DeepSeek Releases Image-Video Large Model Janus-Pro-7B, Competing with OpenAI DALL-E 3
DeepSeek’s multimodal tech revolutionizes AI, with Janus-Pro surpassing DALL-E 3 in performance. Explore efficient, open-source models for advanced image generation and understanding.
"AI Disruption" publication New Year 30% discount link.
The U.S. stock market's NASDAQ has yet to recover from DeepSeek R1, and now DeepSeek's multimodal technology is back.
On Monday, NVIDIA’s market value dropped significantly, losing nearly $600 billion, marking the largest single-day market value drop in U.S. corporate history. The stock plunged by 17%, closing at $118.58.
Just last week, NVIDIA surpassed Apple to become the world’s most valuable publicly traded company. This dramatic drop directly caused the NASDAQ index to fall by 3.1%.
Another major U.S. chip manufacturer, Broadcom, which has seen massive market value growth due to AI, also saw a 17% drop on Monday, evaporating $200 billion in market value.
This sell-off is mainly due to rising concerns about the increased global AI competition pressure brought on by the Chinese AI lab DeepSeek.
Last December, DeepSeek released a free, open-source large language model: DeepSeek—V3. Its training cost was just $5.58 million, and it was completed in only two months using NVIDIA’s lower-powered H800 chips.
In contrast, tech giants like Alphabet, Meta, and Amazon have spent billions on cutting-edge NVIDIA hardware for training and running AI models.