AI Breakthrough: DeepSeek Launches GPT-4 Rival

DeepSeek-R1 Challenges OpenAI with Cost-Effective AI Model

• DeepSeek-R1 is a new AI model from Chinese startup DeepSeek, rivaling OpenAI’s o1.
• The training cost for DeepSeek-R1 is $5 million, compared to OpenAI’s $100 million.
• DeepSeek-R1 requires only 2,000 GPUs, whereas GPT-4o needs 100,000 GPUs.
• The model is open-source under the MIT License, allowing free use and modification.
• NVIDIA’s stock price has declined following the announcement of DeepSeek-R1.

DeepSeek-R1’s introduction significantly reduces AI development costs and resources needed, with its $5 million training cost contrasting sharply with OpenAI’s $100 million investment. This shift has impacted NVIDIA’s stock value negatively.