The AI infrastructure race has mostly been described as a battle over GPUs, chips and data centers . But there is another constraint becoming increasingly difficult to ignore: How do you deliver enough electricity to all those GPUs?
The AI boom is getting dramatically more expensive. Amazon, Alphabet, Meta and Microsoft are on track to spend at least $740 billion on AI infrastructure in 2026 , according to estimates based on the companies' latest earnings guidance. That would represent roughly a 77% increase compared with what the same four companies spent the previous year. The scale is difficult to comprehend.
One of the most frustrating parts of using AI to write and run code is surprisingly simple: The AI keeps asking for permission. “Should I run this command?” “Can I modify this file?” “Do you want me to continue?” “Can I execute this operation?”
Chinese artificial intelligence companies intensified global competition with a wave of product announcements on the same day. DeepSeek’s V4-Flash public beta, MiniMax’s H3 multimodal model, and ByteDance’s Seedance 2.5 video generation model show that China’s AI players are advancing not only on technology, but also on pricing, open-source strategy, and ecosystem reach.