A crippling cyberattack Monday forced DeepSeek, the controversial Chinese artificial intelligence (AI) startup, to temporarily limit user registrations.
The “large-scale malicious attacks” momentarily derailed what had been a historic day for DeepSeek, whose AI assistant supplanted OpenAI’s status as the most-downloaded free app in the U.S. on Apple Inc.’s App Store. Those who earlier downloaded DeepSeek are still able to log in.
“Due to large-scale malicious attacks on DeepSeek’s services, registration may be busy. Please wait and try again,” the company said in a banner note. DeepSeek did not specify the type of attack, but according to its status page, the API was operating with “degraded performance” and the web chat service had a “partial outage.”
The cyberattack capped the disruptive debut of DeepSeek and its R1 reasoning model that not only created a digital land rush of consumers and a wide-scale hack but roiled tech stocks in a worldwide sell-off. Worst hit in the trillion-dollar correction was NVIDIA Corp., which lost nearly $600 billion in market value Monday — the biggest single-day drop in U.S. stock market history — after its shares plunged 17%. Arm Holdings plummeted 10%. Google parent Alphabet Inc. fell 4%. The tech-heavy Nasdaq Composite declined more than 3%.
Silicon Valley and generative AI makers across the globe are still processing the overnight emergence of DeepSeek, a nearly 2-year-old startup that reportedly originated from a Chinese hedge fund’s AI research unit. The company claims R1, developed for $5.6 million, equals or exceeds human intelligence on a variety of tasks, outperforming models from OpenAI, Google, Anthropic, and others – most of whom have poured billions of dollars into their products.
The sudden impact of DeepSeek also highlighted fears over what China is developing in AI, and at what cost, at a time when the U.S. government, private companies, and investors are throwing billions of dollars into the technology. Last week, President Donald Trump and tech executives announced Stargate, a project with the audacious goal of raising up to $500 billion for data center expansion in the U.S. over several years. Microsoft Corp., meanwhile, is on pace to invest $80 billion in fiscal 2025 to build AI-enabled datacenters to train AI models and deploy AI and cloud-based applications around the world. Meta Platforms Inc. is splurging a reported $65 billion on AI this year.
“(DeepSeek) lab’s open source stance hints at a democratized future where smaller firms can feasibly build — or distill — high-performing models,” Red Sift CEO Rahul Powar said in an email. “As seen in DeepSeek-R1’s chain-of-thought demonstrations, transparency in AI design can be as disruptive as performance alone. With demanding benchmarks like Humanity’s Last Exam now in play, the global AI race is sure to intensify, but those shocks could ultimately yield a more innovative and widespread adoption of AI technologies.”
But while the release of DeepSeek showcases the immense potential of open-source AI, it also “naturally attracts diverse threat actors ranging from hacktivists to sophisticated state-sponsored groups seeking to exploit or disrupt this emerging AI platform,” said Stephen Kowski, field chief technology officer at SlashNext Email Security.
And the speed and approach with which R1 was constructed may have exposed it to hackers, warn security experts.
“It appears that not only did DeepSeek skimp on the number of GPUs, but failed to design with security in mind,” Kevin Kirkwood, chief information security officer at Exabeam, said in an email. “Back doors, open gateways, and other easily avoidable security flaws make this product a threat actor’s dream for compromising the data that a user puts into it.”
