SpaceX seeks $40 billion financing for Nvidia chip purchase
publishers 17articles 19first reported 6 Oct, 22:28 UTCdeveloping since 7 Oct · 10 editions
SpaceX is in talks with Apollo and several banks to arrange about $40 billion in financing for a purchase of Nvidia graphics processing units, according to reports by the Financial Times, Bloomberg and CNBC.
CNBC, citing a person familiar with the discussions, reported that the financing would rely primarily on investment-grade debt and would likely use the GPUs as collateral. Apollo is expected to play a leading role in arranging the deal. Apollo declined to comment to CNBC, while Nvidia did not immediately respond to the network’s request for comment.
The talks come months after SpaceX tapped debt markets following its June initial public offering. CNBC reported that the company raised $25 billion through bonds of varying maturities two weeks after the IPO, with strong investor demand at the time. Since then, yields on SpaceX debt and other bonds linked to artificial-intelligence investment have risen as credit spreads widened.
A credit-industry source told CNBC that the market currently assumes high-performance GPUs will retain value for about seven years, supported by scarce computing capacity. The proposed financing follows large debt offerings by technology companies including Meta, Amazon and Google to fund AI infrastructure.
HOW THIS STORY WAS MADE
Written from 7 articles, 1 with the publisher's own text; 7 independent newsrooms once syndicated copies count as one; this version written 12 h after the record first saw the story; the editor kept it as written.
- Publishers
- 17
- Source articles
- 19
- Given to the writer
- 7, 1 with the article's own text
- Independent newsrooms
- 7
- This version written
- 12 h after the record first saw the story
- Second model (editor)
- kept as written
- Publication gate
- passed
- Human review
- none
WHAT THE EDITOR DID
The second model checked the writer's draft against the same sources and published it as written.
Written by a language model from the sources above, then checked by a second model that may only cut, attribute or correct. How it works →