Trusting Masayoshi Son is less reliable than trusting Elon Musk.

On June 23rd, at SoftBank's annual general shareholders' meeting, Masayoshi Son poured cold water on Elon Musk's idea of a space data center. He said that moving data centers to space "doesn't make much sense," and the outcome of the AI competition will be determined by computing power on Earth. Son's reasoning was quite clear. Electricity costs only account for about 7% of a data center's operating costs. The major cost is chips, and when you add up launch fees, in - orbit maintenance, and communication delays, it just doesn't add up.

He might have done the math right, but the problem is that Son is still just doing the math.

Let's talk about SoftBank in recent years. In the investment circle, Masayoshi Son's persona has been pretty much set in stone: the industry's jinx, the master of missing out, and a lifelong consumer of the "pill of regret." The spectacular failure of WeWork, the defeat of the Vision Fund, selling all of NVIDIA shares in 2019 and missing out on hundreds of billions of dollars in potential earnings, and fake - crying in Huang Renxun's arms at the Tokyo Artificial Intelligence Summit in November 2024...

My colleague Pu Fan wrote an article at that time titled "After Crying in Huang Renxun's Arms, Masayoshi Son Invested Another $10 Billion." It evaluated Son as someone who "finds missing out more painful than losing money." This sentence can also summarize the psychological motivation behind all of Son's actions in recent years. In an exclusive interview with Son in June 2023, he said he cried for several days, reflecting on whether he was a qualified entrepreneur. Finally, he figured it out and believed that he was an "architect" who wanted to "sculpt the future of humanity," and that future is AI.

Then Son turned around and poured tens of billions of dollars into OpenAI, and now he has become the second - largest external shareholder of OpenAI.

A person who calls himself an "architect of humanity's future" says, when facing another person who is actually trying to send humans to Mars, that "space is meaningless, and the AI victory will be decided on Earth." I, for one, think the gap is quite large.

Let's start with the term "architect." Although I don't know much about architecture, I'm also not sure if Son really thought about the weight of this word when he said it. So I asked an AI to search for famous architects and buildings in the world:

Tadao Ando designed the Church of the Light. He cut a cross - shaped slit in the plain concrete wall, making natural light the embodiment of faith. That building carries the idea that "light is sacred."

Antoni Gaudí spent more than forty years on the Sagrada Familia. He translated the columns of the forest, the vaults like bones, and the colored glass like the ocean into praise for creation. The building has been under construction for more than one hundred and forty years and is still unfinished. That building carries the idea that "nature is the language of God."

Le Corbusier designed the Notre - Dame du Haut. With its concrete roof curved like a sail and colored light windows of different sizes, he translated religious experience into a modern - space poem. That building carries the philosophy that "space is experience."

So I think architects and their buildings should first carry ideas. There should be a judgment about "how people should live" in them, and there should be people and their thoughts. So architects are artists. Mu Xin said, "Artists carry real things in fake pockets." The same goes for architects. The form is the pocket, and the idea is the real thing.

What about the "future" that Masayoshi Son is sculpting? Anyway, you can't see people or ideas in it. It's just a bunch of data centers and perhaps a nicer - looking financial report. If you want to compare capital with art, you should at least make it look a bit more like it.

My colleague Master Dong wrote in "Is There Truth?" "Believe in technology and gain immortality; believe in money and gain freedom." He said that there are two types of people in the world. One type believes that AI will liberate humanity, and the other type believes that making money will liberate themselves. Elon Musk belongs to the former. No matter how you evaluate him, he is at least trying to let technology carry something beyond capital.

What about Masayoshi Son? He says he wants to "sculpt the future of humanity," but all he does is "believe in money and gain freedom." He invested in OpenAI not because he has a vision for humanity's future, but because "missing out is more painful than losing money." His $60 billion investment has nothing to do with faith or humanity. It's just an indulgence for his fear of missing out (FOMO).

So Masayoshi Son's self - positioning as an "architect of humanity" is, more accurately, that of a more pure - bred traditional capitalist. Note that he is not a capitalist like those on Wall Street who use financial engineering to leverage. Son disdains such delicate work. He always bets all - in with a large - scale approach. But he shares the same underlying belief with Wall Street elites. Capital appreciation is the ultimate goal, and technology, narratives, and humanity's future all have to serve this goal.

I remember that Qin Xiao, the former chairman of China Merchants Bank, once said a sentence. The general idea is that China Merchants Bank introduced a valuation model in internal management. Maximizing shareholder equity is not as good as maximizing profits, and maximizing profits is not as good as maximizing valuation. Why did Masayoshi Son rant about Elon Musk at the shareholders' meeting? Actually, he was doing market - value management for SoftBank Mobile.

Elon Musk has said many "bold words" before, emphasizing that currency will become invalid in the future, which is almost like overthrowing capitalism. And Masayoshi Son's statement that "the victory will be decided on Earth" seems like a financial and industry judgment on the surface, but it's actually Son's instinctive defense. He has to keep AI on the ground and prevent it from escaping the framework of accounting, investment, and capital appreciation. Once the AI narrative goes to space with Musk, what's left for Son?

