AI行业巨头纷纷推迟新一代模型发布，AI技术发展进入瓶颈期了吗？Last summer, Musk stated that the next generation AI model Grok 3 from xAI would be launched by the end of 2024. However, it is now 2025, and Grok 3 has still not been released, nor are there any signs of an upcoming launch. Aside from xAI, Other AI companies are facing similar situations: Anthropic has canceled the planned release of the Claude 3.5 Opus model, originally scheduled for the end of 2024; the flagship model developments at Google and OpenAI have also encountered setbacks.
Source: Hard AI.
Author: Jiang Zihan
Are major players in the AI Industry delaying the release of the next generation of models, indicating a bottleneck in the development of AI Technology?
Last summer, Musk posted on social platform X that xAI's next generation AI model Grok 3 would be trained on 0.1 million H100 GPUs and was set to be released by the end of 2024, stating it would be "very special". However, it is now 2025 and Grok 3 has not yet been released, nor are there any signs of an imminent launch.
Interestingly, in addition to xAI, other AI companies are experiencing similar situations:
Anthropic has canceled the release of the Claude 3.5 Opus model originally scheduled for the end of 2024;
Reports indicate that flagship model development by Google and OpenAI has also encountered setbacks.
Analysts suggest that the small size of the xAI team may be one of the reasons for the delay in releasing Grok 3.
But more importantly, the fact that AI companies are postponing releases may reflect the limitations of the current scaling law in AI—previously, significantly enhancing AI model performance was possible by using vast computational resources and larger datasets; however, recently, the performance improvements of each new generation of models have begun to shrink, prompting technology giants to seek alternative technologies.
Elon Musk himself hinted at this in an interview: when asked if Grok 3 would reach industry-leading levels, Musk replied, "Hopefully, that's our goal, but we may fail."
