2024 Future of U.S. and China Conference: The U.S. and China Innovation Race

We're now going to session seven, this is our final session of the day: US-China and the Innovation Race. This panel will be broadcast and live streamed to our audience in San Francisco and to our Northern California members. Our speakers are Ganesh Iyer, Katie Strear, and moderating again is our chair Gary Rieschel. Please join me in welcoming them.

So I'm looking forward to whoever in this audience is the first one to hand me a business card that says 'shaper' on it. Yet another job that I will never be qualified for, and that list is getting longer and longer every day. Last panel of the day. As you know, I've given a lot of talks about innovation in China. I think there are a couple elements to it that we want to drill down on today, and we have a great group of people to do that. So there's what's happening within China, there's the idea of competition for innovation and technology between the US and China, but there's also the rest of the world not sitting still. One of the things when you look at what Trisha just talked about, the immense energy resources required to power these systems, if a country's entering this world now it's going to have to totally revamp its energy infrastructure to support a 21st-century society. So that's one of the things we'll talk about. With us today, I'll let each of the speakers introduce themselves very briefly, and then I'll kick off with some questions. Ray, you want to introduce yourself?

Hi everyone, my name is Ray and I'm currently co-founder of a generative AI startup, so I guess I will talk a little bit about AI. But previously I started TechNode China, which is a tech media focusing on China tech primarily directed towards English-speaking investors. Actually that is still running. I'm also consulting right now.

Hi, my name is Katie Strear. I'm founder and CEO of Bloom Energy. We produce on-site reliable power using a fuel cell technology for the Fortune 100 and other companies, but ultimately for distributed generation anywhere. We also produce hydrogen electrolyzers that can make the molecule from the electricity, as opposed to making the electricity from the molecule like natural gas or hydrogen.

So what they didn't say: first of all, no one wants to be on a geopolitical panel, so we're going to stay away from geopolitics. I won't necessarily promise to do that. Second thing, when you go through each of the backgrounds, Ray is both a serial launch entrepreneur as well as a serial investor, and has done work both in the US and China, so a really deep understanding of the core behind the innovation society in China as it evolved the last 15 to 20 years. Ganesh survived four and a half years with Elon Musk, and so he doesn't have a badge but I think he just didn't bring it because I'm pretty sure they give you badges if you survive four years working with Elon Musk, especially in the pre-production phase for the Model S, which was pretty heavy. And KR, while Elon talks about going to Mars, KR actually was part of a team that put one of the rovers on Mars. So we have a pretty diverse group of people here in terms of backgrounds to talk about what we're doing with innovation. First of all, I want to draw a couple distinctions. I want to draw the distinction between invention and innovation, and I think this is really important especially when we're talking about the US and China. There is an obsession in the United States with invention, so we're always looking for what is that next great thing, what is something that no one's ever thought of before. When you look through history, there are notable examples of that: Edison, Henry Ford, Alexander Graham Bell, and so on. But the sad fact is that invention captures 10% of the value of a successful product; innovation captures the other 90%. Innovation is dependent upon how rapidly you can turn the cycle time on that original product and improve it. That's one of the things we're going to start talking about right now: how well China does that in particular. One of the things I wanted Ganesh to talk about initially, since you were instrumental at Tesla but also now at NIO, walk us through what's the mindset for the Chinese consumer, what's the mindset at NIO in terms of how you bring new products, services, technologies to market.

Okay, that's really interesting. We'll get back to that and the EV market. Ray, so you've now started a new company, part of a new startup. What's the journey from what you were doing in China as an investor and entrepreneur to doing this company now in Silicon Valley? And how do you think about the two locations in terms of starting an AI-based company here in the valley now?

