Challenges & Opportunities of AI in Sustainability_Silicon Valley Asian Business Talk_Alfred_Chuang

Thank you, Roger. Thank you for inviting me. This is my honor and really it's a privilege to be here today. And thanks Heist and thanks Tha to actually have me speak today. I think the last time I actually spoke at a Heist event was maybe 22 years ago, which someone just found. I think I went to a dinner in Catino and I spoke at the dinner. It was very well attended, and so I have extremely fun memories. It's been a long time, so far for to be back. So I was born in Hong Kong and I graduated from University of San Francisco with a computer science degree. And then I went to UC Davis for my graduate studies and then I got my master's degree from them. And then halfway from my PhD program I quit and I went to work for Sun. I was at Sun for about just under nine years and I founded my own company called BEA Systems. BEA was a very successful venture. We went public in four years and pivoted the company into the web space in the late '90s and reached $58 billion market cap and eventually got to just under $2 billion in revenue and sold to Oracle for $9 billion, just under $9 billion during the financial crisis. And then I've done other companies and been very active in investing in early stage companies for the past decade plus. And so about five years ago I started a firm with two other partners focusing on seed stage investment in infrastructure technology. So that's what I've been very focused on in the past few years. I'm having a grand time. This is like 2000 all over again for this AI era. So very thankful to be here.

Actually, there are very interesting statistics. I don't know if any of you actually follow energy studies. I do. Actually, I spoke at a conference called One Earth, which is a green conference, two months ago. So kind of preparing to go to the conference, I was kind of surprised they invited me, really kind of like a software AI stack guy, to go speak in a conference to save the planet Earth. And I was looking at the International Energy Agency's numbers, right? And they're estimating how much electricity we're actually using in AI aggregately. And then we know you kind of have to look at, actually, I don't look at just what we're using for the H100, A100, and the competitive, you know, MI300X and that related type, you know, super-scale stack. I'm actually looking at how much those things are now driving a data center use of energy, right? It's amazing. The number will say in less than two years, so before we enter into 2027, we'll be consuming more than 10,000 trillion watts per hour of electricity. So just to put this in perspective, right, that will equate to or maybe slightly over what Germany, with 83 million people, is consuming. So AI would consume, just AI in aggregate in data centers, will be consuming more energy than all of Germany with all the 83 million people on an hourly basis. This is a very frightening thing. So you can almost look at in 2030 likely it will match or exceed any country in terms of consumption just on the data center itself. But yet, us techies, I assume most of us in this room here, we don't think about this stuff, right? We just say, ah, don't worry, we somehow we would just either take a bigger hole, you know, somewhere in Saudi Arabia or we harness, you know, more solar and it'll be all okay. But the destruction we're doing to the Earth is actually quite serious. So when I got on the stage at One Earth, I thought, oh my God, I'm the bad guy, right, in the conference. So what do I say? Can I please these people so they think we're conscious? So this is actually what I honestly believe. So I think for the longest time, you know, everything can only be fixed if you have data, right? I'm a very data-driven person. I think our world has become very data-driven. Everything is very data-driven. But yet in energy, we're terrible at it. So we don't really measure the data of energy use well. We only measure the energy use because of cost. So moving to places that electricity is cheaper, you know, like you look at all the crypto mining people, they're constantly moving because they're looking at hydro, they're looking at different things because they're looking at lower cost. But we're not really looking at the full picture about what damage it's doing to the environment itself. So how do you fix it? It's actually quite simple because historically up to this point, we don't build any of that in our software stack. From the operating system up, from even the microcode all the way very up to the application, there's no such thing as energy consciousness in this program. This time we must, as we're building the stack out, the bottom of the system software, you know, people are building specific operating systems to operate the next generation of new network-driven applications, all the way to all the security software, the code generation, the operation software, definitely the application have to be energy conscious. That's the only way you can fix this problem. And use AI to curve it for most of the time. I don't know how many of you, like I do, we don't turn our computer off at night, do we? Right? It's idling and we're not using the resources either. So we don't really do edge computing to a point where we can recycle the resources so we can reduce the compute use and also hence reducing the energy use. Those are the kind of things I think humans literally are incapable of. We must have, I would say, resources that never get tired, which is the AI part of it, to help us do some of these things. I think the opportunity is now we can align green tech, energy tech alongside with AI as we are progressing so quickly to solve this problem at the same time. But in this driver, obviously, I think the people that are in the energy field are extraordinarily committed to leverage this time around. So I think there's hope.

It has to be. It has to be. Because depending on human alone, we know that we didn't get there for all this time. So we have to use a different method.