Masayoshi Son, the SoftBank founder who bet billions on WeWork, poured money into Uber, and then staked his conglomerate’s future on Arm Holdings, is now creating a standalone AI and robotics company. The company will be listed on a US exchange, with a target valuation of $100 billion, before it has shipped a single product or published a revenue figure.
The new entity, to be called Roze, will focus on building data centres and using robotics to improve the efficiency of AI infrastructure construction. SoftBank declined to comment when contacted by multiple outlets.
Robots in the server room
Roze would seek to make data centre construction in the US more efficient by deploying autonomous robots to help build server farms. It is a striking concept. The construction of hyperscale data centres is one of the most labour-intensive undertakings in modern infrastructure, an activity that still depends heavily on skilled electricians, pipe fitters, and HVAC engineers.
The US construction industry was short approximately 439,000 workers as of November 2025, according to the Information Technology and Innovation Foundation, with skilled trades representing the most acute gap. Satellite imagery analysis suggests up to 40% of AI data centre construction sites face delays.
Roze is Son’s argument that the labour wall blocking the global AI buildout can be broken by machines rather than people. Instead of backing software alone, SoftBank is leaning into the infrastructure layer; land, power, machines, and the systems needed to bring massive AI facilities online faster.
The new company is set to include SoftBank’s recent $5.4 billion purchase of ABB’s robotics division, Ampere Computing for $6.5 billion, and DigitalBridge for $3 billion. SoftBank may also add more land, energy, and infrastructure assets, though the energy unit will stay separate. The plan, in other words, is to roll up a portfolio of hard assets and specialist capabilities under a single listed vehicle. and then let public markets decide what that combination is worth.
A $100 billion question mark
For context, Arm Holdings, a far more mature business with decades of revenue history and an installed base in virtually every smartphone on the planet, currently trades at a market capitalisation of roughly $145 billion. Pricing a company that does not yet exist at two-thirds of that figure is a bold pitch, even by the standards of the current AI investment cycle. Combining AI, robotics, and data centre development into one public company is a complicated idea for investors, who already find it hard to value established AI businesses. Roze has not yet announced a product, shared a revenue plan, or set an IPO date. For now, the company has a $100 billion target, an analyst day in July, and Masayoshi Son’s backing.
That analyst day, planned at a Texas data centre in July to boost interest for Roze’s IPO, is confirmed, with KPMG hired to prepare the necessary financials and documents for the public offering. Bilal Safeer, currently an executive at Arm, is expected to serve as interim CFO.
Son’s long game on AI infrastructure
The Roze announcement does not come in isolation. SoftBank is leading financing for the Stargate project to develop data centres across the United States, a planned $500 billion undertaking pursued alongside OpenAI and Oracle. The proposed listing could also help offset some of SoftBank’s major pledges, including more than $30 billion committed to OpenAI. Taken together, the picture that emerges is of a conglomerate that has placed an enormous directional bet on AI infrastructure and is now trying to monetise that bet through the public markets, reducing its own capital burden while retaining strategic control.
The deeper bet, as Son has structured it, is that the AI buildout has at least a decade left to run, and that whoever owns the construction layer captures recurring economics that the chip and model layers cannot. Whether public investors share that conviction, at a $100 billion entry point, for a company yet to turn a wheel, is the defining question.
Related stories
AI chip crunch continues, as Intel and AMD confront CPU delays
