TECH NEWS – Sam Altman, the creator of ChatGPT, has revealed why AI development is slowing down: “We’ve run out of GPUs”. The OpenAI CEO confirms that GPT-4.5 will roll out in phases due to this issue.

OpenAI remains the undisputed leader in the AI industry. Recently, CEO Sam Altman urged young professionals to prepare for an AI-driven job market, predicting that artificial intelligence would usher in a golden age for humanity. However, just days after making these bold claims, Altman had to acknowledge a critical issue: OpenAI does not have enough graphics cards to continue its rapid expansion.

According to TechCrunch, Altman confirmed that the company lacks the necessary GPUs to roll out GPT-4.5 at full scale. As a result, OpenAI has no choice but to introduce the model in phases, starting with ChatGPT Pro subscribers. Altman admitted that GPT-4.5 is both an immense technological leap and an expensive one, which significantly complicates its widespread release.

ChatGPT Plus subscribers will get GPT-4.5 next week

To expand access, OpenAI must secure tens of thousands of additional GPUs in the coming days. Compounding the issue, running GPT-4.5 is significantly more expensive than previous versions: the reported costs are €70 per million input tokens and €140 per million output tokens. Due to these limitations, OpenAI will first make GPT-4.5 available to ChatGPT Pro users, with ChatGPT Plus subscribers gaining access next week once additional GPUs are in place.

Altman acknowledged that predicting demand surges is extremely difficult, making GPU shortages a persistent challenge. To address this, OpenAI is exploring the development of proprietary AI chips to reduce reliance on external suppliers. Additionally, the company is investing in a vast network of data centers to support future growth. While this crisis is certainly a roadblock, it wasn’t entirely unexpected—Altman has repeatedly warned that computing capacity constraints could delay OpenAI’s technological progress.

Source: TechCrunch