Chinese AI assistant DeepSeek has become the top rated free app on Apple’s App Store in the US and elsewhere, beating out ChatGPT and other rivals. It’s powered by the open-source DeepSeek V3 model, which reportedly requires far less computing power than competitors and was developed for under $6 million, according to (disputed) claims by the company. At the same time, it offers performance that’s on par with Claude-3.5, GPT-4o and other rivals, DeepSeek said last week.
Available on web, app and API, DeepSeek is similar to AI Assistant like ChatGPT with features like coding content creation and research. Its first DeepSeek-R1 release is available under an MIT license, so it can be used commercially and without restrictions.
The company is headquartered in Hangzhou, China and was founded in 2023 by Liang Wenfeng, who also launched the hedge fund backing DeepSeek. To develop the tech, he reportedly stockpiled NVIDIA A100 chips prior to the US export ban and paired those with less powerful chips that can still be imported, according to MIT Technology Review.
However, DeepSeek was still at a significant hardware disadvantage next to rival models from OpenAI, Google and others. That forced the company to be more efficient with its AI models, and it has supposedly been able to build and train them at a far lower cost than previously thought possible.
Analysts from Citi and elsewhere have questioned those claims, though, and pointed out that China is a “more restrictive environment” for AI development than the US. Still, the rise of DeepSeek has raised concerns about the potential profits of rivals like OpenAI that have already invested billions in AI infrastructure. In fact, the news that DeepSeek topped the App Store charts caused a sharp drop in tech stocks like NVIDIA and ASML this morning.
If you buy something through a link in this article, we may earn commission.