Nvidia (NASDAQ: NVDA) has partnered with six of the world’s largest financial institutions to mobilize more than $500 billion in capital for artificial intelligence infrastructure.
The partnerships involve Apollo Global (NYSE: APO), BlackRock (NYSE: BLK), Blackstone (NYSE: BX), Brookfield (NYSE: BAM), Goldman Sachs (NYSE: GS), and KKR (NYSE: KKR), creating new financing platforms for prospective AI customers.
The move is designed to help smaller enterprises and AI labs afford Nvidia systems, which carry price tags that have historically put them out of reach for many buyers.
Nvidia itself may backstop up to $125 billion, representing 25% of deals facilitated through the new financing arrangement.
CEO Jensen Huang described the initiative in clear terms, stating: “We are helping create a new class of productive, investable infrastructure: AI factories.”
Huang added: “These financing platforms will help customers access scarce compute at scale and build the AI factories that will power every industry and country in the age of AI.”
By removing the financial barrier for smaller customers, Nvidia effectively expands its total addressable market at a time when AI infrastructure demand continues to accelerate globally.
A separate development is also reshaping how analysts think about the long-term value of Nvidia hardware, specifically the useful lifespan of its GPU products.
Neocloud CoreWeave (NASDAQ: CRWV) recently signed a contract to rent Nvidia A100 GPUs, which were introduced in 2020, through 2029, suggesting these chips last far longer than previously assumed.
Analyst Kyle Reidhead recently explained the dynamic during an interview with Schwab Network, noting the extended lifecycle of Nvidia GPUs across multiple use cases.
“What we are realizing is that the GPUs are used first for training on frontier models for about two or three years, then you get another three years on inference, and another three, four, or five years on niche training and niche work for enterprises,” Reidhead said.
This stands in contrast to the assumption held by some market participants, including notable short-seller Michael Burry, that server hardware useful life runs only two to three years.
Burry, best known for shorting subprime mortgages ahead of the 2008 financial crisis, currently holds short positions in Nvidia and other AI companies based partly on that depreciation argument.
If the actual useful life of Nvidia chips is closer to nine years, as the CoreWeave contract implies, hyperscalers may actually be understating their return on AI infrastructure investment.
Longer chip lifespans could also strengthen Nvidia’s pricing power, since customers would theoretically be willing to pay more for a product that delivers value over a longer operating period.
Nvidia shares have surged more than 1,300% since January 2023, yet most Wall Street analysts continue to view the stock as undervalued relative to its growth trajectory.
Wall Street currently estimates Nvidia’s earnings will grow at 44% annually over the next three years, making its current valuation of 32 times earnings appear modest by growth standards.
As of August 24, the median 12-month price target among 67 analysts tracked by The Wall Street Journal stands at $300 per share, implying 44% upside from the current price of $208.
