ChatGPT CEO Sam Altman has a $100m Meta problem as 'safety issues' delay OpenAI masterplan

OpenAI CEO Sam Altman has announced a delay to a much-hyped update, with the long-awaited open-weight model once again stumbling at the starting blocks. The artificial intelligence wars are heating up, with OpenAI's ChatGPT and xAI's Grok going head to head with two very different strategies.

While Elon Musk is leading the charge with the controversial Grok, Sam Altman is steering OpenAI's artificial intelligence ship.

Altman has recently warned about how he thinks AI will impact the future of his children, and as OpenAI evolves, it threatens Google's internet dominance with its own internet browser that could rival Chrome with generative AI.

Despite all of the progress OpenAI is making, there have been a few missteps along the way. Taking X, Altman announced that the open-model update has been delayed from its planned release next week.

Saying that OpenAI needs to run "safety tests and review high-risk areas," there's also no timeframe on when the open-model update will be rolled out. Admitting that the team doesn't know how long it will take, Altman added: "While we trust the community will build great things with this model, once weights are out, they can’t be pulled back. This is new for us and we want to get it right."

The OpenAI overlord first teased the open-weight model in March 2025, but while it was originally slated for a June launch, it then slipped to 'later' this summer.

Open-weight models allow the core settings (weights) to be accessed by the public. This means those with the right knowledge can download and use these models on their machines.

This is in contrast to closed-weight models, which can only be accessed by a small group of developers or organizations. Given that open-weight can effectively be accessed by anyone, it makes sense that Altman wants to make sure things are watertight before opening the floodgates.

Importantly, open-weight models are said to promote transparency and accountability due to independent researchers being able to review and test. It's also hoped that biases and unethical behaviors can be reduced because these models are open to public inspection.

OpenAI's open-weight rollout has been spurred on by the likes of China's DeepSeek and Meta embracing open-weight. In particular, Meta's Llama models have been downloaded over a billion times and prove the demand for more open AI tech.

Developers hope that OpenAI's open-weight GPT-4-class system can reshape the industry that's largely dominated by Llama.

Still, the recent outrage surrounding Grok praising Hitler could be a reason why Altman has pushed the release back. Although the OpenAI CEO doesn't link the two events, the entire AI industry has been monitoring Musk's sassy chatbot closely.

While Altman is right in his caution, the second delay of OpenAI's open-weight model in a month is sure to have its rivals smirking.