TechCrunch News 01月28日
Nvidia drops $600B off its market cap amid the rise of DeepSeek
index_new5.html
../../../zaker_core/zaker_tpl_static/wap/tpl_guoji1.html

 

中国AI初创公司DeepSeek发布的R1模型,以较少的资源和算力,展现出与美国同类模型相媲美的功能和准确性,这被视为人工智能行业的胜利。然而,这一成就对半导体巨头英伟达造成了冲击,其股价在短短两个交易日内暴跌近17%,市值蒸发近6000亿美元。市场猜测,DeepSeek模型的成功表明,构建强大的AI模型不一定需要昂贵的高端芯片,这直接威胁到英伟达等芯片制造商。尽管英伟达承认DeepSeek的进步,并强调其推理仍需大量GPU,但DeepSeek的出现无疑给美国在AI领域的领导地位敲响了警钟。

🚀 DeepSeek R1模型在资源和算力受限的情况下,展现出与美国先进模型相当的性能,这标志着AI技术发展的新方向。

📉 英伟达股价因DeepSeek的成功而大幅下跌,这表明市场开始质疑高端芯片在AI模型构建中的必要性,对芯片制造商造成冲击。

⚖️ DeepSeek的案例突显了“测试时扩展”技术的重要性,表明可以通过利用现有模型和算力,构建出强大的AI模型,打破了对高端硬件的依赖。

🌍 美国对华出口高端AI芯片的限制,反而可能促使中国等国家发展出更高效的AI技术,美国若想保持AI领域的领先地位,需关注更多层面,而不仅是芯片。

DeepSeek, a Chinese AI startup, became the talk of the proverbial AI town when it released its R1 model on Friday. R1’s functionality and accuracy compared to its U.S. counterparts, despite using fewer resources and less compute power, seems like a win for the overall AI industry. But it isn’t necessarily good news for everyone.

Semiconductor giant Nvidia finds itself in the headwinds of DeepSeek’s recent achievement. The chip giant saw its stock plummet 16.9% from the close of Friday’s public markets to the close of public markets on Monday, according to Yahoo Finance data. Nvidia lost nearly $600 billion off of its market cap. Nvidia’s stock closed at $142.62 a share on Friday afternoon. On Monday, it closed at $118.58.

There is speculation that the reason why DeepSeek’s model release would impact Nvidia’s stock is that R1 provides a clear example that AI models don’t necessarily need expensive, high-end chips or hardware to build an impressive model, which isn’t exactly great news for a chipmaker like Nvidia.

“DeepSeek is an excellent AI advancement and a perfect example of Test Time Scaling,” an Nvidia spokesperson told TechCrunch over email. “DeepSeek’s work illustrates how new models can be created using that technique, leveraging widely-available models and compute that is fully export control compliant. Inference requires significant numbers of Nvidia GPUs and high-performance networking. We now have three scaling laws: pre-training and post-training, which continue, and new test-time scaling.”

The timing of all of this is interesting because this comes one week after former president Joe Biden signed an executive order that made further restrictions on the export of U.S.-produced advanced AI chips to certain countries, with near-blanket restrictions on sending chips to countries like China, where DeepSeek is based.

At the time, Nvidia said that the executive order was “unprecedented and misguided” and that it would “derail” innovation and economic growth worldwide.

President Donald Trump has since reversed Biden’s executive order and has signed a different executive order to create the Stargate Project, an infrastructure program that will invest up to $500 billion into AI data centers.

The release of DeepSeek shows that if the U.S. wants dominance over the global AI market, it may need to pay attention to more than just chips and AI hardware.

Fish AI Reader

Fish AI Reader

AI辅助创作,多种专业模板,深度分析,高质量内容生成。从观点提取到深度思考,FishAI为您提供全方位的创作支持。新版本引入自定义参数,让您的创作更加个性化和精准。

FishAI

FishAI

鱼阅,AI 时代的下一个智能信息助手,助你摆脱信息焦虑

联系邮箱 441953276@qq.com

相关标签

DeepSeek 英伟达 AI模型 芯片 测试时扩展
相关文章