Have you ever had to sell something you absolutely loved just to chase an even bigger dream? That’s pretty much what happened to one of the most legendary investors alive.
Last month the tech world did a collective double-take when it learned that SoftBank had quietly unloaded its entire position in Nvidia – a move that brought in close to six billion dollars. For most people that sounds like the smartest trade of the decade. Nvidia has been the poster child of the AI boom, after all. But the man behind the decision says he hated every second of it.
He was, in his own words, crying when the sale went through.
The Billionaire Who Didn’t Want to Let Go
Few people in finance have swings as dramatic as Masayoshi Son. The SoftBank founder once turned 20 billion dollars into 100 billion during the dot-com era, then lost almost all of it when the bubble burst. He’s the guy who put 100 million into a tiny company called Alibaba back in 2000 and watched it become one of the biggest paydays in venture history.
So when he says he was emotional about selling Nvidia, you listen.
Speaking at a major conference in Tokyo, he was surprisingly candid. He didn’t want to sell a single share, he told the audience. The only reason he did it was because he needed the money for something he believes will dwarf even Nvidia’s incredible run.
I don’t want to sell a single share. I just had more need for money to invest in OpenAI and other projects. I was crying to sell Nvidia shares.
Masayoshi Son
That quote stopped me in my tracks the first time I read it. This is a man who has made and lost fortunes most of us can’t even imagine, and he’s openly admitting tears over a stock sale. That’s not normal billionaire behavior.
What Could Possibly Be Bigger Than Nvidia Right Now?
Simple answer: the next wave of artificial intelligence infrastructure and the companies building actual super-intelligence.
The Nvidia money didn’t go into some diversified basket of blue-chip stocks. It went straight into two massive priorities:
- A significantly larger bet on OpenAI
- Gigantic data-center projects that will power the coming generation of AI models
Think about that for a second. Nvidia makes the picks and shovels of the current AI gold rush – the GPUs everyone needs. But Son is effectively saying the gold rush itself is only the opening act. He wants to own the mines, the railroads, and eventually the entire new economy that artificial general intelligence will create.
In my view, that’s either genius or the kind of overreach that has burned him before. Maybe both.
The OpenAI Bet Is Already Paying Off – Massively
Here’s the part that makes the tears almost funny in hindsight.
Just weeks after the Nvidia sale, SoftBank reported its latest quarterly results. Net profit more than doubled to roughly 16.6 billion dollars. The biggest driver? Valuation gains on its OpenAI holdings.
So the shares he cried over selling helped fund a position that is already printing money at a historic rate. That’s the kind of irony only the venture capital world can produce.
Sources close to the company say SoftBank could actually increase its OpenAI stake further if future funding rounds come at attractive valuations. In other words, they’re not done yet.
Data Centers: The Invisible Empire
Most of the coverage has focused on the OpenAI angle, but I think the data-center push might actually be the sleeper story here.
Training frontier AI models now costs billions of dollars and requires facilities that consume power on the scale of small cities. The companies that control that infrastructure will have enormous leverage in the decades ahead.
SoftBank isn’t just investing in someone else’s data centers. They’re building their own at scale and partnering on projects that sound almost science-fiction – think “Stargate” level supercomputer clusters.
They also recently bought Ampere Computing, a chip designer focused on energy-efficient processors perfect for – you guessed it – massive AI data centers.
It’s a full-stack play: chips, power, real estate, cooling, networking, and the software layer on top. Very few players have the balance sheet to even attempt something this ambitious.
The AI Bubble Debate Just Got Personal
Everyone and their brother is arguing about whether we’re in an AI bubble. Valuation debates rage daily. And here comes Son with the most Son response possible.
People who talk about the bubble are not smart enough.
Masayoshi Son (paraphrased)
He went further, predicting that “super AI” and AI-powered robotics will eventually add at least 10% to global GDP. That’s tens of trillions of dollars in new economic value.
When you frame it that way, today’s multi-trillion-dollar investments look almost conservative. The question is whether we get there in five years or fifty – and whether the companies spending the money today survive long enough to collect.
History says most of the pioneers die broke while the infrastructure owners get rich. Son clearly intends to be the infrastructure owner this time.
Why This Feels Different From the Vision Fund Disaster
A lot of people hear “Masayoshi Son massive AI bet” and immediately flash back to the Vision Fund’s hundred-billion-dollar catastrophe. WeWork, anyone?
Fair concern. But there are some crucial differences this time around.
- He’s concentrating on infrastructure and foundational AI, not consumer apps with questionable unit economics
- The bets are more synchronized – everything feeds into everything else
- He actually has real profits flowing in now instead of burning cash on hype
- The technical progress in AI is measurable and accelerating, not theoretical
None of this guarantees success, of course. But it’s not the same movie with different actors.
What This Means for Regular Investors
If you’re sitting there wondering what Son’s moves mean for your own portfolio, here’s my take.
First, it’s another data point that the smartest (and craziest) money in the world still believes AI is in the early innings. When someone who could retire to his own island tomorrow instead chooses to bet the house again, that’s worth noting.
Second, the Nvidia sale actually created buying pressure in a weird way. Institutional investors now know there’s a massive player who would have happily held those shares forever if he didn’t have even bigger plans. That’s about the strongest vote of confidence you can get.
Third, pay attention to the picks-and-shovels beyond just GPUs. Power generation, cooling tech, specialized real estate, copper, optics – the list of second-order plays is long and still underappreciated.
Finally, remember that Son has been early before – sometimes catastrophically early. Being right eventually doesn’t always mean being right on your preferred timeline.
The Human Side of Billion-Dollar Bets
I keep coming back to that image of one of the world’s richest men literally tearing up over a stock sale.
In an industry that often feels cold and calculated, it’s a reminder that even at the highest levels, these are still human beings making enormous decisions with real emotions attached.
Son didn’t cry because he thought Nvidia was going down. He cried because he genuinely loves what that company represents and hated parting with it. The fact that he did it anyway tells you everything about how strongly he believes in what comes next.
Whether history remembers this as visionary or delusional probably won’t be clear for another decade. But one thing is certain: very few people on Earth are willing to make bets this large with this much conviction.
And honestly? In a world that increasingly feels short-term and risk-averse, there’s something almost refreshing about watching someone go all-in on a future most of us can barely imagine.
Even if it makes him cry.
