AI is moving on three fronts at once: new model releases, bigger compute deals, and tighter regulation. Here are the updates worth watching right now. [2][3][5]
🧵 1/6
Model watch: recent trackers list GPT-5.2 Codex, Gemini 3 Flash, Kimi K2.5, and GLM-4.7 Flash, while Price Per Token highlights Gemma 4, Claude Opus 4.6 Fast, and GLM 5.1. The pace keeps accelerating. [2][9]
🧵 2/6
Compute is still the bottleneck: Broadcom is set to supply custom tensor processing units for Anthropic and Google. More models mean more hardware pressure, so who gets capacity first may shape the next wave. [3]
🧵 3/6
AI regulation is getting more operational. Ctrl+AI+Reg says its Global Updates feed now shows every change added in the last 7 days, and border systems are being pushed toward more auditability and transparency. [5][6]
🧵 4/6
On the infrastructure side, IREN says its Blackwell deployment with Dell could lift annualized run-rate revenue from $3.7 billion to $4.4 billion once it is live. Time-to-compute is becoming a business metric, not just a tech one. [8]
🧵 5/6
Which of these AI shifts do you think will matter most over the next 6 months? Reply with your pick or retweet for others to weigh in.
🧵 6/6
Sign Up To Try Advanced Features
Get more accurate answers with Super Pandi, upload files, personalized discovery feed, save searches and contribute to the PandiPedia.