SoftBank Wants $100 Billion to Buy Companies and Automate Them

Masayoshi Son has a new pitch, and it is not about finding the next startup. According to the Financial Times, as relayed by Bloomberg and others, SoftBank's founder is seeking as much as $100 billion from Gulf investors for a vehicle that would buy established companies and apply AI and machines to their operations.

Son has held early talks with potential backers, including investors in the UAE. SoftBank declined to comment. No commitments have been confirmed.

Not a Vision Fund

The scale invites comparison with 2017, when Son raised close to $100 billion for the first Vision Fund, mostly from Saudi Arabia's Public Investment Fund and Abu Dhabi's Mubadala. The strategy is close to the opposite.

The Vision Funds bought minority stakes in fast-growing startups and waited for them to become giants. This vehicle, as described, would acquire companies that already work and change how they operate. It is closer to private equity with an automation thesis than to venture capital. One account suggests targets could include profitable mid-sized firms in manufacturing, logistics and financial services, though that reading comes from a separate outlet rather than the FT itself.

The key operating asset is Roze, SoftBank's robotics arm, which reports say is expected to play a central role. SoftBank also plans to list Roze in the United States.

The theory

The logic is straightforward, and it is the logic many AI investors have been circling for a year. Selling AI tools to incumbents is slow: every customer has to integrate, retrain staff and decide how much to change. Owning the incumbent removes the sales cycle. If AI and robots can lift margins in a warehouse, a factory floor or a back office, the owner captures that gain directly instead of charging a subscription for it.

It also gives SoftBank a place to deploy its own technology. A fund that owns dozens of operating businesses is a captive customer for Roze robots and for the AI models SoftBank has backed.

The balance sheet behind the pitch

SoftBank comes to this with very large exposure already on its books. It has invested almost $65 billion in OpenAI, which is reportedly seeking a further $30 billion at a valuation of around $1.4 trillion. Its SB Energy unit is building an 8.8-gigawatt data center in Ohio and preparing a Nasdaq listing. SoftBank also faces rising borrowing costs.

That context showed up in the stock this week. SoftBank shares fell as much as 7.3% in Tokyo on Friday on concerns about OpenAI's revenue growth, after a report that OpenAI's September annualized revenue was nearing $50 billion rather than the roughly $70 billion that had circulated. For a company whose value increasingly tracks one AI lab, a new pool of outside capital aimed at a different strategy is also a diversification argument.

Why the Gulf, and why it is harder now

Gulf sovereign and quasi-sovereign funds remain the deepest pools of patient capital for projects this size, and Son has a long relationship with them. But the region he is courting is under economic pressure from the ongoing Iran war, with oil above $100 a barrel this week and regional risk elevated. Big checks are still possible. Fast ones are less certain.

Gulf investors are also more sophisticated counterparties than they were in 2017. Many now run their own technology investment arms and demand co-investment rights, local operations or both. A $100 billion vehicle that buys companies worldwide may need to show what it builds in the region, not just what it returns.

What to watch

The first anchor commitment. A fund at this scale usually needs one or two cornerstone investors before the rest follow. Until a named backer appears, $100 billion is a target.

The target list. Mid-sized, profitable, operationally heavy businesses are the obvious fit. The more specific the first acquisitions, the easier it will be to judge whether the automation thesis works outside a pitch deck.

Roze's listing. Taking the robotics arm public in the U.S. while also making it the engine of a buyout fund creates both a showcase and a potential conflict. Investors in each will want to know who gets the economics.

The read

The Vision Fund was a bet that the future would be built by startups. This fund, if it closes, is a bet that the future will be retrofitted into companies that already exist, by an owner who controls the robots and the models. It is a more operational, more grounded idea than Son's last $100 billion pitch. It is also asking for a great deal of money at a moment when his biggest existing bet is under scrutiny.