Yeah, so I was in China from 2007 to 2015 and worked in investing there, actually in a couple different industries, not just tech, but real estate and media, which is highly restricted, much more restricted than tech, and then tech itself. I still keep tabs on a lot of cross-border transactions. I actually still angel invest. One of the things that's interesting about AI that is very different from a lot of the trends in past years is that you see a much larger crop of Chinese entrepreneurs who are trying to go global from day one. Part of the reason is, I won't go into the details like the last presentation of how AI is working, but that was sort of a consumer demonstration. Actually, most of the value is being unlocked in enterprises, or maybe prosumer, higher-value transactions. One of the things we were discussing earlier during the break is that China has a very weak demand for enterprise software. There are many reasons; I think we came up with at least seven or eight between the three of us, all very valid reasons. But as I was doing research for this panel earlier, just to give you a sense of scale, one of the larger legacy enterprise software companies, Kingdee, is only about $6 billion in market cap. That's very small compared to the enterprise software companies we would have listed in the US or UK or any number of other exchanges. So I think what's happening now is that while on the invention level, the US obviously with this newest age of AI, generative AI, is sort of the leader, and China is very fast making its own follow-on large language models, the foundational models. You saw earlier, there are four unicorns in that space right now and a couple of others in related areas, but at least four AI unicorns in the foundational layer space. On the application level, though, there's much more revenue happening abroad outside of China. In fact, there are two of the more well-known generative AI companies, really like ours only a year old or so, that have more than eight figures of annualized revenue. There are maybe a few dozen of these from China. But you can see at least one of them, and this is public, a company called HeyGen, primarily based in China, raised money from Sequoia China, and recently kicked out Neil Shen as their board member and converted to a totally US structure because they find that all of their revenue is coming from the US and they're now seeking follow-on investors exclusively from the US. So I think this is something that's a little bit different for the AI space that you don't really see happening with other sectors or more legacy sectors. In e-commerce and traditional mobile internet, what you typically saw was large platform companies in China expanding abroad, or a company like Shein taking supply chain advantages and exporting it abroad, but ultimately they're really taking advantage of core manufacturing advantages in China. But with AI, the invention technology is actually US, and also the market is US. It's just the engineering talent or the leadership itself that might have been Chinese-born or China-based for the moment. It's interesting.

So Kai-Fu Lee in his book AI Superpowers, one of his fundamental premises was that China would have a huge advantage in data, that the amount of data in China would actually allow it to exceed or leapfrog what the US had. Part of that was tied to the consumer data you see in e-commerce and payments markets, which are all very large there. How do you reconcile those two things given what you were just saying in terms of entrepreneurs in China saying the revenue looks like it's coming from overseas initially? How are those two things going to reconcile, especially if, and I won't ask you to comment on the US-China side, I'll comment on that later, but it's not going to be smooth sailing in terms of just an open market where technology is going to flow back and forth in the next several years.

No, I mean, I'm not a geopolitical expert, but it's very clear that at least on the foundational model level, China has no intent to rely on US technologies. They're really trying to home-grow. Actually, I was looking on my phone at the beginning of the session because I wanted to make sure I got the number right. In October of last year, there was a report that said China had nearly quadrupled its number of large language models from about 70 to almost 200. I think the exact number is 238 in just four months. So there are a lot of people working on AI technologies in China right now. They're not as good as what we have here cutting edge. I think the latest that was released by one of the unicorns, DeepSeek, they openly say they're 90% as good as GPT-4, but GPT-4 is like half a year old. So anyway, for how to reconcile what Kai-Fu said, it's funny. I think overall, no one in China that I talk to in the industry really believes that China is number one in AI generally, only in specific subfields. So I think what Kai-Fu said could be true for fields like payment data, machine learning for fraud, or facial recognition, for example. So again, the previous generation of AI had very different technological paradigms versus generative AI, where really you're combing through the whole internet. The level of data that we're talking about is still different. It can also be true in the future for other applications of AI. So one of the more popular areas where entrepreneurs are spending a lot more time and getting a lot more funding than here in the US is embodied AI, where you're using generative AI but also using it to interact with the physical environment.

You mentioned robotics in particular?

Exactly. Although I actually just did look up the VC funding last year, and the US did still lead China in robotics, but it was primarily because of a few healthcare, very high CapEx types of things.

So the first two panelists, Ganesh talked about cars, it takes years to develop a car and put it in the manufacturing facility. Ray talks about now we're in the AI side, ChatGPT for 5, 12, 18 months now. We go to something that's really hard. And what KR did when he launched Bloom Energy, which was 22, 23 years ago, he dealt with a new material and a new form of generating electricity. So I'm going to ask KR to talk about when we think in time frames of invention and in time frames of innovation, the energy sector trumps everything else in terms of the sheer scale and difficulty. But none of these other happy things we're talking about, charging our cars at night and running large language models, are going to exist if there's not enough energy. So KR, maybe you'd like to offer some comments in that regard.

