xAI, the AI company founded by Elon Musk, has missed its promised release window for the Grok 3 model, originally set for the end of 2024. In the past, Musk described Grok 3 as a significant advance in AI, claiming it would be a tremendous stride forward, powered by a gigantic cluster of GPUs in Memphis. But as of January 2, Grok 3 is not here, and no immediate signs indicate that its launch is imminent.
In a twist, some code observed on xAI’s website points to the release of an intermediate model, Grok 2.5, before the much-anticipated Grok 3. The delay comes as part of a trend where AI companies are missing their launch deadlines. Last year, AI startup Anthropic also missed the launch of its next-generation model, Claude 3.5 Opus, while other companies like Google and OpenAI have experienced setbacks with their flagship models.
Repeated delays point to the scale-up challenges for AI models. For years, as AI capabilities have grown, so, too, have the benefits of adding more computing power and larger datasets started to hit diminishing returns. That caused companies to look elsewhere in improving their models. Setting aside the setbacks, Musk tweeted that he is optimistic the Grok 3, while acknowledging in a recent interview that it may take many cycles until expectation targets are met.
xAI also has a smaller team size compared to other companies such as Google DeepMind. Perhaps this, combined with advancing AI technology becoming increasingly complex, could be a significant factor in the delay. However, the missed deadlines add to concerns that there may be limits to conventional AI scaling approaches.
