has on it solar panels, a Starlink receiver so that it's not subject, so it has a mesh network and whatever you want to put on it, which includes cameras and other things, not weapons, that is we can't sell them with weapons. They can put weapons on later if they want. And then say, okay, the admiral I was standing next to said, you know, I'd like a thousand of these circling Taiwan so that I know everything that's going on around Taiwan at every minute. And a thousand of these at $100,000 we can afford. A thousand Saronics at 2 million, we can't. So that's the way these things are working. And Havoc's again one of those that we invested at a hundred million and it did a $250 million round and now it's doing a $700 million round. So, it's making real progress. It's got lots of boats in the water. Not just 10-meter boats there because they're fundamentally a swarm software company. They've now put that swarm software on drones and on surface vehicles on manned cars, unmanned jeeps. Really, space really excited me. We met Andrew Rush. He had already taken a space company public, a company called Made In Space and joined another space company, public company, Redwire. And he said he realized that one of the problems in space is that when you launch a satellite, typically it has two big solar panels for its power. And all too often meteorites, micrometeorites or debris will hit one of them and knock it out. And now you have half the power and that means you have to change your mission or you have to de-orbit the satellite which is extremely expensive. He said but you know those solar panels they can take more than one sun's worth of light and generate more electricity. So if I can shine a laser tuned to the right frequency for that solar panel I can give it two suns, three suns, five suns even. So, I'm going to put a satellite at 1500 kilometers and beam power using lasers and mirrors, smoke and mirrors as we say, to the satellites below us. And if I can do that, I'm the power company in space, Pacific Gas and Electric or Con Edison or whatever the Singapore Power Company is. And we had a lot of interest from commercial companies from DoD and others because if you can beam and beaming power in space is nice because you don't go through the atmosphere so you don't have huge losses of energy when you try to beam power back to Earth and people are doing this and we'll see more of that. You have huge efficiency losses. This doesn't. So we now have a large number of power purchase agreements. We have a billion dollars worth of POCs of people saying they will buy it if we get it built. Quantum computing, I've kind of been a believer in quantum computing and we met the entrepreneur in 2016, Nick Fina, and he was working with a Michigan State University professor and a Princeton professor who had a different way of looking at this. And when you look at what IBM and what Google and the things you see mostly you see IBM started out well we got two qubits working now we got 16 qubits now we've got 100 qubits. IBM is now up to 1132 qubits. But that's not what you need for practical quantum computing. If you want to break Bitcoin you need about a million working qubits and that means you have to design for scale from day one and that's what EeroQ was doing. Now in 2016 it was way too early and I said come back when you got some further and in 2019 he came back and I said yeah still too early but in 2022 finally, which was six years later, I felt he had shown persistence, he was still doing it.

With the people from the university, it now had real promise, and so we funded it then. It's grown dramatically, and we've taped out a 10,000 qubit chip. Sometime in late '27, I think we'll be demonstrating somewhere between a few hundred and 10,000 qubits, which will be a gigantic value boost. So the capital has gone in milestones. First, show us that you can design the chip. Show us that it's electrons on helium. You build a regular CMOS chip with channels, you put liquid helium in it, you float electrons above them, you entangle them, and you get quantum effects. They've been showing us each stage of the way. Now we have a spin qubit. Now we have two qubits to entangle. And we'll do all that. And then Precision Neuro. So there's the brain, and there's their thin-film array. And Neuralink has now been in 13 patients. With what Neuralink has done, we've been in over 70 patients at 13 major hospitals. We got FDA approval so that we can now leave it in when we're doing work at the hospital. Usually when people have a brain tumor and they're taking out the brain tumor, we can do a lot of testing. We show that we can get people to use their brains to move a cursor, to click on things. We have a video that shows you, you may have seen the video Neuralink did a few years ago of an ape playing pong. It was a very popular video. The ape was playing pong and then they unplugged the joystick, but the ape continued to play with the joystick. Then they took the joystick away and it continued to play. We've now done that with humans. Humans are at least as good as apes at playing pong. And so, we've made great strides. We've mapped out all of the fingers and how you move your fingers. We've mapped out speech so that we can decode what you're thinking of saying. If for example you have a neurodegenerative disease like ALS where you can't speak and you can't control your muscles but your brain can still think about what it wants to do. Very exciting and continuing to grow. So what do they share? They share a deep scientific insight. They share a clear market need which is something we all want. Discipline. A discipline about how to spend capital and how to raise capital and founders who are willing and able to commercialize. And that last one very often, the university professors who were doing IROQ were doing interesting experiments in their labs but one of them, Johannes from Michigan State, was willing to take the leave of absence, get the patent license from MSU and dive into it full-time. He was willing to make the leap towards commercializing. Nick was more the business person there. That's what these companies are that are successful there. So if you're a founder, don't lead with your technological breakthrough. Lead with the problem. Lead with urgency. What's the urgency? What's the demand? So you're bridging the research excellence that brought you something new with the venture execution and that's the way you get to build a successful company in this space. Thank you.

