저자: 손정의 · 외국인
요약 보기
결론: 손정의는 AI 특이점과 엣지·클라우드 AI 확산을 강조하며 ARM의 장기 성장을 주장하나, 특정 종목·지수·자산의 단기 가격 방향에 대한 예측은 없다. 주제 1: 과거 빅뱅과 미래에 대한 태도 - 현재를 충실히 살자는 일본 비즈니스맨의 수동적 태도에 비판적 입장 제시 - 미래의 신호를 민감하게 포착해 남보다 빠르고 깊게 생각해야 한다고 강조 - 빅뱅(138억 년 전) 이후 우주가 여전히 팽창 중이라는 비유 사용 주제 2: 특이점(싱귤래리티)과 AI 혁명 - AI 지혜가 인간을 넘어서는 특이점 도래를 주장 - 초지능이 모든 산업을 재정의할 것이라고 전망하며 인류 역사상 최대 혁명으로 규정 - 옥스퍼드 연구의 AI 대체 직업 시점 예측(2024·2027·2030)은 오차가 크다고 비판 주제 3: ARM 인수 배경과 엣지 AI - 소프트뱅크는 최근 10여 년간 통신사로 인식되나 창립 35년 중 통신 사업은 약 1/3 기간 - 정보혁명이 AI 혁명과 같아진다는 관점, 통신도 AI를 위한 통신으로 봄 - 전 세계 스마트폰 약 70억 대 수준 사용, 연간 15억 대 출하, 스마트폰 내 ARM 칩 탑재율 100% 주장 - 누적 출하 1,000억 개 돌파, 2030년 1조 개 전망 - 가전·자동차·공장 센서 등 엣지 기기에 AI 기능 탑재 프로젝트(Gilliam) 추진 주제 4: 클라우드 측 AI 연산 - GPU 중심 클라우드 칩당 연산 능력이 2030년까지 약 200배 증가 전망 - 클라우드와 엣지 AI 칩이 5G·6G·7G 등 초고속 통신으로 연결될 것이라고 주장 - AI 진화 속도가 이차곡선처럼 가속된다고 전망 리스크·조건: 기간이나 가격 목표 언급 없음. 장기 견해만 제시. 출연자 전망: ARM의 엣지 칩 독점적 지위 강조, 장기 성장 확신. 단기 방향 예측 없음.
다듬은 글
Good morning everyone. We just had an opening video, and the future is truly about to unfold before our eyes. Often, people say that we don't know what the future holds, so we should just live our lives to the fullest in the present. But is that really okay? I think many Japanese business people have that kind of passive way of thinking, just looking at the reality that is happening now and thinking about how to deal with that reality, and that it's okay to just live as we are.
I feel a sense of crisis about that. Is it really okay to just stare intently at the present reality and deal with it seriously and diligently? Is it really true that we don't know what the future holds? I think that maybe we think that way because we're not seriously trying to understand it. There are always signs of the future that are right in front of us right now.
We should be sensitive to those signs, take them for ourselves, take them for our own future, and work hard to think about them more quickly, more seriously, and more deeply than others. Whether you approach things with the intention of gaining insight or not, and whether you try to change the current situation, change your own current situation, change the current situation of our world, I think the results will be completely different for people or companies that make such efforts compared to those that don't.
Today I would like to talk a little about the Big Bang, a little about the future, and how the SoftBank Group is trying to approach it. First of all, as I think you all know, it is said that the Big Bang occurred 13.8 billion years ago. I am not a space expert, so I am just saying that it is said that something like that happened around that time, but because of this Big Bang, the universe expanded, and various planets, stars, and galaxies were born, and it is said that it is still expanding even now.
The universe is still expanding even now, so the Big Bang did not end in an instant. Singularity As another Big Bang, I think the Singularity is coming. In a word, this Singularity is a human coach So, the wisdom of artificial intelligence will surpass human wisdom, and when the wisdom of human-coached AI surpasses human wisdom, superintelligence will be born, and that superintelligence will redefine every industry.
If all industries are to be redefined, then how should we approach that? I believe this is the greatest revolution in human history. With the evolution of AI, various calculations will be performed, but I believe that these calculations will not be done only on the cloud side or only on the edge side, but both will work simultaneously in parallel, complementing each other, to perform more advanced AI computations.
It's not that one side and the cloud are more important than the other, but that both are important. First, I'd like to talk a little about the edge side. When I acquired Arm, there were many people in the Utah media and the general public who wondered why a mobile phone company would acquire Arm, why an old company now?
