Sources: SoftBank and OpenAI's $100B Stargate investment hits snags on tariff fears; talks with dozens of lenders started earlier in 2025 have yielded no deals
SoftBank Group Corp.'s plans to invest $100 billion in artificial intelligence infrastructure in the US have slowed …
Context & Ripple Effects
Stargate followed reports that SoftBank was weighing a $15B-$25B OpenAI investment alongside an $18B Stargate commitment. This report shows the larger infrastructure ambition encountering a separate constraint: lenders have not agreed to finance it.
That matters because the initiative’s execution now depends not just on sponsor commitments but on whether its construction economics remain financeable amid tariff uncertainty.
First-order effects
- SoftBank and OpenAI face a slower path to advancing the $100B Stargate plan while lender discussions produce no deals.
- Tariff fears add uncertainty to the project’s cost assumptions, making it harder to settle financing terms for the proposed U.S. infrastructure buildout.
Second-order effects
- The absence of committed lending increases pressure on the sponsors to adjust project scope, timing, or capital structure before construction can proceed at the envisioned scale.
- It reinforces the execution gap between announced AI infrastructure capacity and deployable compute: capital commitments do not by themselves secure financed facilities.
Third-order effects
- If similar projects remain difficult to finance, AI infrastructure expansion will be shaped increasingly by access to durable financing and risk allocation, not only demand for compute.
- The pattern points toward a more selective buildout in which projects with clearer cost visibility and financing support move ahead first.
The trend: AI infrastructure is becoming a financing-and-execution challenge, as much as a race to announce larger compute investments.