DeepSeek, a Chinese artificial intelligence startup, is temporarily limiting new user registrations following a large-scale cyberattack on its services. The company, founded in 2023, has quickly emerged as a formidable competitor to OpenAI's ChatGPT, Google's Gemini, and other leading AI technologies. Its R1 model, released last week, has gained significant traction, topping app stores and industry leaderboards due to its exceptional reasoning capabilities.
Despite the U.S. curbing chip exports to China three times over the past three years, DeepSeek has successfully developed its models. The R1 model, designed as a reasoning tool rivaling OpenAI's o1, is open-source, allowing unrestricted access for AI developers. The startup has managed to dethrone OpenAI's ChatGPT as the most-downloaded free app on Apple's App Store.
In response to the recent cyberattack, DeepSeek announced it would temporarily limit user registrations. However, existing users can continue logging in without interruption. The startup's decision follows concerns about the security of its rapidly growing user base and infrastructure.
DeepSeek's growth trajectory is noteworthy, especially given the cost-effectiveness of developing the R1 model. Estimates suggest the model was created at a fraction of the cost incurred by competitors such as OpenAI, Anthropic, and Google. With an estimated training cost of $5.6 million — assuming a $2 per hour rental cost for GPUs — DeepSeek's efficiency in resource utilization has been commendable.
Emerging from a Chinese hedge fund's AI research unit in April 2023, DeepSeek shifted its focus toward large language models and pursuing artificial general intelligence (AGI). This strategic pivot has paid off, capturing the attention of industry insiders and tech enthusiasts alike. However, questions loom over the AI sector regarding the sustainability of inflated funding rounds and billion-dollar valuations, raising concerns about a potential bubble.