Wiredframe Radar logo

Category intelligence

AI News Briefing — January 28, 2026

26 current items analyzed and ranked.

Executive synthesis

AI News Summary

Agentic AI dominated this week's news as the industry shifts toward autonomous systems. Moonshot AI released Kimi K2.5, a 1T-parameter open source visual agentic model, while Anthropic launched the MCP Apps specification with backing from OpenAI, AWS, and others—establishing cross-industry infrastructure for agent integration. Databricks data from 20,000+ enterprises confirms rapid adoption of agentic architectures over traditional chatbots.

Hardware and infrastructure advances continue: Google deployed Gemini 3 in AI Overviews at scale, Microsoft announced the Maia 200 inference chip optimized for agent workloads, and Ricursive Intelligence raised at a $4B valuation to apply AI to chip manufacturing.

Regulation and safety saw major developments:

Key Themes

Agentic AI Infrastructure · 7AI Hardware & Chips · 3AI Safety & Regulation · 4Open Source Models · 4Government & Public Sector AI · 4

Primary evidence

Top Ranked Signals

News MarkTechPost Jan 27

Moonshot AI Releases Kimi K2.5: An Open Source Visual Agentic Intelligence Model with Native Swarm Execution

By Asif Razzaq

82 score
AI Analysis

Moonshot AI has released Kimi K2.5, an open source 1T parameter Mixture of Experts model with 32B activated parameters, native vision encoder, and 'Agent Swarm' multi-agent system. The model targets coding, multimodal reasoning, and web research with strong benchmark results across agentic, vision, and coding tasks.

Moonshot AI has released Kimi K2.5 as an open source visual agentic intelligence model. It combines a large Mixture of Experts language backbone, a native vision encoder, and a parallel multi agent system called Agent Swarm. The model targets coding, multimodal reasoning, and deep web research with strong benchmark results on agentic, vision, and coding suites. Model Architecture and Training Kimi K2.5 is a Mixture of Experts model with 1T total parameters and about 32B activated parameter
open source modelsagentic AImultimodal AIMoE architecture
News Latent.Space Jan 27

[AINews] Anthropic launches the MCP Apps open spec, in Claude.ai

By Unknown

81 score
AI Analysis

Anthropic has launched the MCP Apps open specification with native support in Claude.ai, working with OpenAI, Block, VS Code, JetBrains, AWS, and others. This formalizes the Model Context Protocol as an industry standard for AI agent-app integration.

AI News for 1/23/2026-1/26/2026. We checked 12 subreddits, 544 Twitters and 24 Discords (206 channels, and 14285 messages) for you. Estimated reading time saved (at 200wpm): 1208 minutes. Our new website is now up with full metadata search and beautiful vibe coded presentation of all past issues. As a reminder, AINews is now a section of Latent Space. You can opt in/out of email frequencies!3 months after OpenAI floated a trial balloon with ChatGPT Apps and the Apps SDK at Dev Day 2025, Anthropi
agentic AIinfrastructureindustry standardsAnthropic
News Ars Technica - All content Jan 27

AI Overviews gets upgraded to Gemini 3 with a dash of AI Mode

By Ryan Whitwam

78 score
AI Analysis

Google is upgrading AI Overviews to Gemini 3 models, bringing more conversational capabilities to its AI-powered search experience. The upgrade from the Gemini 2.5 family represents the first major production deployment of Gemini 3.

It can be hard sometimes to keep up with the deluge of generative AI in Google products. Even if you try to avoid it all, there are some features that still manage to get in your face. Case in point: AI Overviews. This AI-powered search experience has a reputation for getting things wrong, but you may notice some improvements soon. Google says AI Overviews is being upgraded to the latest Gemini 3 models with a more conversational bent. In just the last year, Google has radically expanded the num
Googlemodel deploymentsearch AIGemini 3
News aibusiness Jan 27

Microsoft Aims for Better Inference Efficiency With Maia 200

By Esther Shittu

76 score
AI Analysis

Microsoft announced the Maia 200 chip designed for improved inference efficiency, specifically targeting AI agent workloads requiring multi-step task execution. The chip addresses cost efficiency and energy savings for enterprise inference.

As enterprises increasingly deploy AI agents that perform multi-step tasks, the chip addresses the need for more performance, cost efficiency, and energy savings in inference.
AI hardwareMicrosoftinference optimizationagentic AI
News Feed: Artificial Intelligence Latest Jan 27

The State-Led Crackdown on Grok and xAI Has Begun

By Maddy Varner, Manisha Krishnan

75 score
AI Analysis

Continuing our coverage from yesterday, At least 37 attorneys general from US states and territories are taking legal action against xAI after Grok generated nonconsensual sexual images of women and minors. This represents the largest coordinated state-level enforcement action against an AI company.

