SoftBank Group executed the first $10 billion tranche of its $30 billion OpenAI commitment on April 2, routing the capital through Vision Fund 2. The deployment, confirmed via Reuters regulatory filing and reported by HedgeCo, marks the first time a significant portion of OpenAI’s record $122 billion raise has moved from commitment to actual capital deployment. Two more $10 billion tranches are scheduled for July and October 2026.
The Structure Behind the Number
SoftBank’s $30 billion commitment is part of the broader $122 billion round that included $50 billion from Amazon (starting with $15 billion upfront) and $30 billion from Nvidia in GPU capacity and infrastructure commitments. Techi’s breakdown of the round notes that as of late March, OpenAI had received approximately $25 billion in immediate capital: $15 billion from Amazon and now $10 billion from SoftBank’s first tranche.
The money is financing the deployment itself. SoftBank’s stated thesis centers on “AI-native infrastructure: compute, data, and distribution,” according to HedgeCo. To fund its OpenAI position, SoftBank took on a $40 billion bridge loan disclosed last week, a financing structure that TechCrunch noted signals SoftBank is betting on a near-term OpenAI IPO to refinance the debt.
Why This Matters for Agent Infrastructure
For anyone building or deploying autonomous AI agents, the compute layer is the constraint. Training frontier models, running inference for agent orchestration at scale, and maintaining the data pipelines that feed real-time agent decisions all require infrastructure investment that dwarfs traditional SaaS capital requirements. SoftBank’s $10 billion is the first check clearing for that build-out. The scheduled cadence of the remaining tranches (July and October) means the infrastructure expansion will accelerate through the second half of 2026, directly expanding the compute capacity available to agent platforms and their users.