NEW YORK: Chinese startup DeepSeek has emerged as a formidable player in the artificial intelligence (AI) space, with its DeepSeek AI Assistant overtaking ChatGPT to become the top-rated free app on Apple’s App Store in the United States, according to app research firm Sensor Tower.
Released on January 10, the DeepSeek AI Assistant, powered by the company’s DeepSeek-V3 model, has rapidly gained popularity among U.S. users. DeepSeek-V3, described by its creators as “a leader among open-source models and competitive with the world’s most advanced closed-source models,” has sparked discussions in Silicon Valley and beyond about China’s growing capabilities in AI technology.
DeepSeek’s achievement has sent shockwaves through Silicon Valley, challenging long-held beliefs about U.S. dominance in AI innovation. It has also raised questions about the effectiveness of Washington’s export controls, which aim to limit China’s access to advanced chips and AI development tools.
AI models like ChatGPT and DeepSeek require high-performance chips for training. Since 2021, the Biden administration has expanded export bans on these chips to prevent their use in Chinese AI projects. However, DeepSeek researchers revealed in a paper last month that the DeepSeek-V3 model was trained using Nvidia’s H800 chips, costing less than $6 million—a fraction of the investment required for U.S.-based models.
While some have disputed the specifics of DeepSeek’s claims, including the use of less advanced chips, the revelation has prompted U.S. tech executives to question whether these restrictions are effectively curbing China’s progress in AI.
Founded in 2023 in Hangzhou, DeepSeek is a relatively small startup that emerged as China’s AI sector began to gain traction. The launch of Baidu’s first Chinese AI large-language model that same year set off a wave of AI development across China, with dozens of companies releasing their own models.
What sets DeepSeek apart is the praise it has received from the U.S. tech industry. Experts have lauded the performance of DeepSeek-V3, comparing it favorably to cutting-edge U.S. AI models.
“DeepSeek is on par with, or even surpasses, the best American models,” said Alexandr Wang, CEO of Scale AI, in an interview with CNBC.
DeepSeek’s rise has had far-reaching effects on the global tech landscape. On Friday, all three major Wall Street indexes fell, with the S&P 500 retreating from a record high. Investors took profits amid concerns about heightened competition following DeepSeek’s breakthrough.
The fear of competition has rattled U.S. tech giants like Nvidia, Meta, and Alphabet, which have collectively invested hundreds of billions of dollars into AI development. The relatively modest $5.6 million spent on developing DeepSeek-V3 has highlighted the stark contrast in costs and efficiency between the emerging Chinese model and established U.S. players.
In Tokyo, tech and chip firms saw significant losses, with Advantest falling more than 8 percent and Tokyo Electron losing nearly 5 percent. SoftBank, a major backer of former U.S. President Donald Trump’s proposed $500 billion AI infrastructure project, also tumbled over 8 percent.
Other global markets were also affected, with losses recorded in Shanghai, Singapore, Mumbai, Bangkok, and Manila, while Hong Kong managed to post gains.
DeepSeek’s rapid ascent comes at a critical time for AI, as nations compete for technological leadership in the sector. The success of DeepSeek has underscored China’s growing prowess in AI innovation, prompting U.S. policymakers and tech companies to re-evaluate their strategies.
With China’s AI startups continuing to gain ground and U.S. tech giants feeling the heat, the global race for dominance in artificial intelligence has never been more intense.
