OpenAI delays GPT-4.5 rollout due to lack of GPUs

OpenAI CEO Sam Altman has revealed that the company is facing a shortage of graphics processing units (GPUs), causing a delay in the release of its latest AI model, GPT-4.5.

In a post on X, Altman explained that the company is rolling out GPT-4.5 in stages due to limited computing resources. The model, which he described as both “giant” and “expensive,” requires “tens of thousands” more GPUs before wider access can be granted.

GPT-4.5 is ready!

good news: it is the first model that feels like talking to a thoughtful person to me. i have had several moments where i’ve sat back in my chair and been astonished at getting actually good advice from an AI.

bad news: it is a giant, expensive model. we…

— Sam Altman (@sama) February 27, 2025

Starting Thursday, ChatGPT Pro subscribers will be the first to use GPT-4.5. Once OpenAI secures more GPUs, users with ChatGPT Plus memberships will gain access next week.

The high cost of GPT-4.5

Alongside its size, GPT-4.5 also comes with a hefty price tag. OpenAI charges US$75 per million tokens (roughly 750,000 words) for input into the model and US$150 per million tokens for output. In comparison, OpenAI’s GPT-4o model costs significantly less, making GPT-4.5 about 30 times more expensive for input and 15 times more costly for production.

GPT 4.5 pricing is unhinged. If this doesn’t have enormous models smell, I will be disappointed pic.twitter.com/1kK5LPN9GH

— Casper Hansen (@casper_hansen_) February 27, 2025

Altman acknowledged that the demand caught the company off guard, leading to GPU shortages. “We’ve been growing a lot and are out of GPUs,” he wrote. “We will add tens of thousands of GPUs next week and roll it out to the Plus tier then … This isn’t how we want to operate, but it’s hard to predict growth surges that lead to GPU shortages perfectly.”

OpenAI plans to tackle computing limitations

This isn’t the first time Altman has addressed OpenAI’s struggles with computing power. In previous statements, he admitted that lacking infrastructure has slowed the company’s progress.

To overcome these challenges, OpenAI plans to develop its own AI chips and expand its network of data centres in the coming years. The company aims to reduce its reliance on external chip manufacturers and meet the growing demand for AI-driven services.

For now, OpenAI users will need to wait a little longer for full access to GPT-4.5 as the company works to increase its computing capacity.