When Russia launched Sputnik 1, the first satellite into space in 1957, the West was caught off guard. The USA lagged behind in its own development of spacecrafts, and scrambled into action. This was the start of the space race, and Nasa would be founded the following year.
Nearly 70 years and 12 men on the moon later, and commentators are saying that artificial intelligence is having its Sputnik moment. America is home to all of the top ten AI companies, yet a previously unknown Chinese AI model called DeepSeek crash landed onto the scene last week, soaring to the top of the App Store and sending US tech stocks spiralling. Is this the start of the interface race?.
DeepSeek’s free, AI powered chatbot called V3 was released in December last year, while its new large language model called R1 came out on January 20 and quickly caught the attention of the tech industry. V3 can have conversations, solve complex maths problems, write advanced code and generate poetry just as capably as its competitors, like OpenAI’s ChatGPT, while R1 is a reasoning model, which works through complex problems step by step.
What sets Deepseek’s chatbot apart from its rivals is how cheaply it was developed. DeepSeek’s engineering team claim that they only needed around $6 million and two months to train the AI model. This is peanuts in the tech world: Meta recently spent around 10 times that figure building its latest AI technology.
Yet this relative minnow carved out a $1 trillion hole in Nasdaq, the tech-heavy US stock exchange. The company Nvidia, which develops chips and software which power AI, shed nearly $600bn in market value, losing its standing as the world’s most valuable company.
