Gen Z boos AI, while compute costs tighten

From graduation-stage interruptions to rising hostility in polling, Gen Z’s backlash against AI is intensifying as lawmakers back efforts to slow data-center building. At the same time, service outages and platform rules are pushing companies to rethink how th

For the third morning in a row, some graduates didn’t clap when artificial intelligence entered the room.

In May, students booed speakers at graduation ceremonies across the country when AI was mentioned. Google chairman Eric Schmidt faced boos when he tried to tell graduates at the University of Arizona that AI has world-changing potential. At the University of Central Florida. Gloria Caulfield. vice president of strategic alliances for the investment firm and real estate developer Tavistock. was also met with boos after she compared the rise of AI with the Industrial Revolution. And at Middle Tennessee State University. students shouted at Scott Borchetta. CEO of Big Machine Records. simply for bringing up AI.

The reaction isn’t just scattered behavior. A Gallup poll measuring AI adoption and attitudes among Gen Zers shows that those excited about AI fell from 36% to 22% over the last year. Anger toward AI rose from 22% to 31%. The shift is especially striking because other age groups are also skeptical. but the youngest potential workforce shows the sharpest decline in attitude toward a new technology.

When a new tool meets resistance, companies usually respond with messaging. This time, the pressure is coming from two directions at once—public sentiment and the practical cost of running the systems behind AI.

As anti-AI sentiment grows and midterm elections approach. several politicians have begun supporting efforts to halt or slow the building of data centers. facilities that provide the computing power AI needs. The argument is direct: if capacity can’t keep up with demand. the rising cost of computing power could put real limits on how far industries—including the media—can take AI.

That pressure is already visible inside the AI stack. Anyone who uses Claude regularly knows the frustration of repeated outages. and the service’s status dashboard shows an “embarrassing amount of red” over the past 90 days. At the same time, the Claude Code revolution has driven demand for Anthropic’s AI, with that demand exploding in 2026.

Anthropic is trying to meet it, including by signing a deal to buy computing power from Elon Musk’s SpaceX. Yet the company also stopped allowing builders to leverage their Claude subscriptions for third-party software. The source points out that those setups can sometimes use thousands of dollars of compute for the relatively cheap $200-a-month Claude Max subscription. After the change, builders now have to use Anthropic’s platform or pay as they go.

That decision angered people, but it also forces a rethink of what teams are really doing with compute. The path forward may mean moving to a cheaper model—possibly open source—or finding a way to use the Claude platform. Some builders may simply have to shut down what they built.

The industry’s version of the story is that these shortages and closures are part of a bigger infrastructure bottleneck. As demand for AI rises, free-compute loopholes will keep getting closed. Industry leaders have been urging trillion-dollar infrastructure projects such as OpenAI’s Stargate. arguing that for AI to deliver on its promises. compute must work “like water from the heavens”—meaning more data centers and more power plants.

For Gen Z, there is another reason anger lingers: energy. Environmental concerns are described as one of the chief drivers of Zoomers’ hostility toward AI. The source says AI’s massive energy use became a serious PR problem months earlier and has only worsened since then. It also describes how. in internal surveys at companies advising AI adoption. workers’ concerns about environmental impact come up consistently—and may even be affecting decisions over whether to use AI at all.

Even where the politics of data centers and environmental claims are debated. the cost of compute is presented as a concrete reality. Leaders at the front of AI transformation at companies are past handing every employee a ChatGPT account. They are now trying to help workforces leap forward with agentic processes. automated workflows. and rapid prototyping through “vibe coding.” The source also notes that engineers may be encouraged to get obsessed with “tokenmaxxing.”.

The tension is now straightforward: compute may stay expensive, while lawmakers may slow supply by targeting data-center construction. In that environment, the dividing line between progress and frustration comes down to how companies govern AI internally.

The source argues that governance has to go beyond training staff on the tools and what different models are good at. It means balancing experimentation and progress—giving workers room to innovate their own workflows while making sure the organization’s compute spending is actually being used well. The goal isn’t just “keeping costs down.” It’s understanding that costs might be high. but the outcomes have to justify them.

One example described comes from consulting with media companies and PR agencies. An agency piloted a vibe-coding tool. Usage surged at first as workers tested limits, but multiple teams ended up building similar things. The turning point was internal coordination: the agency held regular internal workshops and project reviews. learning from people and providing direction. That push led the group to use cases that performed well—specifically speeding up and automating media intelligence.

The pilot ended up with a hard choice that many teams never make early enough: the agency adopted an entirely different platform and sunset the initial tool based on what it found through the governance process.

The takeaway is blunt. AI—especially agentic work—requires compute-heavy execution. If compute remains costly. leaders can either impose strict rules that suffocate innovation or define what success looks like. train teams to use tools and models effectively. and build collaboration systems that catch waste before it drains budgets.

The political fight over data centers will “rage on,” the source says. But the argument for companies is to chart a path inside those constraints—reducing harsh limits and strictness that workers feel in day-to-day work. even as the price of running AI rises and public trust becomes more complicated.

AI backlash Gen Z attitudes Gallup poll data centers compute costs Anthropic Claude outages SpaceX deal governance vibe coding tokenmaxxing ESG and energy concerns media AI