Key Takeaways
- Goldman Sachs is discussing the financing structure with banks, insurers, asset managers and private credit firms.
- The bank could provide junior capital and private credit for Nvidia AI projects.
- Nvidia may backstop up to 25% of individual financing deals.
Goldman Sachs has begun talks with institutional investors after securing a central financing role in Nvidia’s planned AI infrastructure program.
Goldman targets institutional investors
Goldman Sachs is speaking with a broad group of potential investors as it works on the financing structure. The discussions include U.S. insurers, asset managers, banks and private credit firms that could fund projects tied to AI computing infrastructure, a financing approach similar to Meta’s AI data center financing package with PIMCO and Blue Owl. Asset managers are expected to retain a meaningful share of the financing.
The effort moves the focus from Nvidia’s partnership announcement to the work needed to turn the plan into investable financing. Goldman is assessing different forms of capital and how private investors can take part in the transactions.
Goldman’s asset management business can provide junior capital and private credit. Its investment bank can also help place debt with private credit funds and later distribute it through public debt markets.
Goldman’s role extends beyond arranging funding
Goldman secured a central lending role in the financing structure, reflecting its long relationship with Nvidia. The bank has advised the chipmaker on several transactions and has worked on technology financing deals involving Nvidia.
The relationship also includes major capital markets work. Goldman was among the lead underwriters for Nvidia’s $25 billion bond sale in June and advised the company on its $6.9 billion acquisition of Mellanox Technologies in 2019.
The latest assignment gives Goldman a chance to combine several parts of its business. Its investment bank can arrange and distribute debt, while its asset management arm can participate directly through private credit and junior capital.
Financing model aims to broaden the investor base
The proposed structure could create a wider market for AI computing infrastructure rather than relying mainly on traditional corporate borrowing. The plan seeks to package financing around AI compute assets so that debt can reach a broader group of investors.
Nvidia has said it may backstop up to $125 billion, or 25% of potential deals. That would leave the majority of the financing with the banks, asset managers, insurers and other investors taking part in the structures.
The approach also reflects the growing need for private capital as companies expand data centers and computing capacity. Goldman Sachs analysts have estimated that the four largest cloud companies could spend more than $5 trillion on technology and data centers through 2030, increasing demand for financing beyond traditional bank balance sheets.
For Goldman Sachs, the immediate task is to build an investor base capable of supporting these transactions. The talks mark an important next step, as reported by Reuters, in efforts by financial institutions to develop a repeatable funding model for large-scale AI infrastructure projects.