Thank you. Well, clones, there are four clones of me, you don't have that, that's the easy way to do it. Look, I have a passion for life, I have a passion for learning, and I have had for the last decade a passion for longevity. So I do things to try to keep my mind and body sharp, proper exercise, proper sleep, proper diet, proper supplements and things that I need. So that part is something I've just always been and most of that is genetic. I'm very lucky. I got good genes from my father and mother. You know, the best way to figure out if you're going to live longer is to figure out how long your parents lived. And you know, unfortunately you can't change that today. In about 20 years with some gene therapy you might be able to change that but right now you can't. So but your other question I think is an important one I think Gary will have much more to say about than I do which is, you know, Singapore is uniquely situated and it is in a part of the world where they have important neighbors in China and Japan and India and Indonesia and defense tech is a little different, has a different meaning here I think because Singapore started from day one, President Lee Kuan Yew knew he needed to be able to keep people from coming over that bridge and so how are we going to build, do that, and that was the model of universal service and having an aware population that was part of it. And I think defense tech is a continuing part of that as well but it can also then be as it has been for Israel a major economic driver. Israel has provided a lot of defense technologies to many countries in the world including Singapore and that's been a huge driver and it's kept their industry going and they need it because they're almost constantly at war. They have been but I think Singapore which had been fortunate not to have been at war also use it that way.

I think it's a very good question. The academic researcher can work on any problem that's of interest, right? And when they find a problem that they think is also of interest to a business or a consumer and they want to become an entrepreneur, then what I encourage them to do is focus. The biggest issue for an academic is, well, I've taken this idea so far but I see 14 other ways I can use it. No, if you're building a company you have to stay focused for the first two, three on the one way and that becomes very important for us. And you also can't be afraid to fail. That's something we didn't talk about. But the reason the US venture world has been so successful, the US entrepreneurial world, is because we allow people to fail and then start something new. And for many decades, Europe was far behind because it couldn't do that. If you failed in Europe, you didn't get another chance. Similarly in Japan, that's changing in Japan. My partner Karan is also here, spent a lot of time recently in Japan and we're looking at, we've done some venture there but it means encouraging the entrepreneurs to focus, encouraging professors who want to be entrepreneurs to really focus on their technology or if they don't want to focus then find somebody, one of your students presumably, to go and take that thing forward.

Well, you know, obviously the zero to one is the raw founder. When you look at the founders who have stayed in control of trillion dollar companies like Bill Gates, the reason he did was that very early on he brought in Mike Maples from IBM. He brought in a seasoned executive to help him. I mentioned Sheryl Sandberg coming into Facebook. Eric Schmidt came into Google. And Michael Dell brought in Glenn Henry from IBM. So it means that your idea, get the idea right, do that zero to one or maybe 10 million and then as you really want to scale then you have to put in process. So you're an entrepreneur and you have 20 people and you hired every one of them. And then you walk in one day you've got 50 people and you look at somebody and you say who hired her? what is she doing? And then you get to a hundred people and it's like I don't know these people. Who are these people? I know they're helping me but so that's the mindset change when you realize you have to delegate and that delegation is very hard for some founders who are used to micromanaging and because they want to know everything and it's very difficult. That's the key change. But if you then bring in, we used to call it bringing in an adult in the valley. But if you bring in the adult, they can help you through that and particularly if you bring in one who is clearly not going to try to take your job. We want to leave you as founder. We want to leave you as CEO, but we want you to be educated as to what those duties are and what other things you have to do other than just coding every day or doing the engineering design every day. So I think that's the key which is be open to bringing in experienced help and that otherwise as in many companies you get thrown out, you know, company starts to grow, the board sees that you're really not able to grow with it so they make you chief strategy officer or CTO or something or they throw you out completely but hopefully not but some do. But you do need to have that ability to bring in the right kind of help at the right time, not in that first year when you really need to go full blast on everything the way you're doing it.

No, I understand. So my PhD was in operations and but my undergraduate degree was in physics so I have a lot of background in basic science and in physics. And operations research is really modeling and mathematical modeling and decision-making and that's something that we use all the time in venture capital. We are always looking at the models of the company, models of the market. As I mentioned, we have a big AI team now at B Capital which is helping us to build models and to analyze companies. So I use all of that stuff that I learned there but I use the physics to talk to people on space technology or on quantum computing technology. And I'm a lifelong learner. So even though my PhD was in those areas, computer science and operations research, I continue to read very broadly in science. Science magazine comes out every week and I've gotten it every week since 1962. Okay? And I read through it every week. So I learn about genomics. I'm on the board of Cold Spring Harbor Laboratories, which is a major genomic research center in microbiology. So I keep trying to refresh myself and stay ahead. I don't want to be doing what I did in the 60s and 70s and 80s. I want to start doing what I'm going to do in the 2030s if I live there. So, yeah.

