OpenAI is scrambling to handle a surge after launching new image generation in ChatGPT. Just two days in, CEO Sam Altman admitted the company’s GPUs were “melting” under demand. The new feature transforms photos into art styles, including a popular Studio Ghibli mode.
The issue started March 25 with overwhelming user traffic. On March 31, Altman revealed 1 million new users signed up in just one hour. The next day, he warned ChatGPT risked slowdowns or outages.
Sam Altman stated: 
"Our GPUs [graphics processing units] are melting. We are going to temporarily introduce some rate limits."
This rapid growth spotlights AI’s massive energy needs. A new International Energy Agency (IEA) report warns data centers could double electricity use by 2030, topping 945 terawatt-hours. That beats current Japan electricity consumption. Data centers may use nearly 3% of global power.
The IEA added:
"In the United States, power consumption by data centers is on course to account for almost half of the growth in electricity demand between now and 2030."
In the US, data center construction is exploding. Donald Trump announced Stargate, a $500 billion public-private fund for up to 10 new centers. Meta and Microsoft plan to link data centers to nuclear plants to power AI’s growth.
With AI now in Bing, WhatsApp, and soon Google, this surge in demand is only going to grow — raising fresh concerns about AI’s environmental cost.
