Nvidia has reportedly invested $150 million (about 220 billion won) in Baseten, a startup specialized in artificial intelligence (AI) "inference."
On the 20th (local time), the Wall Street Journal (WSJ), citing sources, said Baseten raised $300 million (about 440 billion won) through a funding round led by venture capital IVP, among others, and Nvidia invested $150 million of that amount. In this funding round, Baseten's corporate value was assessed at $500 million, double the previous value.
Nvidia is actively investing in startups focused on AI inference. Earlier, in Dec. last year, Nvidia signed a technology licensing deal with Groq, a startup specializing in AI accelerator chip design. Groq said, "This agreement reflects a shared goal of expanding access to high-performance, low-cost inference technology." The two sides did not disclose the contract amount, but foreign media reported the size at $20 billion (about 29 trillion won).
AI models are generally divided into "training" and "inference." Training is the process of "learning" patterns from large amounts of data, and inference refers to the process of "deriving" predictions or conclusions about new data using a trained model. While most AI models released so far have been developed through training, mainly by global big tech, there is an outlook that the inference market will rise further in the future.