Anthropic buys more room for Claude

Claude’s limits were a capacity problem

Anthropic just tied Claude’s usage limits to a very physical constraint: enough power, chips, and data center space.

The company says a new SpaceX agreement gives it more than 300 megawatts of capacity at the Colossus 1 data center, including over 220,000 NVIDIA GPUs. It is also doubling Claude Code’s five-hour rate limits for Pro, Max, Team, and seat-based Enterprise plans, removing peak-hour reductions for Pro and Max users, and raising Opus API limits.

This is the useful part. AI companies talk about intelligence, but customers feel capacity when a request slows down, gets capped, or fails during a busy workday.

My read is that Anthropic wants to make Claude feel less rationed before enterprise buyers standardize around someone else. The SpaceX deal also shows how much leverage now sits with whoever can deliver power and hardware quickly.

The orbital compute mention is worth watching, though it still sounds early.

For now, better limits may matter more than bigger claims.

Snap’s AI search deal quietly ends

Snap had a $400 million AI deal on paper, then walked away before it reached most Snapchat users.

The company said its Perplexity partnership “amicably ended” in Q1, after earlier plans to bring Perplexity’s AI search into Snapchat’s Chat interface. The integration was tested with select users, but Snap said in February there was still no agreed path to a wider rollout.

That matters. Snap needs AI features that fit inside a fast, visual app where people tap through messages, lenses, and maps under bright phone light.

My read is that this deal probably looked cleaner in an earnings announcement than inside the actual product. Conversational search can be useful, but Snapchat has never been a place where users go to slow down and ask detailed questions.

Snap still grew daily active users 5% to 483 million, so this is not a panic story. It does suggest the company is being more careful about which AI bets deserve real distribution.

What else gets cut before Specs arrive on June 16?

Google’s AI Search gets link-heavy

Google is trying to make AI answers feel less like a dead end.

The company announced five updates to AI Mode and AI Overviews, including more inline links, previews on desktop, subscription labels for news sources, and discussion snippets from public forums and social platforms. It also says query fan-out will help surface more relevant sites across the web.

This is practical. Publishers and creators have been worried that AI answers give users enough information to skip the click, and Google clearly knows that tension is getting harder to ignore.

My read is that these changes are useful, but they also keep Google in control of the door. A hover preview, a “Subscribed” label, or a forum quote can send traffic out, but Google still decides which sources sit next to the answer.

The real test is behavior, not screenshots. Do people click more when the links are closer to the sentence they just read?

That is the number publishers will watch.

Prohuman team

Free Guides

Explore our free guides and products to get into AI and master it.

All of them are free to access and would stay free for you.

Feeling generous?

You know someone who loves breakthroughs as much as you do.

Share The Prohuman it’s how smart people stay one update ahead.