WATCH NOW: Elon Musk Speaks at G20 Innovation Ministerial Summit in North Carolina | AI15

So, our first speaker is Elon Musk, CEO of Tesla and SpaceX. Elon has shown that overcoming the challenges of physics and engineering may be easier than getting past bureaucratic red tape. I believe his dedication to doing both continues to inspire people around the world. So, thank you, Elon, for joining us. We were just concluding our introductory session where we made it clear that under this presidency, we believe that economic growth is one of the most important drivers for all of us, and we believe that technology is a big part of that. I've been at the forefront of this field for a long time, and I think one of the key questions for the group, and something we wanted to discuss with you, is that since you've led innovation in multiple countries around the world, what do you think separates countries where innovators can successfully translate achievements into applicable technologies from those where progress actually stalls? Because I believe there are many lessons to be learned about what people and governments should do, and perhaps what they should not do.

Well, I think a lot of people have talked about this, and I think some of it is quite straightforward, which is that you should have a relatively regulation-free environment. In other words, new things should be legal by default rather than illegal by default. In the European Union, for example, we find that the level of regulation is very high, and things are generally illegal by default. This hinders progress in new technologies. It slows it down. It does not stop it in the end, but it slows it down very significantly. Of course, you need investment capital and a supportive environment for new companies. You can think of new companies as small seedlings in a forest. Most countries tend to provide a lot of support for the large trees in the forest, and not enough for the small seedlings. But large trees do not need this support. It is the small seedlings that are needed, i.e., startups. Therefore, the system should generally be biased towards supporting small trees rather than large ones. But this is rarely the case. It is because large corporations usually have access to the leaders of countries, while small startups do not. Therefore, you really need to promote the growth of startups and take effective steps in this regard. As I said, make things legal by assumption, not illegal by assumption.

One of the questions that many countries here face is a question related to adoption. I think there is a general recognition that adopting emerging technologies can be very beneficial to economic growth, regardless of the shape or size of any country. How do you view the issue of adoption? What should countries do to encourage adoption? Is it down to the same organizational structures, or how do you view adoption as a key driver of growth?

I think you should have a proactive approach and be willing to try new technologies. Instead of remaining somewhat stuck in the past. Yes. Naturally, new technologies need encouragement, and I believe they should be adopted. We will see, and indeed we are seeing, huge productivity gains from artificial intelligence, and we will see enormous productivity gains from robots. Things like a self-driving Tesla car would be, or are already, a huge advantage for users. I believe that humanoid robots will bring about an incredible change. To give you an idea of the scope, I think artificial intelligence will increase the global economy by between 20 and 30 percent.

This is my rough estimate. That is, approximately $20 to $30 trillion annually. Artificial intelligence will be able to do anything digital and anything that does not require manually shaping atoms. Most likely by the end of next year. As everyone knows, artificial intelligence is already incredibly good at software. It is reaching a stage where artificial intelligence will not only be good at software, but will reach what I call the "Stockfish" level. "Stockfish" is a chess program that can easily beat the best chess players in the world.

In fact, you can now play "Stockfish" on your phone and defeat Magnus Carlsen at chess. My prediction is that at some point next year, software artificial intelligence software will become so good that it will reach the level of "Stockfish," meaning that it will be impossible for humans to compete with artificial intelligence in writing software. But artificial intelligence will completely surpass all humans in the field of software. I think it will be very good, perhaps on the level of "Stockfish," but it will certainly be very skilled in all forms of engineering and anything digital. Literally within 12 to 18 months. Important. One of the statistics is actually, let me add one more thing to that. I apologize. An increase of 20 to 30% in the overall global economy. So we are talking about a very large amount of prosperity here. Only digital artificial intelligence. But from robots, from humanoid robots, think of artificial intelligence for a multi-purpose robot.

I think we will see a multiplier effect on the global economy. This means that the economy can be increased by 10 times or more. Important. These are mind-boggling numbers. Yes, I was about to say that; I mean, one of the statistics we've given to the group here is that in the four years since the launch of "ChatGPT," I think there are over a billion people around the world already using artificial intelligence. There has been very rapid adoption of this across the country and around the world. I think the question directed at you is: Many people think of the next chapter as being about applied or physical artificial intelligence. Where do you see robots heading? How quickly can this be done? I think I remember in the first Trump administration we were talking about automated factories, and now it's been 10 years since that. So, how quickly do you think these changes will occur in the world of physical artificial intelligence?

