Nvidia is the central bank of AI
Summary
Financial engineering drives Nvidia's growth.
Jensen Huang uses guarantees and income top-ups to help customers finance data centers, such as a $105bn backstop for an Ohio facility.
Nvidia agreed to provide a backstop worth up to $105bn
Custom chips threaten Nvidia's dominance.
Hyperscalers like Google and Amazon design internal silicon that costs one-fifth to one-third of Nvidia's hardware, potentially capturing 50% of the market by 2030.
Nvidia subsidizes the AI ecosystem.
The firm takes equity stakes in startups and invests in open-weight model platforms like Hugging Face to create customers independent of tech giants.
The intention behind such investments is to fuel demand for Nvidia’s chips
Neocloud backstops lower borrowing costs.
Nvidia promises to floor neocloud revenues for up to six years, making compute an investable asset class for lenders.
Slow demand triggers massive liabilities.
If AI growth disappoints, Nvidia may be forced to honor $300bn in potential liabilities, including buying unused compute or covering price shortfalls.
merely disappointing growth could conceivably spell trouble
Highlights
Mr Huang frames this as a way to make compute “an investable asset class”.
Some have taken to calling Nvidia the “central bank of AI ”
the more of this boom they finance, the more of any bust they will have to bear.
Nvidia is walking a fine line between “enabling demand” and “creating it”.
Read the original ↗Save to your Sponge
Summarized with Sponge. Quotes are the author’s own words.