BEIJING, Jan 27 (Reuters) — Chinese startup DeepSeek is making waves in the global AI landscape with the launch of its DeepSeek-V3 and DeepSeek-R1 models. The company claims its AI technology matches or even surpasses the performance of industry-leading models from U.S. giants like OpenAI and Meta, but at a fraction of the cost, shaking up the competitive dynamics in artificial intelligence.
Why is DeepSeek Gaining Global Attention?
DeepSeek’s rapid rise follows the 2022 launch of OpenAI’s ChatGPT, which triggered a race among Chinese tech firms to develop their own AI chatbots. However, initial releases from major players like Baidu left many disappointed, as their AI models struggled to match the capabilities of U.S. counterparts.
Now, DeepSeek has flipped that narrative. The DeepSeek-V3 model is generating buzz for its superior quality and cost-efficiency. DeepSeek’s AI Assistant, powered by DeepSeek-V3, recently became the top-rated free app on Apple’s U.S. App Store, overtaking ChatGPT.
Notably, DeepSeek-V3 was trained with just under $6 million worth of computing power, using Nvidia H800 chips. In comparison, the costs for U.S.-developed models can often run into the hundreds of millions. This stark price difference is prompting experts to reconsider the financial rationale behind massive AI investments by companies like Nvidia and others.
DeepSeek Models: Cost-Effective & High-Performance
The DeepSeek-R1, launched just last week, has already been praised for being 20 to 50 times cheaper than OpenAI’s GPT-4 for many tasks. This breakthrough is a game-changer for businesses and developers looking to integrate AI while keeping costs down.
Despite the praise, some experts remain skeptical about DeepSeek’s cost claims. Scale AI CEO Alexandr Wang raised questions about the company’s access to advanced Nvidia chips, citing U.S. export restrictions that prevent the sale of H100 chips to Chinese companies. However, DeepSeek has not yet publicly addressed these concerns.
Who Is Behind DeepSeek?
DeepSeek was founded in Hangzhou by Liang Wenfeng, co-founder of the Chinese hedge fund High-Flyer. Liang’s investment fund has a strong focus on AI, and in March 2023, it shifted its resources to develop a new research group aimed at creating Artificial General Intelligence (AGI). The formation of DeepSeek followed soon after, backed by the same fund.
While the exact investment figures from High-Flyer into DeepSeek are undisclosed, it is known that the hedge fund operates a cluster of 10,000 A100 chips, which are crucial for training AI models.
Beijing’s Interest in DeepSeek’s Success
DeepSeek’s breakthrough has not gone unnoticed in Beijing. On January 20, the day DeepSeek-R1 was launched, Liang Wenfeng attended a closed-door symposium hosted by Chinese Premier Li Qiang, signaling the importance of DeepSeek’s AI success to China’s broader goals. The country aims to overcome Washington’s export controls and establish AI self-sufficiency, which could be accelerated by DeepSeek’s advancements.
This marks a growing recognition that AI is one of the key industries in China’s strategy to assert technological independence on the global stage.
Stay Ahead in the Tech Race
For those involved in the tech world, staying updated on the latest AI advancements is critical. Curious about your IP address and online presence? You can easily check it with our What is My IP tool. Whether you’re a developer, tech enthusiast, or business owner, knowing your IP can help optimize security and performance in your AI and tech ventures.
