DeepSeek actually used 1.6 Billion instead of 6 Million.

Posted: Feb 02, 2025

  • DeepSeek, a Chinese startup, claimed to have trained its AI model with low resources, but a report revealed it actually spent $1.6 billion on hardware and has 50,000 Nvidia GPUs
  • DeepSeek operates its own data centers and primarily hires talent from mainland China
  • The company has invested over $500 million in technology
  • DeepSeek's efficient approach to AI model development may reduce the need for high-end GPUs, impacting companies like Nvidia
  • A recent claim that training costs were only $6 million were just for a portion of the total training cost, not accounting for research or infrastructure expenses
  • DeepSeek's success is attributed to strategic investments, technical breakthroughs, and a competitive workforce, challenging industry leaders in AI




Stay ahead of the market with AI stock alerts & AI summaries of the latest earnings, stock ideas for free with Fluid Bot. Sign up now!