Yes, anything physical always takes longer than anything digital. In digital technology, when you sell something digitally, it's software that you can easily copy across other computers. But when it comes to physical things, you have to build an entire massive supply chain. You have to move lots and lots of atoms. Supply chains are extremely global at this stage. So, that's why it takes longer. However, when you think about humanoid robots, the way you think about them, I think the correct framework is to consider that the utility of a humanoid robot, i.e., a multi-purpose robot, will be roughly the product of the AI program multiplying by the quality of the AI chip in the robot, by its electromechanical dexterity, especially the hands. Well, all three of these things are improving exponentially. The benefit of the robot is the product of these three things multiplied together. Then, when robots are made, robots will start making robots. Therefore, you will get a repetitive effect. It starts very slowly, but then grows at an explosive rate. So, if I were to say 10 years from now, I would say there will be more than a billion humanoid robots. Wow.

The productivity of a single robot will likely be five times that of a human. That is, productivity within 10 years of humanoid robots, and I think that's a conservative estimate by the way. This is something I would be willing to bet real money on. That there will be at least one billion robots within 10 years. And that those robots will produce at least five times the output of humans. This means that one billion humanoid robots would be more productive than all humans combined. Wow. Turning a little to a question facing the United States today. As you know, data centers have been a big political issue for the past 6 to 8 months here in the United States. I think there is a general understanding that in order to lead and operate the next AI revolution, we need to provide electricity and data center computing capacity.

startups. Therefore, the system should generally be biased towards supporting small trees rather than large ones. But this is rarely the case. It is because large corporations usually have access to the leaders of countries, while small startups do not

. Therefore, you really need to promote the growth of startups and take effective steps in this regard. As I said, make things legal by assumption, not illegal by assumption.

One of the questions that many countries here face is a question related to adoption. I think there is a general recognition that adopting emerging technologies can be very beneficial to economic growth , regardless of the shape or size of any country. How do

you view the issue of adoption? What should countries do to encourage adoption? Is it down to the same organizational structures, or how do you view adoption as a key driver of growth?

I think you should have a proactive approach and be willing to try new technologies.

Instead of remaining somewhat stuck in the past. Yes. Naturally, new technologies need encouragement, and I believe they should be adopted. We will see, and indeed we are seeing, huge productivity gains from artificial intelligence, and we will see enormous

productivity gains from robots. Things like a self-driving Tesla car would be , or are already, a huge advantage for users. I believe that humanoid robots will bring about an incredible change . To give you an idea of ​​the

scope, I think artificial intelligence will increase the global economy by between 20 and 30 percent.

This is my rough estimate. That is, approximately $20 to $30 trillion annually. Artificial intelligence will be able to do anything digital and anything that does not require manually shaping atoms.

Most likely by the end of next year. As everyone knows, artificial intelligence is already incredibly good at software. It is reaching a stage where artificial intelligence will not only be good at software, but will reach what I call the

"stockfish" level. " Stockfish" is a chess program that can easily beat the best chess players in the world.

In fact, you can now play "Stockfish" on your phone and defeat Magnus Carlsen at chess.

My prediction is that at some point next year, software artificial intelligence software , will become so good that it will reach the level of "Stockfish," meaning that it will be impossible for humans to compete with artificial intelligence in

writing software. But artificial intelligence will completely surpass all humans in the field of software. I think he will be very good, perhaps on the level of "Stockfish" , but he will certainly be very skilled in all forms of engineering and anything

digital. Literally within 12 to 18 months. important. One of the statistics is actually, let me add one more thing to that. I apologize. An increase of 20 to 30% in the overall global economy. So we are talking about a very large amount

of prosperity here. Only digital artificial intelligence. But from robots, from humanoid robots, think of artificial intelligence for a multi-purpose robot.