I won't go into detail here, but what I'm about to say will explain why I acquired them. I think you'll understand a little better why I think ARM is so important. For the past decade or so, SoftBank has been seen as a telecommunications company, but in SoftBank's 35-year history, we've only been a telecommunications company for about a third of that time, just over a decade.
SoftBank has been a company that has been driving the information revolution since its founding. One of the core businesses in that information revolution is telecommunications. And I believe that this telecommunications is telecommunications for AI. I think we are now heading towards a situation where the information revolution equals the AI revolution.
ARM will become an indispensable chip on the edge side. I think everyone here is more advanced than the average person, but I think there is probably almost no one here who does not have a smartphone in their pocket. Right now, people all over the world, last year alone, 1.5 billion smartphones were sold.
If we average the number of smartphones used per person over two or three years, that means 3 to 4 billion smartphones are being used by 706 people all over the world. Of course, among those 7 billion people... There are babies, young children, and the elderly, but I think that most teenagers and adults are now using smartphones.
And inside those smartphones there is something that is 100% present, 100%, not 99%, not 98%, but 100%, 100.0%, and that is the ARM chip. Moreover, this is not a component that waits for the surrounding components, but the equivalent of the engine in a car. In a world where people can no longer live without it, in reality, ARM is responsible for designing the engine, the heart of the machine, 100%.
I think this is an extremely important component. To use an analogy from Othello, I think it is equivalent to one of the four corners in the game of Othello. This ARM chip is not only in smartphones, but is now starting to be found in everything, from cars, gaming devices, and all sorts of things, including sensors in homes and factories.
It has been shipped out, and the cumulative number shipped up to last year exceeded 100 billion. 100 billion chips. As I mentioned earlier, the Earth's population is 7 billion, so ARM chips have been deployed on a scale far exceeding the Earth's population of 7 billion. Moreover, it is predicted that in 12 years, by 2030, there will be 1 trillion chips.
ARM chips will be scattered all over the Earth, and they will be connected in a network. Furthermore, these will begin to play an active role as edge devices for artificial intelligence. ARM, through a project called Gilliam, has decided to equip almost all of its chips with AI functionality and has already begun supplying them.
This means that, on the device side, people's movements will be monitored. For example, if you press a button on your phone, your ID will be verified below. Or, even without pressing a button, just by looking at your face, it will recognize your face as the owner of the smartphone and use that as your ID.
This is also artificial intelligence. All of this is currently being done with ARM chips. Or, Siri. It speaks to you, its backside recognizes your voice and tries to understand it. All of this is thanks to ARM's checks, which are currently working hard inside your smartphones. The next project will be to further evolve the AI functionality and make it easier to utilize its capabilities.
This will first be installed in A-series chips like those in smartphones, but this AI functionality will be installed in tiny devices that go into IoT, like the M-series. In other words, it will be installed in various surveillance cameras, sensors that detect various things in Rolo, drones that fly in the sky, and so on.
There will be a more detailed explanation after the song, but ARM chips will also be increasingly used on the H side in ride-sharing and autonomous driving. This is not just the chip, but also the software platform, such as Surfloader and Cafe, and so on. ARM creates the software and layers that connect these things to ARM chips, and collaborates with companies like Nvidia, Google, and Amazon, so that ARM chips can perform AI.
So, they're working to ensure that each software platform works well together. And on the device side, the fact that almost all of these chips will be ARM is a key point, in my opinion. So, the computing power of AI on the H side and the proliferation of 1 trillion AI chips on the H side will become extremely important.
Furthermore, the cloud is also extremely important. On the cloud side, GPU capabilities are central, and it is predicted that the computing power per chip will increase by about 200 times by 2030. Of course, this is the power per chip, but the number of chips will also increase rapidly on the cloud side, and the AI chips on the H side and the ARM chips on the H side will be organically connected at extremely high speeds, such as 5G, 6G, and 7G.
As communication continues to evolve, the evolution of AI will be at an incredible pace, like a quadratic curve from now on. As I said at the beginning, we don't know what the future holds. Because we don't know, let's stop thinking about it and just live fully in the present. I have doubts about this way of thinking, but without a doubt, AI will increase its power exponentially.
Without a doubt, the edge chips will increase. At least the people inside the arms are assuming, designing, and planning their business so that our arms will be shipped. I think that is an important fact. In other words, if AI's capabilities evolve to that extent, then all industries will be redefined.
For example, there are discussions where scholars at Osford University have made predictions about what jobs will be overtaken by AI and when. They write 2024, 2027, and 2030 respectively, but whether it's 23 years too early or too late, or whether it's really not overtaken, is in my opinion an error. It's 23 years too early or too late.
Comments 0