Data-Center Disenchantment

Data centers are incredibly unpopular among Americans.

In a March survey of Americans by Gallup, 71% of respondents said they would oppose the construction of one in their local area. By comparison, just 53% said they would oppose a nuclear energy plant.

Political pressure is also mounting, and politicians who once championed data centers are now bashing them. Pennsylvania Gov. Josh Shapiro, for example, who once embraced AI, on Tuesday signed an executive order placing strict guardrails on new data centers.

The public-relations crisis is starting to overwhelm the tech giants.

Companies like Microsoft, OpenAI and Amazon need the public’s buy-in to avoid speed bumps, maintain political support and curb blowback. To shore up support, they are hosting listening sessions, offering job guarantees and writing big checks to invest in local communities, my colleague Meghan Bobrowsky and I reported this week.

OpenAI in July held an open house at Effingham County College & Career Academy, near the company’s planned data center in Georgia. It featured educational booths manned by OpenAI employees, posters showing what the facility might look like—and even a taco bar.

It’s only the beginning of a new movement in the massive AI infrastructure buildout, but it seems like tech companies will soon need more than tacos to change the public’s tune.

—Isabella, who covers Amazon, is a tech reporter based in New York.

As parents and teachers grow increasingly anxious about students’ use of chatbots, OpenAI is rolling out ChatGPT for Teens starting Tuesday. The company intends for it to help students think through problems, rather than just spitting out answers. It is increasing the guardrails against emotional dependence, too.

I love a bargain and hate waste, so I’m an easy sell on secondhand tech. With prices on the rise because of memory-chip costs, people don’t need convincing to buy used devices. But doing so comes with a few hazards. Here’s how to avoid them and still score a great deal.

Today’s AI large language models might excel at coding, but proponents say doing real physical work requires large action models, aka world models. Despite the relatively primitive state of world models, tech luminaries are betting they can take AI to places where LLMs can’t venture. Pioneers hope to merge these two schools of artificial thought.

Founders have long put in punishing hours in the name of building the next big thing. But the growing capabilities of AI agents—and the speed at which the models powering them are evolving—give new meaning to working yourself to the bone.

OpenAI told investors its revenue grew by 18% from the first to the second quarter, while its losses deepened, results that disappointed some shareholders who had hoped the startup would show more progress catching up to rival Anthropic.

Hidden Debt: Nine top tech companies had $3 trillion of off-balance-sheet commitments mostly related to AI, according to a Wall Street Journal analysis.

American DeepSeek: Nvidia is planning to use a $6 billion deal it struck this week to build one of the world’s most powerful open-weight AI models, one that would compete with Chinese heavyweights like DeepSeek and Kimi K3.

Gigawatt Gamble: OpenAI has signed a 10-gigawatt data-center deal in Ohio with SoftBank’s SB Energy that is partly backed by Nvidia, a project expected to become one of the largest AI hubs to date.

…when AI companies are using an actual guillotine on rare books.

The Technology newsletter is a weekly digest of tech reviews, columns and headlines from Anne Steele and Wilson Rothman. Write to Wilson at wilson.rothman@wsj.com and Anne at anne.steele@wsj.com. Got a tip for us? Here’s how to submit.