

28 Jan 2025
DeepSeek, a Chinese AI chatbot, performed better than ChatGPT with fewer resources.
DeepSeek, founded by Liang Wenfeng in May 2023, specialises in efficient and affordable open source large language models (LLMs). Open sourcing makes it available to researchers and developers. It has two models DeepSeek-V3 and DeepSeek-R1.
DeepSeek uses Reinforcement Learning (RL), allowing its models to comprehend and improve more efficiently. Its models were built using 50,000 Graphic Processing Units (GPUs) compared to OpenAI’s 500,000 GPUs.
DeepSeek’s cost-effective offerings caused the US tech indices to lose around $1 trillion. Nvidia’s stocks plummeted by 17% on Monday, and the company lost $589 billion in one day, the biggest single-day loss to a US company ever.
Nvidia saw a meteoric rise in net profits from $4.8 billion in 2022 to $66.7 billion in 2024 mainly due to its GPUs that sell for $25,000 per piece. DeepSeek’s model used Nvidia’s GPUs for training.
With this, Apple became the most valuable company globally again.
Â
Never miss another post from SalestorrsNews150. Follow Salestorrs on WhatsApp, LinkedIn, Facebook, X, and Instagram.