DeepSeek's cheaper models and older chips call into question trillions in AI infrastructure spending
Briefly

China's DeepSeek AI model has emerged as a formidable competitor to US AI firms. The DeepSeek-V3 model, which mirrors ChatGPT, operates at a fraction of the cost of its American competitors, being 20-40 times cheaper. Its performance is striking considering it employs a modest hardware configuration compared to the expansive resources typically used by firms like OpenAI. Analysts suggest that this efficiency challenge could have significant implications for U.S. investments in AI infrastructure, questioning its sustainability amid DeepSeek's rise.
DeepSeek's model dramatically reduces costs for AI deployment, making it 20-40 times cheaper than alternatives from OpenAI, raising questions about the viability of US AI investments.
DeepSeek's remarkable efficiency stems from its effective utilization of limited hardware, enabling it to outperform larger models despite having significantly fewer parameters.
Read at Business Insider
[
|
]