July 16, 2026
Ctrl+Alt+Delusion
Generative AI Is an Engineering Disaster
AI’s hunger is driving up gadget prices, and the comments are absolutely on fire
TLDR: The article says generative AI is so expensive to run that it may be driving up the price of memory, storage, and even low-cost computers. In the comments, people split between “this is rich-company resource hoarding” and “calm down, every new technology looks wasteful before it matures.”
The Atlantic basically dropped a fire alarm: the giant chatbots everybody keeps using may be gobbling up so much computer memory and electricity that regular people are now paying the price. The article claims AI companies could be buying a huge chunk of the world’s top-tier memory, while laptop prices climb, hard drives vanish from stock, and cheap computers risk becoming an endangered species. Add in data centers multiplying fast enough to make jet-engine power sound normal, and commenters reacted like they’d just watched Silicon Valley eat the electric grid for breakfast.
But the real show was the comment-section cage match. One camp said this is exactly what happens when rich companies hoard resources: ordinary people get squeezed out first. That line hit hard, with one commenter turning the story into a broader class-war warning about wealth, power, and everyone else getting priced out. The other camp pushed back hard, arguing the article is unfair because AI is still new and messy, like every big technology before it. In their view, comparing today’s AI boom to polished internet businesses is like mocking a toddler for not running a marathon.
And then came the hot takes. One commenter basically said, relax, every giant AI training spree has paid off so far, so maybe this spending frenzy is just the cost of unlocking the next big thing. Another tossed in a deadpan line that felt tailor-made for meme status: in a successful civilization, non-computing energy use should be a rounding error. Translation: some readers think the power bill apocalypse is not a bug, but the future.
Key Points
- •The article says generative AI companies may be purchasing about 70 percent of the world's supply of high-end computer memory, contributing to shortages and higher prices for memory, storage, and some laptops.
- •It reports that AI-related data-center expansion is accelerating, with planned U.S. data-center capacity growth of roughly eightfold over the next few years.
- •The article states that electricity demand at some data centers is so high that companies are repurposing jet engines for power generation.
- •It argues that generative AI scales poorly compared with other major computing technologies such as streaming, smartphones, the Internet of Things, and cloud software.
- •The article says leading AI models have grown from about 175 billion parameters in 2020 to more than 1 trillion today, while returns from increasing model size are diminishing.