Sure. Good afternoon everyone. Let me give a shout-out to Gary Rieschel for your leadership in building US-Asia relationship and being the chair and running Asia Society. Thank you. That's a bot that Trisha put up on this stage to say that. So can the real KR please come out? Thank you very much. I truly mean it. Look, I think Gary, you're asking the right question. Energy at the core of it: there's nothing since you got up this morning that you could have done in your normal course of work if energy didn't exist in some form or another. The water in the faucet would not have come for you to brush your teeth or use the toilet. Everything you do, the fact that you're able to see me, there are electrons shining photons that come to my face and reflect to you, and that won't happen without electricity. So everything we talk about takes energy, and we are in a unique place right now: energy transition in the world. It's not just in any country. It's particularly meaningful for emerging economies because there is no country that's energy rich but economically poor. There's no country that's economically poor that's energy rich. They're directly correlated. So it gets into the heart of national security, economic security, well-being. For the first time in the Industrial Age, even in developed countries like the US, availability is becoming the key issue: reliable availability of energy. You can just look at me and say, what are you talking about? We launched an energy transition. The way I try to tell people, when we try to shut things off before new things come on, which gets to your question on invention to innovation to commercialization and affordability, is we are flying a plane. The plane is called human civilization, human development, human well-being. We have to swap engines while we are flying it because we can't stop it for a little while to keep flying. There are people advocating we shut the engine off before the new engine is ready. That's what's going on. If you just look at the US, we have retired 290 gigawatts worth of electricity generation in the country by shutting down nuclear and coal. We have added 250 gigawatts of renewables at a rapid pace. It's a fabulous thing, it's good for the planet, but 250 gigawatts of renewables give you about 100 gigawatts equivalent of baseload power that these things give over 24/7. So we have a deficit of 150 gigawatts over the last 10 years in this country. In the meanwhile, what happens? The last 40 years, the CAGR of demand growth in the country has been 5%. EVs alone will require a 6 to 8% growth. AI will require 6 to 10% growth. Electrification of everything in the homes will require another 8 to 10% growth. And if we take industries and try to electrify it, it'll take 100% growth. We have a 40-year history of doing 5%. We have dropped our generation, and we have not built transmission distribution infrastructure. That's America. Then you go to emerging economies, and I'll speak about India since I know that better than China. There are 350 million people between the ages of 0 to 14, equivalent of the US population. They're getting educated, they're ambitious, they have to grow. It's equally distributed almost between male and female, 185 million to 165 million. They've got the demographics. I was just recently there, and it looks like what it felt like when you visited China about 25, 30 years ago. Little towns are getting infrastructure, roads are getting built, airports are coming, planes are coming in places where people haven't seen planes other than in the sky. Things are changing. Just think about that: 1.4 billion population. Last 20 years it's been growing at 7.7%. Let's assume that it's only 5% averaged out for the next 20 years. They'll have to double their energy use. Now let's get to the timeline for invention to innovation to commercialization. Cambridge has done a study of all energy technologies. The median time to commercialize something and make it viable is 40 years. Bloom's trying to do that in 30 years, and that'll be a record, at the lower end of what we do. So if you're opening your Scientific American and reading about this great invention, think about the median age it takes to impact. So this is the energy construct, and you're not going to have leadership in any field if you don't have security of energy. So therein comes the issue of what competition and energy mean. In the short term, it means going and securing as much energy as you can from whoever you can to build your country. This is why countries like China and India still work with countries that we think they shouldn't be working with, because they don't have the luxury like us of being endowed with all these energy assets. Midterm, both these countries realize fossil fuel will be there only for so long, and they need to transition. They're making a great job on renewables. If I take India again as an example, I was there speaking to three different CEOs who all have two-wheeler industries: Hero MotoCorp, Honda in India, and TVS. Between them, 60% of two-wheelers. I'll tell you in the next five years, there will not be a single two-wheeler that consumes oil in India. It'll all be electric. That's going to happen at a lot faster rate than cars are ever going to have in India. So for them to figure out that innovation chain and do what they need to do, and then be able to export to Africa and other places, that's a play for them. You look forward to 2050, the competition is going to be the molecule: how to make hydrogen using renewables. But if you don't sow the seeds now, given that 40-year time frame, you're going to mess up. These countries have wised up. China will have a huge investment in hydrogen infrastructure, so will India, so will the US. That's the competition.

