Answer from China with Deepseek in artificial intelligence! rose to number 1

DeepSeek, founded by hedge fund manager Liang Wenfeng, released its R1 model on January 20 and announced its model, which can automatically learn and improve itself without human supervision.

After Washington banned Nvidia from exporting its most powerful chips to China, local AI companies were forced to find innovative ways to maximize the processing power of the limited number of local chips. Due to US sanctions, China is unable to import H100s and is instead importing H800s with lower data transfer speeds, Reuters reported.

In the statement made today, DeepSeek engineers said that they were able to produce the models with lower quality Nvidia chips and less capital. DeepSeek stated that it used 2048 Nvidia H800s and only 5.6 million dollars to train a model with 671 billion parameters, which is far below the amount OpenAI and Google spent to train models of similar size.

After President Donald Trump took office in the USA, OpenAI announced that it had launched a joint venture with the federal government to spend $500 billion on artificial intelligence infrastructure over the next four years. A little-known Chinese company closing the gap on the world’s largest technology companies with far fewer resources could undermine U.S. efforts to remain a leader in artificial intelligence, experts say.

U.S. stock futures and Asian shares outside China fell on Monday as DeepSeek weighed the implications of launching a free, open-source artificial intelligence model to rival OpenAI’s ChatGPT.

Spokespeople of Deepseek, Meta and OpenAI have not yet made a statement on the issue.