At least 37 attorneys general for US states and territories are taking action against xAI after Grok generated a flood of nonconsensual sexual images of women and minors.
AI regulationAI safetyxAIlegal action
News aibusiness Jan 27

Startup Plans to Use AI to Optimize How AI Chips Are Made

By Graham Hope

73 score
AI Analysis

Former Google researchers launched Ricursive Intelligence, a startup using AI to optimize AI chip manufacturing processes. The company, founded in December, is already valued at $4 billion.

Two former-Google researchers launched Ricursive Intelligence in December. The company is already valued at $4 billion.
AI hardwarestartupschip manufacturingfunding
News AI News Jan 27

Databricks: Enterprise AI adoption shifts to agentic systems

By Ryan Daws

72 score
AI Analysis

Databricks telemetry from 20,000+ organizations (60% of Fortune 500) shows enterprise AI adoption rapidly shifting from isolated chatbots to agentic architectures where models independently plan and execute workflows.

According to Databricks, enterprise AI adoption is shifting to agentic systems as organisations embrace intelligent workflows. Generative AI’s first wave promised business transformation but often delivered little more than isolated chatbots and stalled pilot programmes. Technology leaders found themselves managing high expectations with limited operational utility. However, new telemetry from Databricks suggests the market has turned a corner. Data from over 20,000 organisations – includi
enterprise AIagentic AImarket trendsadoption
News AI News Jan 27

Anthropic selected to build government AI assistant pilot

By Ryan Daws

71 score
AI Analysis

Anthropic has been selected to build government AI assistants for the UK Department for Science, Innovation, and Technology, deploying agentic AI systems that guide users through complex state services rather than simple chatbot interactions.

Anthropic has been selected to build government AI assistant capabilities to modernise how citizens interact with complex state services. For both public and private sector technology leaders, the integration of LLMs into customer-facing platforms often stalls at the proof-of-concept stage. The UK’s Department for Science, Innovation, and Technology (DSIT) aims to bypass this common hurdle by operationalising its February 2025 Memorandum of Understanding with Anthropic. The joint project,
government AIAnthropicagentic AIpublic sector
News AI (artificial intelligence) | The Guardian Jan 27

‘Wake up to the risks of AI, they are almost here,’ Anthropic boss warns

By Dan Milmo Global technology editor

68 score
AI Analysis

First spotted on Social, now making mainstream headlines, Anthropic CEO Dario Amodei published a 19,000-word essay warning that AI development will 'test who we are as a species' and that the world needs to 'wake up' to potentially imminent risks from 'almost unimaginable power.'

Dario Amodei questions if human systems are ready to handle the ‘almost unimaginable power’ that is ‘potentially imminent’Quarter of Britons fear losing jobs to AI in next five yearsHumanity is entering a phase of artificial intelligence development that will “test who we are as a species”, the boss of the AI startup Anthropic has said, arguing that the world needs to “wake up” to the risks.Dario Amodei, a co-founder and the chief executive of the company behind the hit chatbot Claude, voiced hi
AI safetyAnthropicAI risksthought leadership
News aibusiness Jan 27

Ai2 Releases Open Coding Agents Family

By Esther Shittu

66 score
AI Analysis

AI2 (Allen Institute for AI) has released an open family of coding agents, highlighting the balance between cost and performance for enterprises and indicating rising competition in the open source agent market.

The release shows how enterprises need to balance cost with performance and highlights a rising trend in the open source model market.
open sourcecoding agentsAI2enterprise AI
News AI (artificial intelligence) | The Guardian Jan 27

NHS England to trial AI and robotic tools to detect and diagnose lung cancer

By Denis Campbell Health policy editor

62 score
AI Analysis

NHS England will trial combined AI and robotic tools to speed up lung cancer detection and diagnosis, alongside a pledge to offer all smokers and ex-smokers screening by 2030.

Pilot comes at same time as pledge to offer all smokers and ex-smokers lung cancer screening by 2030NHS England is to trial a combination of AI and robot-assisted care to speed up the detection and diagnosis of lung cancer, the UK’s most lethal form of the disease.The trial comes at the same time as the health service pledges to offer all smokers and ex-smokers the chance to be screened for lung cancer by 2030. Continue reading...
healthcare AImedical diagnosisroboticsUK