So going back to the fundamental premise of the panel, which was about innovation in US and China, if we look at the idea behind invention and innovation, recently even our current president said 10 years ago that China couldn't innovate, and I think we've kind of set that to bed, that that's not true. But how do we think, and I'll start with Ray and then go to Ganesh, how do you in your experience think about China as a place that can invent? And the innovation on the consumer side, looking at Xiaomi, looking at what NIO has done, that cycle when you have something and then it's a question of incremental improvement, the Chinese have done a phenomenal job of that. But if we take a step back and look at the core of invention, how would you think about the capabilities that exist today in China and the ability to do that and execute on that?

I think it's a lot better than before, because before the society was so poor that a large part of it was about catching up and getting to so-called middle-income status. But I think the learning is just because an invention was not made in China does not mean China can't invent. There's a big difference. A very simple example I like to use is that people in the past, less so now, would say Chinese developers don't know how to make anything, they just copy. If you actually work in Chinese companies, their reasoning tends to go like this: it's very practical. It's not whether or not I can make that thing; it's if it has already been made and it is open source, for example, why should I make a more original solution for the sake of being original, rather than just take something that has already been working? Now sometimes that extends to methods that are not as clean as just using open source software. But you basically see what has happened is when the incentives change, then the inventiveness grows. Taking open source software as an example, in the past, 10 years ago, there were very few contributors from China. But once it became a commercial KPI, for example Alibaba made it a KPI for some of their departments to publish open source software and get GitHub stars or whatever, then you see a lot more contributions. So what I would say is right now, I think the market forces in general still reward the innovation, or I guess innovation we're describing as the scaling up part, much more robustly than they do the invention, which you yourself said invention captures only 10%. So that's why I think we don't see as much over there. But once the knobs are turned a little bit, then you see more. For example, I'll throw in one more anecdote relevant to energy: patents in China. Before, a lot of Chinese patents were because of this sort of fake KPI where the more patents you have, the more bonuses you get inside your company, or you get special citizenship bonuses in certain cities. But the Chinese company with the most number of patents is China State Grid. Presumably they're not all related to energy; they're probably some other things. I was just looking up AI patents, and it is one of the top 10 contributors even for AI patents. But again, I think that is because there are certain incentives in society or in that corporation itself to be inventive. You learned early in the sales game: you do get what you pay for. So if you're setting the incentives a particular way, that's what you're going to get.

Ganesh, you have an interesting view of Tesla and NIO. How do you think about when you hear people say the Chinese can't innovate, or the innovation at Tesla was far more innovative than the Chinese firms, and you're sitting where you sit today, how do you think about that?

Yeah, I think so. I'll wait to the end to summarize things. KR, one thing I'd like you to comment on, not necessarily right in your wheelhouse in terms of what Bloom does, but ultra-long-haul energy transmission. China's had a huge investment in that area. When you think about grid capability and places like India or other developing countries, how important is that ultra-long-haul transmission of energy going to be compared to the traditional grid structures we've dealt with in the past?

My personal view is that ultra-long haul doesn't work for most countries in the world because of energy security, because of how strongly cyber is connected to the future grid and pure security issues. The future wars will be fought on turning your power off, not about putting boots on the ground. So there is a national security element that prevents you from doing those things in a meaningful way. If we fought for water and we still do coming out of rivers, give me a break, that ain't happening. That's utopian on a piece of paper, it's not happening. But the grid existing in a big way is a flywheel that energy needs, so the grid will exist and that can be made a lot more efficient. How do you in a modern world where, God forbid, we were in a surgery room, a robot is going to conduct the surgery with digital data, and if there is no power, a doctor is not going to know how to perform that surgery. That's the future of medicine. Self-driving cars are going to be paralyzed on a street if data is not available. So electricity being off is not an option. That reliability factor has to go to a much higher level, and the only way you secure that is using microgrids and macrogrids. Microgrids are where the future is, with making the grid a lot better in terms of transmission distribution because renewables need to come from there. That combination is how I see it.

Let me just because I think that's right. Part of the reason I brought that up is when Ray brought up the number of patents at State Grid, a significant number of those are due to long-haul transmission of energy, which has disappeared from the Chinese discussions on energy about four years ago for many of the reasons you just talked about. They realized if they can shut off someone's grid, someone can shut their grid off, and from a security perspective it just doesn't make sense.

