WASHINGTON: US President Donald Trump has described DeepSeek’s emergence as “a wake-up call” for American technology companies.
US Tech Giants Face Challenges
DeepSeek’s R1 AI model, made at far lower costs than rivals, has shocked the global tech industry. Nvidia shares plummeted, wiping out nearly $600 billion in market value, signaling concerns over America’s AI dominance.
DeepSeek claims its R1 model achieved similar results to leading US models but at just $6 million in training costs.
Rapid Rise of DeepSeek
Launched earlier this month, DeepSeek became the most downloaded free app in the US within a week. Its quick success sparked comparisons to OpenAI’s models and intensified scrutiny of its methods.
Trump Responds to DeepSeek’s Rise
Speaking aboard Air Force One, President Trump called China’s AI progress “a positive challenge” for US companies.
“If we achieve the same result with less cost, it’s good for us,” he stated. Trump remained confident about US leadership in AI.
Asian Stock Markets React
Japanese AI-related stocks, including SoftBank and Advantest, faced sharp losses on Tuesday, contributing to a 1% drop in the Nikkei 225 index.
Other Asian markets, including Taiwan and South Korea, remained closed for Lunar New Year celebrations. Mainland China’s financial markets are shut until February 5.
DeepSeek’s Innovative Approach
DeepSeek’s open-source R1 model relies on data efficiency and reduced hardware requirements. The company claims its model requires only 2,000 chips, compared to the 16,000 estimated for competitors.
This innovation stems from Chinese AI developers sharing research and finding efficient methods, bypassing US export restrictions on advanced chips.
Mixed Reactions from Experts
Sam Altman, CEO of OpenAI, called DeepSeek’s model “impressive,” especially considering its affordability. However, he added that OpenAI’s models would remain superior.
Not everyone shares the enthusiasm. Elon Musk suggested DeepSeek may have exaggerated its hardware capabilities, claiming it likely uses 50,000 Nvidia chips.
