China’s cost-effective DeepSeek AI takes world by storm
ChatGPT falls from prominence, Nvidia records highest daily market cap decline ever, while DeepSeek reveals AI development possible with far fewer resources
ANKARA
China’s cost-effective and free DeepSeek artificial intelligence (AI) chatbot took the world by storm due to its rapid progress rivaling the US-based OpenAI’s ChatGPT with far fewer resources available.
DeepSeek gained unprecedented popularity in mobile app stores across the globe, dethroning ChatGPT in certain regions, the US included.
The rapid progress of the large language model (LLM) gained center stage in the tech world, as it is not only free, open-source, and more efficient to run, but it was also developed and trained using older-generation chips due to the US’ chip restrictions on China.
The world’s leading AI firms use over 16,000 chips to train their models, while DeepSeek only used 2,000 chips that are older, with a less than $6 million budget. OpenAI has reportedly spent over $100 million for the most advanced model of ChatGPT, the o1, which DeepSeek is rivaling and surpassing in certain benchmarks.
DeepSeek’s high performance against competing models from Google and OpenAI, which use much more advanced and expensive chips and infrastructure, pressured Big Tech stocks.
American chip giant Nvidia’s share fell as low as $117.26 on Monday and the firm’s one-day loss edged close to 18%, the highest daily loss since March 2020.
At the same time, Nvidia’s market cap decreased by over $500 billion, the largest daily decline in the US stock market history.
The launch of DeepSeek ignited a new debate over the costliness of LLMs, as the rise of low-cost models threatened the dominance of Western Big Tech in AI.
Billionaire Mark Zuckerberg, the CEO of Meta, announced to spend $60 billion to $65 billion on AI infrastructure this year, while President Donald Trump said the US would invest $500 billion.
Trump also said DeepSeek has been “a wake-up call” for Silicon Valley, telling American firms to focus on competing.
OpenAI CEO Sam Altman said on X that DeepSeek 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,” he added.
* Writing by Emir Yildirim in Istanbul.
