Artificial intelligence company DeepSeek AI has gained popularity globally, challenging the US control over the AI industry. Founder Elon Musk has questioned its microchip claims, saying it’s “obviously” that 50,000 specialized chips are needed to train DeepSeek’s model.
DeepSeek AI, powered by open-source models, requires fewer chips than leading models, with only 2,000 specialized chips from Nvidia needed for training, compared to 16,000 or more. This has raised questions about the computing power required for developing AI systems.
The company’s AI assistant has overtaken rival tool ChatGPT and become the top-rated free application available on Apple App Store in the US. Industry leaders and the White House are expected to respond to DeepSeek’s rise in popularity, which could alter the global AI market.
Musk’s comments follow posts about DeepSeek AI on X, where CEO Alexandr Wang mentioned 50,000 NVIDIA H100s used for training due to US export controls. Salesforce CEO Marc Benioff also praised DeepSeek, stating its potential lies in data and metadata rather than UI or model costs.
As DeepSeek gains traction, it could boost China’s position in the AI race. Experts are divided on the implications of this shift, with OpenAI CEO Sam Altman warning about a “strategic choice” between an open access global AI or an authoritarian one that restricts its benefits to certain nations.
Musk’s skepticism has sparked debate among AI enthusiasts and experts. With DeepSeek’s rise, the US government is likely to take a more involved approach to AI regulation. The future of AI development hangs in the balance as this emerging technology continues to reshape global markets.
Source: https://www.newsweek.com/deepseek-elon-musk-microchip-claims-2021394