Leidi.com Rakuten August 6
The popularity of ChatGPT has ushered in a wave of investment in the AI industry, and Nvidia has benefited greatly.
At present, NVIDIA A100/H100 series accelerator cards are highly sought after due to their powerful performance, but the price remains high, and it even costs more than 100,000 yuan to buy one. This has allowed a startup company in the United States to successfully obtain a large investment.
CoreWeave, a professional cloud provider for large-scale GPU acceleration workloads, recently announced the completion of $2.3 billion in debt financing, mainly using the NVIDIA H100 accelerator card in hand as collateral.
This round of financing was led by Magnetar Capital, led by Blackstone, with strategic participation from Coatue, DigitalBridge Credit, and participation from funds and accounts managed by BlackRock, PIMCO and Carlyle.
CoreWeave has unique access to the most advanced and short-supplied Nvidia chips, giving it an edge over traditional cloud providers such as Microsoft, Amazon and Google, which face supply constraints as they work to develop their own chips.
CoreWeave's massive loan also marks a growing market for private equity financing, as private equity firms turn to less risky loans backed by hard assets and take on more corporate debt at a time when banks are slowing down.
“We negotiated with them on a timeline of how much collateral we need to put in, a depreciation plan, a payout plan,” said Michael Intrator, CoreWeave’s chief executive. “For us, going out And borrowing money on an asset basis is a very cost-effective way to access the debt market.”
In April 2023, CoreWeave announced the raising of $221 million in Series B financing, led by Magnetar Capital and funded by NVIDIA, Nat Friedman and Daniel Gross. A month later, CoreWeave also received US$200 million in Series B financing led by Magnetar Capital. After this financing, the company’s valuation exceeded US$2 billion.
Founded in 2017, CoreWeave is a professional cloud provider that provides large-scale GPU computing resources on the industry's fastest and most flexible infrastructure.
CoreWeave builds cloud solutions for compute-intensive use cases (Machine Learning & AI, VFX & Rendering, Life Sciences, Metaverse, and Live Streaming).
CoreWeave is powering the LLM (Large Language Models) and generative AI craze with purpose-built ultra-high performance cloud infrastructure at scale.
CoreWeave said the new financing will be used to grow its high-performance computing team to serve its customers, open new data centers, and grow CoreWeave's world-class workforce.
CoreWeave CEO and co-founder Michael Intrator said: "As long as the infrastructure is in place to deliver performance at scale, AI has the potential to transform the way we engage with technology and power the industries of the future , and make socially important services more efficient.”
"CoreWeave is leading the industry forward by meeting this unprecedented demand with the most reliable, flexible and high-performance computing resources available. These new resources from world-class investors are a vote of confidence in what we have achieved to date , and validate our future strategy.”
Earlier this month, CoreWeave announced the construction of a new $1.6 billion data center in Plano, Texas, a milestone as the company continues to aggressively expand its capacity and infrastructure footprint. CoreWeave expects to have 14 data centers by the end of 2023.
Last month, CoreWeave unveiled the world's fastest AI supercomputer built in partnership with NVIDIA, which can train the full GPT-3 LLM workload in 11 minutes, as measured by an industry-standard benchmark called MLPerf.
Inflection AI creates one of the world's most complex large language models (LLMs) on CoreWeave Cloud using infrastructure that supports MLPerf submissions. In short, CoreWeave allows customers to train the most complex models in the shortest time possible.
———————————————
Lei Di was founded by Lei Jianping, a media person. If you reprint, please indicate the source.
