DeepSeek suffers cyberattack as popularity surges

28 Jan 2025

Image: © Valery/Stock.adobe.com

DeepSeek AI should be a ‘wake-up call’ for US industries, US president Donald Trump said.

For nearly a day, Chinese AI start-up DeepSeek has halted new registrations on its V3 AI chat platform due to “large scale malicious attacks” on its services, and while the company said it has since resolved the issue, registrations on the platform might still take time to fully recover.

Last week, the one-year-old start-up caused a flurry in Silicon Valley with the release of its latest reasoning model, the R1, which boasts capabilities on a par with industry heavyweights such as OpenAI’s GPT-4 and Anthropic’s Claude 3.5 Sonnet, while needing only $5.6m to train the model – a fraction of what it costs its US competitors.

The start-up has received much praise from industry leaders and direct competitors, including from OpenAI’s CEO Sam Altman, who wrote on X: “Deepseek’s R1 is an impressive model, particularly around what they’re able to deliver for the price.

“We will obviously deliver much better models and also it’s legit invigorating to have a new competitor! We will pull up some releases,” he added.

Moreover, weighing in on the Chinese start-up’s achievements – which came despite increasing US sanctions on AI chips to the country – US president Donald Trump said: “The release of DeepSeek AI from a Chinese company should be a wake-up call for our industries that we need to be laser-focused on competing to win.”

He explained that he saw DeepSeek’s advancements as a “positive”, adding, “instead of spending billions and billions, you’ll spend less, and you’ll come up with hopefully the same solution”.

However, Kela, a cyberthreat intelligence organisation, in a new report, said that DeepSeek’s R1 is significantly “more vulnerable” than ChatGPT.

The organisation said that its team was able to jailbreak, or bypass the model’s in-built safety measures and ethical guidelines, which enabled R1 to generate malicious outputs, including developing ransomware, fabricating sensitive content, and giving detailed instructions for creating toxins and explosive devices.

Following R1’s release, Nvidia, the world-leading chipmaker, lost close to $600bn in market cap yesterday (27 January). Although, the company is slowly recovering from the monumental loss as share prices slowly inch upwards.

In a statement yesterday, an Nvidia spokesperson praised DeepSeek, calling it an “excellent AI advancement and a perfect example of Test Time Scaling”.

“DeepSeek’s work illustrates how new models can be created using that technique, leveraging widely-available models and compute that is fully export control compliant.”

However, a former DeepSeek employee told MIT Technology Review that in order to train R1, the start-up had to use Nvidia GPUs specifically designed for the Chinese market that caps its performance at half the speed of its top products.

Don’t miss out on the knowledge you need to succeed. Sign up for the Daily Brief, Silicon Republic’s digest of need-to-know sci-tech news.

Suhasini Srinivasaragavan is a sci-tech reporter for Silicon Republic

editorial@siliconrepublic.com