SoftBank Wants $100 Billion to Buy Boring Companies and Make Them AI Smart

The biggest bet in technology is about to get bigger. Masayoshi Son, the SoftBank founder who once bankrolled WeWork and Arm, is seeking to raise up to $100 billion from Gulf investors for a new wave of AI investments, the Financial Times reported on October 9, in a story carried by Reuters.

Son has held discussions in recent weeks with senior figures, including in the United Arab Emirates, about the potential fundraising, the report said, citing people familiar with the matter. Reuters said it could not independently verify the report, and SoftBank did not respond to a request for comment.

If the fund materializes, it would be one of the largest single fundraising efforts in the history of technology investing, and it signals something important: the AI boom is entering a new phase. The era of simply financing AI model companies may be giving way to something stranger and more ambitious.

A different playbook

Unlike the Vision Funds, which backed startups like WeWork, Uber, and DoorDash with famously aggressive checks, the proposed vehicle would reportedly acquire established businesses and use AI to improve their operations. That is a different theory of value creation: not betting that startups will invent the future, but betting that AI can squeeze new efficiency out of the present.

The logic is easy to see. A decade of enterprise AI hype has left most old-economy companies, in logistics, manufacturing, retail, and professional services, only partially transformed. Buying them outright, then running their operations through modern AI systems, could unlock gains that incremental consulting projects never reach. Whether SoftBank can execute that at scale is a different question entirely.

The AI boom is entering a new phase. The question is no longer which startup wins the model race, but what happens when AI gets deployed inside the old economy, at scale.

The context: an arms race in capital

The proposed $100 billion fund does not exist in a vacuum. Last week, SoftBank completed its $30 billion investment in OpenAI as part of the ChatGPT maker's last fundraising round. A month earlier, SoftBank raised $11.1 billion in what was billed as the largest high-yield corporate bond sale in the world, specifically to fund its bet on OpenAI.

And Son is not alone in thinking at this scale. SpaceX is reportedly pursuing approximately $40 billion in financing for Nvidia AI chips, roughly $30 billion in bonds and $10 billion in bank loans. Broadcom and Oracle are involved in similarly large financing arrangements tied to AI computing demand. Morgan Stanley estimates AI infrastructure could require approximately $1.5 trillion in external financing by 2028.

The pattern is clear: the AI boom is becoming a capital markets story as much as a technology story. Money is being raised through bonds, bank loans, sovereign wealth, and equity alike, all converging on the same bet, that demand for AI compute will keep growing for years.

Why the Gulf

The Gulf connection is not incidental. Sovereign wealth funds in the UAE, Saudi Arabia, and Qatar have been among the most aggressive investors in AI infrastructure in the world, and they have a structural reason to be: many see AI as the next long-dated asset class, the thing that generates returns for decades after oil. Son has a history of raising mega-funds with Middle Eastern money; the first Vision Fund, launched in 2017 at $100 billion, drew major commitments from Saudi Arabia and the UAE.

That history is both an asset and a warning. The Vision Fund era produced spectacular wins, like Arm, and spectacular losses, like WeWork. The new fund's strategy, buying established companies rather than speculative startups, reads like a lesson learned from that era.

Reasons for caution

It is worth stating plainly what is and is not confirmed here. The fundraising discussions remain preliminary, according to the reporting, and Reuters could not independently verify the Financial Times account. A $100 billion target is an aspiration until it is a signed commitment.

There are also signals that investors are growing choosier. This week, Firmus Technologies, an Australian AI infrastructure company backed by Nvidia and Blackstone, withdrew plans for a $5 billion IPO targeting a valuation above $30 billion, after investors balked. Reuters reported that only 42 megawatts of its planned 1-gigawatt capacity was operational. The gap between present business and projected growth is becoming the central question of the AI investment era.

That tension applies to Son's plan too. Acquiring established businesses and transforming them with AI sounds cleaner than betting on moonshots, but operational transformation is slow, messy, and hard to scale. The history of private equity rolling up old-economy companies is a history of mixed results, even before adding AI hype to the mix.

What it means

Still, the signal is real. A $100 billion fund aimed at AI-transforming established companies would be the clearest indication yet that the center of gravity in AI investing is shifting from building models to deploying them, from the lab to the warehouse, the call center, and the back office.

If Son pulls it off, the next era of AI will not be defined by whoever trains the smartest model. It will be defined by whoever owns the boring companies that AI makes smarter. That is a very different industry, and a very different set of winners, than the one venture capital has been betting on for the last decade.