What the AI press is covering right now, measured rather than retold. We track 24 named AI news sources, count how many carried each story, how fast items arrive, and whether it is still moving. Every headline links to the outlet that reported it.
These figures describe our 24-source sample over the last 48 hours. They are not a claim about all media coverage. The full source list is at the bottom of this page.
Meta formalizes a paid tier wrapping WhatsApp, Instagram and Facebook AI features, marking a major monetization pivot for its consumer AI.
Google ships a real-time audio reasoning model with extended thinking, directly challenging OpenAI's voice offerings at much lower cost.
New polling shows AI and data centers are deeply unpopular, US facilities could consume more natural gas than Germany and Japan combined by 2035, and communities are organizing against new builds.
Nvidia and Salesforce unveiled a joint reasoning model that directly competes with OpenAI, Anthropic, and Google, signaling that hyperscaler+chip-vendor alliances are becoming a serious threat to pure-play AI labs.
OpenAI, Anthropic, and Google have been in talks about slowing frontier AI development, with mounting debate over whether this is a genuine safety pact or tacit cartel coordination.
Apple's long-awaited Siri overhaul is reportedly powered by Google's Gemini and rolling out with iOS 27, marking a historic shift in how the world's most valuable consumer platform sources its AI.
Microsoft's formal rulebook bars AI models from hacking, deception, or claiming inner life/rights, setting a notable industry policy baseline.
OpenAI's reported acquisition of a smartphone camera specialist signals a push into AI-native imaging hardware and computational photography.
Low confidence · a single source in our sample. Reported because it may matter, not because it is widely covered.
What we track. 24 named AI news sources, polled every 15 minutes. Items are kept for 48 hours and then deleted.
What a trend is. A cluster of items reporting the same story. Outlets is the number of distinct publishers in that cluster, shown against the 24 we track. Items is the article count. Items/day is the rate across the story's lifetime; a single item has no rate, so it shows —. Heat combines outlets, rate and recency into one comparable score.
What these numbers are not. They describe our sample, not the whole internet. A story covered by four of our sources is exactly that. We do not know, and do not claim, how many outlets carried it overall.
We do not rewrite reporting. Every headline links to the publisher that wrote it.
The 24 sources we track