Well, there's a lot of history from Israel in building global companies from a small platform and the main thing is to send over an advanced force of one or two people who can tell you whether the US market is ready for what you're building if you're going to the US. But to build a solid base in the local markets in particular here Singapore and maybe Indonesia or maybe India and show that you can expand and bring your technologies there. And that's something that I think can make it work really well if you have strong technology in your local place, whether it's Singapore or Israel or Paris or Munich. But then you build marketing and sales forces that know the local markets that you're going into. And don't expand to, the biggest mistake is expanding too fast, is trying to get to 20 countries all at once. Now there are cases where you want to do that if it's a consumer play. Uber, one of the things that not the founder Garrett Camp, but Travis Kalanick who came in did was to get us into 300 cities in three years. That was an insane expansion. He figured out a process to do it. We created basically a guidebook which said here's how you start a city. Here's what you do. Eventually what you do with a manufacturing platform is say okay here's how we move into a country, here's what we learned in doing the first two and now we can scale it but you got to do two before you can do 50.

I do think you should pursue something at the start because really what the patents are good for is keeping the big guys off your back. You know, you're not going to license, you're not going to sell them, you're not going to make a lot of money from them, but it means two things. One, it's not so easy for them to just try to compete against you because you can threaten them a little bit. It's too expensive to do the actual lawsuits. So, but the second thing is that if you build up over time a patent portfolio, an IP portfolio, that has value in itself and investors feel a little more comfortable if you have a defense. We want a defensible moat. Patents are one element of a defensible moat, of a better technology, strategic customers, but patents are one. You just can't spend too much on them at the beginning, but you really if you can get one key patent about what you're doing, that's pretty important.

Well, basically it gives us customers to talk to and we, you know, then what we do is we go to those customers and we say you did a POC with Appronic. You did a POC which said, you know, if they get it right you would buy 500 to a thousand robots for a particular application. What makes you think it can handle that application? How committed is your company to doing this? Have you thought through? So basically it gives you a way into customer thinking about a product and about a deep tech product. With Starcatcher we've talked to people in the Air Force, Space Command, we've talked to people in places like Planet Labs, people putting up a lot of satellites. The other thing they worry about is well why doesn't Starlink do this? And the answer is they might but they're going to do it for themselves first and there's 25,000 other satellites we can as customers. So, but that's the real point. Real value of POCs is it shows customers think that what you're doing is valuable and you can dig in as a venture capitalist on to why that is.

Yeah, I mean what I told the B Capital team yesterday is Bill Gross at Idealab gave an interesting talk at DLD in Munich this year and I took it to heart and what he said was last year with AI he would start doing something and he'd say I wonder if I can get some help from AI on this. This year he starts by saying okay AI do this and ask me any questions you have. You start and to future proof yourself as an investment professional, you got to start thinking that way. It's not how can AI help me. It's how can AI do this and how can I help the AI do it because then you're future proofing yourself because you're becoming the driver of the thing. You're not the passenger waiting keeping back and forth. You tell the AI what to do. And that's what the agentic world is getting us into very much so. By the way, as an aside, I'm very pleased to be talking in the Ascent series because the name of my fund, the early stage fund at B Capital is called the Ascent Fund. So we agree on the name at least.

On the first, of course, there are many companies in my anti-portfolio, the ones we turned down. One of the most memorable was Brian Chesky and Nathan coming to see Josh and I and saying here's our idea. And we each went home to our wives. And we said to our wives, would you be willing to rent out our sofa? We can make a hundred dollars each evening. And they both said, no way are we going to allow people. So we did not do Airbnb. We also had invested in the prior company before Twitter called Odeo and they gave us back the money because Apple started to do the same thing and he said I'll come to you with the next company and he came to us with Twitter but it was a $20 million valuation and we had set a $10 million limit and so we said no we love it but it's too expensive. So two years later when Jack Dorsey came to us with Square at a $40 million valuation we said yes we learned from our mistake and that was a 25x for us. But yeah we've had a lot of problems. I don't know that I would change my life. I mean I love my life. How could I, how would I change it? I was always in a hurry, right? I mean I did high school in three years and college in three years and PhD, I had a PhD at 22. I was the youngest professor at Cornell. I would sit down. Most of my students were older than me. My PhD students were all older than me. And I just keep learning. So I don't know that I would want to change that. I don't like the sight of blood. So I don't want to be a medical doctor. Other than that, there's not much else I would have rather been. And the only other change, I stayed in the universities for 14 years, tenured full professor. Well, maybe I would have left a little earlier to do some of this stuff. Other than that, I've been very happy with my life.

Well, you know, we have a special team in the healthcare area headed by Dr. Robert Mittendorf. And we do some mental health things. And there we don't think of them as niche markets. I mean, they would, as you say, with depression in particular, it's huge worldwide. It's gotten worse after COVID, for sure. And but we look for a unique solution. So, you know, looking at the biomarkers for it is something we haven't seen. We've seen a couple of, we have one company called Isaac Health that helps people with PTSD. I have a company in Israel that uses organoids of your nerve cells to test against the various depression medicines. So rather than give you an SSRI or one of the possibilities, what they do is they grow your nerve cells. They test them against the various and because typically somebody, they'll prescribe one for you, you'll be on it three months and it doesn't work. Then they'll prescribe the next one, you'll be on it for three months, it doesn't. So we now try to find in two weeks which one it will work against. So you try to find interesting technological advances that can fit a particular area.