I think we will see a multiplier effect on the global economy. This means that the economy can be increased by 10 times or more.

important. These are mind-boggling numbers . Yes, I was about to say that; I mean, one of the statistics we've given to the group here is that in the four years since the launch of "ChatGPT ", I think there are over a

billion people around the world already using artificial intelligence . There has been very rapid adoption of this across the country and around the world. I think the question directed at you is: Many people think of the next chapter as being about

applied or physical artificial intelligence. Where do you see robots heading? How quickly can this be done? I think I remember in the first Trump administration we were talking about automated factories, and now it's been 10 years since that. So, how

quickly do you think these changes will occur in the world of physical artificial intelligence?

Yes, anything physical always takes longer than anything digital. In digital technology, when you sell something digitally, it's software that you can easily copy across other computers. But when it comes to physical things, you have to build an entire

massive supply chain . You have to move lots and lots of atoms. Supply chains are extremely global at this stage . So, that's why it takes longer. However, when you think about humanoid robots, the way you think about them, I think the

correct framework is to consider that the utility of a humanoid robot, i.e., a multi- purpose robot, will be roughly the product of the AI ​​program multiplying by the quality of the AI ​​chip in the robot , by its

electromechanical dexterity, especially the hands. Well, all three of these things are improving exponentially. The benefit of the robot is the product of these three things multiplied together. Then, when robots are made , robots will start making robots. Therefore,

you will get a repetitive effect. It starts very slowly, but then grows at an explosive rate. So, if I were to say 10 years from now, I would say there will be more than a billion humanoid robots. wow

. The productivity of a single robot will likely be five times that of a human.

That is, productivity within 10 years of humanoid robots, and I think that's a conservative estimate by the way. This is something I would be willing to bet real money on . That there will be at least one billion robots within 10

years. And that those robots will produce at least five times the output of humans. This means that one billion humanoid robots would be more productive than all humans combined . wow . Turning a little to a question facing the United States

today. As you know, data centers have been a big political issue for the past 6 to 8 months here in the United States. I think there is a general understanding that in order to lead and operate the next AI revolution, we need to

provide electricity and data center computing capacity to perform the training and reasoning processes for this AI. How do you think about this particular issue ? Where are we on the curve of what we have built versus what we need

? As government leaders here consider how to prepare their economies and build the right infrastructure for energy and data, how should they think about computing and energy needs over the next few years?

Well, there is indeed a real energy crisis. So , this is actually something you can see if you only follow the topic of artificial intelligence on the " X" platform, which, by the way, is where almost all the discussion of artificial intelligence takes place.

I think you'll get a very good idea of ​​where things are headed. This is how I get my news, and it's very good. Every important person in the field of artificial intelligence publishes on " X". Therefore, I advise you to

go to the topic of artificial intelligence on "X" and you will understand all these things and get daily details about them.

The consensus at this stage is that there will be a significant energy shortage next year. So, it's not in the distant future. It is expected, and I believe the consensus estimate among analysts who closely follow the field of artificial intelligence, is

that there will be a power deficit of at least 15 gigawatts in 2027 for AI chips.

This is perhaps to be expected, as the rate of production of artificial intelligence chips is increasing at an enormous rate.

It is rising at a rate of 40 to 50% annually. But available energy outside of China is increasing at a rate of 10 to 20% annually. It is clear that the fastest-growing thing will eventually overshadow the slowest-growing thing.

In fact, I would say at this point, there are energy challenges even before next year, which is why Google, Anthropic, and many other companies are leasing computing capabilities from SpaceX. Because we have managed to operate artificial intelligence better

than anyone else so far . But this was achieved by building our own power plants. It's the only way we were able to do it.

China possesses a huge amount of electricity. But because of the export ban on graphics processing units (GPUs), one cannot build data centers with the latest chips in China. So, in fact, the consideration here is what is the size of

electricity growth outside of China? Nations, and this currently represents a significant deficit compared to the production of artificial intelligence chips . So, I think there is an incredible opportunity for countries around the world to say that if they are

interested in AI data centers , they can build a lot of capacity and offer it to AI companies . In return , of course, these data centers will be taxed and will have to pay reasonable fees, etc.

important. But it creates an opportunity for many countries. certainly. Thank you very much for your time. It means a lot to us that you've joined us here, and we really appreciate it.

Thank you very much. You're welcome . Thank you. Thanks. amazing. Now, we will hear from David Sachs , in his capacity as the former White House official in charge of artificial intelligence, and my colleague in chairing the President’s Council of Advisors on

Science and Technology. He is a successful person