Masayoshi Son has gone back to the place that funded his last big gamble. In recent weeks the SoftBank founder has been in contact with investors in the Gulf, including the United Arab Emirates, to raise as much as $100 billion for a new push, according to a Financial Times report that Reuters and Bloomberg later matched.
The money would be used to acquire companies and remake their operations with artificial intelligence and robotics, the report said. SoftBank declined to comment. The plan is the latest in a long line of enormous wagers from a man who has spent decades betting that technology would reshape whole industries, sometimes correctly, sometimes not.
The timing is pointed. Last week SoftBank completed the final tranche of a $30 billion investment in OpenAI, the largest single outside bet on the AI lab that now anchors Son’s portfolio. The Gulf has already been the deepest well for that kind of capital, with sovereign funds financing AI and data-center projects around the world.
The Gulf relationship runs deep. Saudi Arabia’s Public Investment Fund committed $45 billion to SoftBank’s first Vision Fund in 2017, the single largest pledge in a vehicle that raised about $100 billion. That fund made its name on Alibaba and nearly lost it on WeWork, and the memory of both outcomes still shapes how Gulf investors weigh Son’s next pitch.
Son’s pitch, in essence, is that AI will not just build new companies but transform old ones, and that SoftBank is the vehicle to buy those companies and install the technology. It is an expensive thesis, which is why he is back in the Gulf asking for the money.
The Gulf’s sovereign funds have reason to listen. Abu Dhabi and its neighbors have spent years trying to convert oil wealth into a diversified portfolio, and AI has become the preferred destination. The Emirates launched a dedicated AI investment vehicle, and Gulf money has flowed into data centers, chip ventures and the labs themselves. A $100 billion commitment to SoftBank would be large even by their standards, but the relationship is already warm.
Elon Musk greeted the news with a familiar jab, writing on X that “only a fool” would give SoftBank the money. Musk and Son have a history, and Musk’s own AI and robotics ambitions make him a direct competitor for the same capital and the same talent.
The contrast between the two men is instructive. Musk builds companies and spends his own equity to fund them. Son raises other people’s money and places bets. Both are now competing for the loyalty of the same Gulf investors, and both need the same thing: the belief that AI will repay whatever is poured into it.
SoftBank’s record cuts both ways. Its early bet on Alibaba made its reputation, and its Vision Fund’s bet on WeWork nearly destroyed it. Son has spent the years since trying to prove the second outcome was an aberration, and his OpenAI stake is the centerpiece of that argument.
The new fund, if it comes together, would mark a return to Son’s favorite scale. A hundred billion dollars is more than the first Vision Fund raised in outside commitments, and it would position SoftBank to buy companies outright rather than take minority stakes, changing the shape of its investments.
For the rest of the AI industry, a SoftBank war chest of that size would be a signal as much as a financial event. It would confirm that the capital flowing into AI is still accelerating, and that the largest pools of money in the world have concluded the boom is not yet near its peak.
There is a second meaning in the destination. Son is not courting Silicon Valley’s limited partners or Japan’s banks this time. He is courting sovereign wealth, the only pools deep enough to write checks at this scale, and the only investors with the patience to wait out a technology cycle that could take a decade to pay off.
Son has been making the case for years that artificial general intelligence, and eventually artificial superintelligence, will arrive within his lifetime, and that the companies controlling it will be worth more than anything that has come before. The OpenAI stake is the financial expression of that conviction, and the new fund would be its operating arm, buying the businesses AI will remake.
The scale of the ask also reflects the scale of the infrastructure now required. Training and running frontier models demands tens of billions of dollars in chips, power and data centers, and no single balance sheet can carry that alone. Son is positioning SoftBank as the bridge between the Gulf’s capital and the companies building the machines.
Whether the Gulf writes the check is an open question. The talks are early, and $100 billion is not a number anyone commits lightly. But Son has asked for impossible sums before and, more than once, the money has shown up.
