DeepSeek AI Surpasses ChatGPT as Most Downloaded App, Challenges Nvidia with Cost-Effective Technology

The Chinese startup DeepSeek has introduced a new AI model that’s reshaping the competitive landscape.

This week, their chatbot soared to the number one spot on the US Apple App Store, surpassing OpenAI’s ChatGPT as the most downloaded free app.

DeepSeek’s technology, based on their recently unveiled R1 reasoning model, boasts comparable capabilities to leading AI models but at a significantly reduced cost and lower hardware requirements.

On January 20th, DeepSeek released the R1 model, designed for complex problem-solving.

It reportedly matches the performance of OpenAI’s offerings on some benchmarks, yet it required fewer resources to develop.

According to the company, their V3 LLM model, a precursor to R1, cost less than $6 million to develop—far less than the reported $100 million for OpenAI’s GPT-4.

The financial implications are considerable, with Nvidia’s stock dropping by over 12 percent in pre-market trading following the news.

DeepSeek’s claim of needing only 2,000 Nvidia chips for training, compared to the 16,000 typically required, has sparked a debate over the efficiency of AI development methods used by major tech firms like Nvidia, Microsoft, and Meta.

Investors are now reevaluating the billions poured into AI data centers, including the massive $500 billion Stargate Project, questioning whether these investments are as future-proof as once thought.

The developments at DeepSeek suggest that alternative approaches to AI might lead to similar or even superior results without the hefty price tag.

This shift could have lasting effects on the industry, especially under the looming shadow of trade restrictions aimed at maintaining US dominance in the AI sector.