Of course, Son also invested in OneWeb, a British satellite - internet operator, which was once called a "strong rival to SpaceX," and he invested quite a lot of money, about $2 billion. But it eventually went bankrupt and was reorganized. Anyway, one person wants to transcend capital, and the other person wants to defend capital. You tell me, who is more like an "architect of humanity"?

Does OpenAI need space computing power? No. Sam Altman wants on - ground data centers, more GPUs, and cheaper electricity. Masayoshi Son's investment in OpenAI is, in essence, a bet on the on - ground computing - power route. This is a business he can understand and calculate.

But Elon Musk's space data center is not an isolated product concept. It is a part of SpaceX's entire closed - loop system. If you take the puzzle piece out and do the math, of course, it doesn't make sense, just like taking the engine out of a rocket and weighing it, and then saying it can't fly.

There is also a difference in the cognitive dimension here. Elon Musk once said a passage. The general idea is that all problems with action stem from a lack of accuracy. There are only 27 key steps in rocket manufacturing. When you improve the accuracy of your assessment of the objective world, things that seem out of reach will become simple. This is the thinking of an engineer, breaking down the unimaginable into executable steps.

Masayoshi Son's thinking is exactly the opposite. In his eyes, the future is a financial statement where he can calculate the ROI. One is about breaking down, and the other is about compressing; one turns the "impossible" into steps, and the other turns the "uncertain" into numbers. These two different ways of thinking determine that the two people see completely different futures.

It's not the first time that Masayoshi Son and Elon Musk have disliked each other.

At the beginning of 2025, for Trump's "Stargate" project, Masayoshi Son invested $500 billion to take a stance. Elon Musk publicly undermined it, saying that the project had no money. In June 2025, Masayoshi Son proposed a trillion - dollar robot plan, and Elon Musk said on Twitter that his nonsense had reached an "epic scale."

Masayoshi Son may not have changed and still takes the large - scale route. But look at where Elon Musk's thinking is at. Questions like whether humanity should become a multi - planetary species, whether AI will replace capitalism, and whether currency will disappear. Whether you think he's just making promises or not, when Musk goes against Son, it's definitely a high - dimensional attack on a low - dimensional one.

So if you want to challenge Elon Musk, you need to choose the right angle. If you want to deny him from a business perspective, it's a low - dimensional attack on a high - dimensional one, and it will make people think you're being overly attached. Even if you choose a good angle, like Master Dong, who from a moral and philosophical perspective, from a human - centered perspective, said that Musk's technological authoritarianism is worthy of vigilance, and that Musk uses technology to transform society but doesn't take responsibility for the consequences, he still got a lot of scolding, some of it quite harsh. For Masayoshi Son to try to challenge Musk by doing the math is like discussing the budget with a prophet.

Speaking of this, I remember a small thing. Around 2016, when the concept of large - scale models hadn't emerged yet, machine learning and reinforcement learning were popular in the industry. Masayoshi Son invested in a company whose technology was to wear a pair of glasses and scan a car's body and chassis to know its condition. He was so excited at that time and said he must invest.

At that time, Son was extremely willing to pay for technology. I even thought it was a bit ridiculous because the product seemed so unreliable. A person who was willing to pay for a pair of glasses to scan a car, but when facing Elon Musk, who is really trying to take humanity to space, says that "the victory will be decided on Earth." I think maybe Son is afraid of the inevitable struggle between capital and AI?

When I was a kid, I loved a book that was probably about the love story between a person from the tenth planet that disappeared in the solar system and an Earth - person. The book talked about how the moon was formed, how Saturn's rings were created, and why Uranus rotates on its side. I was so fascinated that I read it many times and wanted to be an astronaut. Later, I got nearsighted, and that dream faded away. Then, when super - rich people could go to sub - orbit, I no longer had that desire. But I still remember the feeling of being struck by the idea of space.

So I can understand why so many people have become Elon Musk's followers. Because no matter how big and illusory the promises he makes are, at least he can make people remember this feeling again. Master Dong wrote in "Is There Truth?" "The Thiel - Musk brothers are okay. They respect themselves, dare to speak their minds, and actually take action. The hypocrisy of capital is exposed. Capital doesn't allow for true theoretical argumentation. Capital has been surpassed by technology. Can technology carry the truth? No one knows. We'll find out when we get there."

Yes, we'll find out when we get there. Masayoshi Son can't say these four words in his lifetime. Because all of his "going" has to go through the accounting process first. He says he's an "architect" who wants to sculpt the "future of humanity." Don't even mention whether it comes from the heart; it doesn't even reach the kidneys. If I had to choose between the space - related promises that Elon Musk makes and the "future" that Masayoshi Son is sculpting, I'd choose Musk. At least he might fulfill a middle - aged person's childhood fantasy.

This article is from the WeChat official account "China Venture Capital News." Author: Zhang Nan. Republished by 36Kr with permission.