DeepSeek suffers large-scale cyberattack and says it is limiting logs

DeepSeek said on Monday it would temporarily limit user registrations “due to large-scale malicious attacks” on its services, although existing users will be able to log in as usual.

The Chinese artificial intelligence startup has generated a lot of commotion in recent weeks as a fast-growing rival to OpenAI’s ChatGPT, Google’s Gemini, and other leading AI tools.

Earlier on Monday, DeepSeek took rival OpenAI’s coveted spot as the most downloaded free app in the US. Apple‘s App Store, dethroning ChatGPT by DeepSeek’s own AI assistant. Helped inspire a significant liquidation in global technology stocks.

Rumors about the company, which was founded in 2023 and launched its R1 model last week, have spread to analysts, investors and technology developers, who say the hype, and the resulting fear of being left behind in the ever-changing cycle of AI hype, may be justified. Especially in the era of the generative AI arms race, where both tech giants and startups compete to ensure they are not left behind in a market. expected to exceed $1 trillion in income in a decade.

DeepSeek reportedly emerged from a Chinese hedge fund’s artificial intelligence research unit in April 2023 to focus on large language models and achieve artificial general intelligence, or AGI, a branch of artificial intelligence that matches or It surpasses human intellect in a wide range of tasks, which OpenAI and its rivals say they are rapidly chasing.

Rumors around DeepSeek began to spread especially last week, when the startup launched R1, its reasoning model that rivals OpenAI’s o1. It’s open source, meaning any AI developer can use it, and it has skyrocketed to the top of the app stores and industry league tablesand users praise its performance and reasoning abilities.

The startup’s models were notably built despite the US. curb chip exports to China three times in three years. Estimates differ on how much exactly DeepSeek’s R1 costs or how many GPUs are included in it. Jefferies analysts estimated that a recent version had a “training cost of just $5.6 million (assuming a rental cost of $2 per hour and 800 hours). That’s less than 10% of the cost of Goal‘s Flame.”

But regardless of the specific figures, reports agree that the model was developed at a fraction of the cost of rival models by OpenAI, Anthropic, Google and others.

As a result, the AI ​​sector is awash with questions, including whether the industry’s growing number of astronomical funding rounds and billion-dollar valuations are necessary, and whether a bubble is about to burst.