A DeepSeek is redefining the global competition of artificial intelligence when developing Large scale models with lower cost and lower dependence on advanced chipsindustry experts said on Sunday.
Second Pan Helinfrom China’s Ministry of Industry and Computerization, Deepseek “has opened a new technical path” that avoids dependence on state -of -the -art chipsdribbling the restrictions imposed by the US. Your model R1launched in January, It cost only $ 5.6 million to be trainedless than 10% of the cost of Llamada Meta.
Even with US attempts to block the company, technology giants as Microsoft e Nvidia They are already adopting the Deepseek model. THE Microsoft integrated R1 support on its Azure cloud platform, while the Nvidia It has made the model available to users of your NIM service, highlighting its efficiency and accuracy.
A Huawei and SiliconFlow also announced that they already operate the R1/v3 in your cloud ascendensuring stability for commercial applications. To Pengfrom Beijing’s Academy of Social Sciences, DeepSeek differs in prioritizing the “opening and benefit to humanity”, in contrast to the US geopolitical control approach to AI.
Source: https://www.chinahoje.net/deepseek-contorna-restricoes-dos-eua-e-impulsiona-inovacao-global-em-ia/