Given that's one right, but think about the other. If electrons coming to your premises, whether it's home or companies, is the lifeblood for everything you do, knowing everything you know about climate change that's already baked in, but with more violent storms, hurricanes, typhoons, more wildfires, more hot days, more cold days, would you design a piece of wire on a pole as the way you get that lifeline delivered to you? Which one of you would? So yeah, we need to think differently about it.

So we're going to have 10 minutes for questions. Let's see, Mark, you got yours up quick, so we'll go to Mark Cohen first and then get everyone else's questions ready.

This question is primarily directed to Ganesh. While you were talking, I looked up the PTO website and your company has about 400 patents in the US, 428, which by the way is four times State Grid, which is an interesting phenomenon. State Grid files a lot in China, very little in the US, very typical for state-owned enterprises. Why are they innovating if they have a monopoly? It's an interesting question. What is the quality? Hard to say. But NIO is an interesting profile because it looks like you're innovating all over the world. I saw in the patents inventors in Germany, several from different places in the US, and of course China largely in Ningde, which itself is a cluster for battery technology. So we hear a lot in the US, particularly in EVs, that the US can invent but we can't manufacture, and that all these great ideas are ultimately being applied primarily in China manufacturing in China. So I'm wondering how you view the ecosystem and the challenges facing the US to compete in this area, or just to innovate in this area and contribute back to China, because that looks as if just reading the patents very quickly, only a few of them look a little bit like what NIO is doing.

Okay, question. Someone in the back over there, Mr. McGee? Jim? Okay, all right. Let's make this geopolitical.

I want to ask about talent. America's had this wonderful innovation because the talent of the world comes here. We have two people from India and a person from China on stage talking about innovation. Where our policies have gone in recent years, as we all know, it's hard for people with STEM talent to get visas to come. We're scaring professors out. What is the future of that? We've had a nice run here in America because we've been open to talent from around the world, and we've become kind of paranoid and worried about everybody from outside. Will other countries start to surpass us in 5, 10, 20 years because the talent will be going elsewhere or staying at home?

Sure. Well, I know that Tsinghua University, from where I have a degree, just put out a report that said something like 86% of their alumni go back to China or just stay in China to work. Actually, the number might be higher, I don't remember exactly. I think there have been a lot of reports recently about how PhDs in STEM from China were stopped at the border and were not allowed to continue their research here. I think that is definitely a loss for the US. I don't know that China is really picking up the talent though. So I think it's really just us stabbing ourselves in the foot repeatedly. It's our game to lose.

Yeah, so look, as we sit here, Congress and the White House are trying to figure out a budget plan. There is no single policy that has zero costs and gives so much to the country than immigration. Net net, it is a positive. I think the reason it is stuck is because we have combined talent migration and comprehensive immigration together in one bucket, and neither party has the political will to separate it out and do what's right for the country. This country was built to where it is today because of a steady stream of immigrants. I don't think the future will be any different if we still want to have our position in the world.

Yeah, the immigration issues would be an entire day or week or month. I've never been so frustrated in all the meetings in DC as when the issue comes up about foreign students coming to the United States, primarily focused on China, and the comments like 'well, what if they're spies?' These come from senators, these don't come from idiots. So I append the last comment. But the idea that you can look at the entire history of this country, 65-70% of the CEOs of AI companies are immigrants. It does not take a rocket scientist, no offense KR, to actually look at this and say we have a serious problem. If we want to be successful going forward on these very long-term intensive competitions, you have to have the best and brightest people in the world wanting to be here. So that's going to be the end of things. I want to thank our panel for the conversation and thank all of you for participating. Thank you.

That was a perfect way to end our conference. Thank you. It's a good day when Gary wears a tie. You've been a really terrific audience. We hope you learned some new insights, met new friends and contacts. This concludes part eight of our Seeking Truth Through Facts US-China series, which we started back in July. We still have many programs to cover, so stay with us. All of our videos will be up on our website. Thank you to our fantastic speakers, many who traveled here to spend the day with us, to our fantastic and dedicated sponsors, and to our team, our staff, our interns, and our volunteers. Let's please hear it for our team: Nina Udagawa for taking the lead, Man Wong, Christy Young Smith, Jason Fong, and Aaron Wen. There's been hours of rehearsals and preparation. Our volunteers and interns, thank you so much again. You are so talented. We host monthly programs in San Francisco, Silicon Valley, virtually, and now in Seattle, our new center there. Please do come join us. We make a lot of them available to both audiences. Join us as a member, and on behalf of the Asia Society Northern California, thank you for coming out today.