DeepSeek—OpenAI's rival from China—hits top spot on Apple App Store
What's the story
DeepSeek, a Chinese AI start-up, has taken the tech world by storm. Its app, powered by the company's new reasoning model, R1, has quickly captured the attention of users and industry experts alike.
Within days of launch, it topped Apple's US productivity apps and ranked third in India. It also topped China's "Efficiency" category.
The R1 model has garnered significant attention from global AI community for its ability to rival or even surpass models developed by AI leaders like OpenAI.
Cost efficiency
Cost-effective approach challenges major tech giants
DeepSeek has been able to build models that match the performance of its competitors at a fraction of cost.
This affordability is a far cry from US companies, which have spent billions on purchasing NVIDIA H100 chips and data for their models.
As NVIDIA's flagship GPU, the H100 is off-limits to China due to US sanctions.
Instead, DeepSeek built and trained its V3 model for under $6 million using 2,000 NVIDIA H800 chips—a downgraded version with slower data transfer speeds.
Global impact
DeepSeek's success could impact US AI dominance
The success of DeepSeek, a relatively unknown Chinese start-up, in matching the capabilities of global tech giants with lesser resources could potentially disrupt US efforts to maintain an AI edge over international competitors.
The development comes as OpenAI announced a joint venture with the federal government to invest $500 billion in AI infrastructure over the next four years.
Currently, OpenAI's ChatGPT holds the second position on Apple's US chart where DeepSeek's R1 model reigns supreme.