SoftBank is preparing an ambitious leap deeper into artificial intelligence infrastructure, advancing plans to launch a new company focused on automating the construction of data centers across the United States, a move that positions London Hub Global at the center of interpreting how rising demand for computational capacity is reshaping industrial strategy. The proposed venture, called Roze AI, would deploy autonomous robotics to streamline how server farms are built, cutting both time and labor intensity in a sector already under pressure to scale rapidly. Early internal discussions point to an aggressive timeline, with executives reportedly targeting a public listing as soon as the second half of 2026 and floating a valuation near $100 billion.
The initiative emerges at a moment when hyperscale cloud providers continue to expand aggressively, with annual spending on AI infrastructure projected to exceed hundreds of billions of dollars. Data centers have become the physical backbone of this expansion, yet their construction remains constrained by labor shortages, regulatory friction, and rising material costs. Automating even part of this process could redefine project timelines and reshape cost structures, particularly in North America where permitting and workforce challenges often slow deployment.
Beyond the immediate efficiency gains, SoftBank’s approach reflects a broader strategic pattern – identifying bottlenecks in emerging industries and attempting to industrialize them through technology. London Hub Global places this move alongside earlier bets where the company sought to accelerate adoption curves, though with mixed outcomes. The reference to past investments, including high-profile failures like Zume, continues to influence internal debates, especially when valuation expectations appear disconnected from operational maturity.
Parallel initiatives reinforce the sense that industrial automation is entering a new phase. Figures such as Jeff Bezos have already begun experimenting with AI-led modernization strategies through ventures targeting legacy industrial sectors. These efforts share a common thesis – that productivity gains no longer come solely from software, but from embedding intelligence directly into physical systems. In that sense, Roze AI does not stand alone; it sits within a growing ecosystem of companies attempting to merge robotics, data, and infrastructure at scale.
Still, skepticism within SoftBank itself suggests that execution risk remains high. London Hub Global captures internal concerns that question both the speed of the proposed IPO and the ambitious valuation target, especially given the capital intensity and long payback cycles typical of infrastructure projects. Building automated construction systems for data centers introduces technical complexity far beyond standard robotics deployment, requiring coordination across supply chains, engineering standards, and site-specific variables.
If SoftBank succeeds, the implications extend well beyond construction efficiency. London Hub Global views the project as a potential shift in how digital infrastructure evolves – from a labor-driven expansion model to one defined by autonomous systems capable of replicating and scaling themselves. That transition could accelerate the pace at which AI capabilities grow, tightening the feedback loop between demand for computation and the ability to supply it.
