CoreWeave, Nvidia sign $6,3 billion cloud computing capacity order
Sept 15 (Reuters) – CoreWeave (CRWV.O), has signed a $6.3 billion initial order with backer Nvidia (NVDA.O), the data center operator said on Monday, in a deal that guarantees that the AI chipmaker will purchase any cloud capacity not sold to customers.
Shares of CoreWeave rose 8% as the deal cements the company’s position as a key Nvidia cloud partner and cushions it against any potential decline in demand for AI computing capacity.
CoreWeave operates AI data centers in the U.S. and Europe, offering access to Nvidia’s GPUs, which are highly sought after for training and running large AI models.
The latest order, which requires Nvidia to purchase any unsold capacity through April 13, 2032, builds on an April 2023 agreement.
The amended agreement serves as a backstop for CoreWeave, ensuring that capacity will be leveraged irrespective of the end customer, analysts at Barclays said, adding that the incremental spending by Nvidia is seen as a healthy diversification away from the company’s largest customers.
We see this as a positive for CoreWeave given concerns from investors around the company’s ability to fill data center capacity beyond its two largest customers (Microsoft and OpenAI).
In March, CoreWeave and OpenAI agreed on a five-year contract worth $11.9 billion, under which the New Jersey-based firm provides the ChatGPT maker cloud computing capacity.
The company signed an additional agreement under which OpenAI has committed to pay up to $4 billion through April 2029.
CoreWeave in August reported a surge in demand for its cloud services in the second quarter, amid the rapid adoption of artificial intelligence tools.
However, operating expenses jumped nearly fourfold to $1.19 billion, highlighting the strain that rapid revenue growth is placing on its finances.
Reporting by Harshita Mary Varghese in Bengaluru; Editing by Sriraj Kalluvila
READ the latest news shaping the AI Chips market at AI Chips News
CoreWeave, Nvidia sign $6,3 billion cloud computing capacity order, source
Add comment