TRENDS

AI trends we’re tracking

Recurring themes in global AI — compute buildout, open weights, agents, regulation, safety and language access — with dated evidence and editorially set momentum.

Momentum (rising / steady / fading) is an editorial judgement reviewed with each daily edition, not a computed score. Evidence links are examples from configured feeds, not a measure of total activity.

RISING First tracked 2026-11-01

Compute & data-centre buildout

Capital, power and land for AI data centres, from hyperscaler capex to national compute programmes.

Latest evidence

Judge allows NAACP lawsuit over xAI data center pollution to move forward TechCrunch · 2026-10-31
Nvidia weighs playing the cosigner for Nscale’s AI data center in West Virginia TechCrunch · 2026-10-30
VIDEO Au Texas, le projet de construction du plus grand data center au monde qui divise le camp de Trump Le Monde · Pixels · 2026-11-01
Backbone of a beast – charting Nvidia's rise to networking titan The Register · 2026-10-31
Supersonic airliner in doubt as company shifts to powering AI data centers Ars Technica · AI · 2026-10-30
Sam Altman says it’s better for RAM to be in a data center than on your desk The Verge · 2026-10-30

Recent headlines from configured feeds, selected editorially as examples. Not a count of activity. Momentum updated 2026-11-01.

StoriesAnthropic files to go public: what its numbers reveal about the AI money race

RISING First tracked 2026-11-01

AI agents at work

Systems that take actions on a user's behalf, and claims about agents 'joining the workforce'.

Latest evidence

DeepSeek Harness Team Open-Sources a Super-agent Prototype Pandaily · 2026-11-01
DeepSeek unveils AI ‘super agent’ in coding race with OpenAI, Anthropic SCMP · 2026-11-01
AI Agents Now Create Over 20,000 Tencent Cloud TDSQL-C Databases a Day Pandaily · 2026-11-01
Verve plans phone tap-to-pay and AI agent payments, CEO says TechCabal · 2026-10-31
AI agents are helping criminals to commit cyber crimes more cheaply and efficiently, say experts The National (UAE) · 2026-10-31
Studie: Große Mehrheit der Fachleute übernahm die Meinung eines KI-Agenten Der Standard · Web · 2026-10-30

Recent headlines from configured feeds, selected editorially as examples. Not a count of activity. Momentum updated 2026-11-01.

StoriesOpenAI and Google launch new AI models with opposite rules on who can use them

STEADY First tracked 2026-11-01

Open-weight models

Releases and policy debates over models whose weights are published, including non-US labs.

Latest evidence

SenseTime Open-Sources SenseNova-Loop-9B-Preview, a Looped Model Pandaily · 2026-11-01
KI-Update kompakt: Mistral ML4, Cohere, KI-Empfehlungen, KI klont Indie-Spiele Heise Online · 2026-10-30
“でぶねこ”の重み公開 1兆パラメータのAIモデル「Mistral Large 4」（Le Chonk）がダウンロード可能に ITmedia AI+ · 2026-10-30
DeepSeek V5 identifiers surface in official Harness code repository TechNode · 2026-10-29
Welcome Le Chonk: Mistral Large 4 on Hugging Face Hugging Face · 2026-10-29
Qwen3Guard: Real-time Safety for Your Token Stream Qwen (Alibaba) · 2025-09-22

Recent headlines from configured feeds, selected editorially as examples. Not a count of activity. Momentum updated 2026-11-01.

STEADY First tracked 2026-11-01

Regulation & governance

Laws, rules and enforcement: the EU AI Act, national strategies, export controls and court cases.

Latest evidence

AI giants promise publishers new way to block some scraping The Hindu (AFP) · 2026-10-31
New York restricts AI chatbot interactions with children, explained The Hindu · 2026-10-31
Disquiet grows in Albanese’s back yard as Grayndler artists unite against ‘insane’ AI proposal The Guardian · AI · 2026-10-31
Prof. Vukosi Marivate on Governing AI in Africa | Deep Learning Indaba 2026 iAfrikan · 2026-10-20
First harmonised standard supporting the AI Act published in the Official Journal EU Digital Strategy · 2026-10-19

Recent headlines from configured feeds, selected editorially as examples. Not a count of activity. Momentum updated 2026-11-01.

StoriesBig AI companies promise publishers one simple way to block AI training

STEADY First tracked 2026-11-01

Safety, evaluation & measurement

How models are tested, who sets the benchmarks and what evaluations miss.

Latest evidence

Introducing Gemini 4 Carbon (staged release to cyber defenders) Google · 2026-10-28
Self-improving AI the key step towards AGI but safety is key: Z.ai founder SCMP · 2026-11-01
I read Anthropic's 92-page risk factor document, so you don't have to – video explainer The Guardian · AI · 2026-10-30
From synthetic enterprise data evaluations to real Microsoft Copilot AI confidence Microsoft Research · 2026-10-30
ShallowDefend: Harnessing Shallow Safety Alignment for Multi-Turn Defender arXiv · cs.CL · 2026-10-29
Your Robot’s Safety Functions Already Work. What If the Input Lies? IEEE Spectrum · AI · 2026-10-27

Recent headlines from configured feeds, selected editorially as examples. Not a count of activity. Momentum updated 2026-11-01.

StoriesOpenAI and Google launch new AI models with opposite rules on who can use them

STEADY First tracked 2026-11-01

Languages beyond English

Models, data and evaluations for under-served languages and regions.

Latest evidence

No recent headline evidence selected yet.

Recent headlines from configured feeds, selected editorially as examples. Not a count of activity. Momentum updated 2026-11-01.