EDITORIAL ARCHIVE
AI news
News, tests and analysis to understand what changed, what is hype and what your business can already use.
Latest from the radar
Page 6 / 25JetStream Launches Clearance, a Zero Trust Authorization Engine for AI Agent Actions
JetStream Security announced Clearance on September 2, 2026, a reasoning engine that evaluates every AI agent action against an approved design and decides whether it can run before execution, blocking dangerous sequences such as data exfiltration attempts. According to Yahoo Finance, Enterprise DNA and Dealroom, the product was demonstrated at the Fal.Con conference in Las Vegas and reaches general availability this fall.
Institute of Foundation Models Launches K2 Horizon, the Largest Fully Open AI Model Release Yet
The Institute of Foundation Models, affiliated with MBZUAI, released K2 Horizon on September 3, 2026, a family of six AI models ranging from 0.9 billion to 375 billion parameters, with weights, code, training data and methodology fully open under an Apache 2.0 license. According to HPCwire, PR Newswire and Moor Insights, it is the largest fully open source model release in AI history, delivering top tier performance at every size class, from watches and glasses to local servers.
US DOJ Backs OpenAI and Microsoft in New York Times Copyright Lawsuit
The US Department of Justice filed a statement of interest on September 1, 2026, arguing that training AI models on copyrighted text qualifies as fair use, a position that favors OpenAI and Microsoft in the lawsuit brought by the New York Times. According to the Washington Post, Bloomberg Law and Deadline, it is the first time the federal government has formally taken a position in this type of dispute, and the filing warns that requiring broad licensing of training data could concentrate frontier AI in the hands of a few wealthy companies.
Claude Code 2.1.263 Arrives Focused on Fixes and Reliability
Anthropic released Claude Code 2.1.263 with a stated focus on bug fixes and reliability. The update announces no new model or command, but reinforces the operational base for sessions, automations, and remote workflows.
Microsoft launches VibeVoice-ASR-Streaming to transcribe and identify speakers live
CONFIRMED: Microsoft launches VibeVoice-ASR-Streaming, an open model combining live transcription and speaker attribution in ten languages, including Portuguese.
Lutnick Says Trump Administration Now Trusts Anthropic After Months of Friction
US Commerce Secretary Howard Lutnick said on September 2, 2026, at the G20 Innovation Ministerial in Chapel Hill, that the Trump administration now trusts Anthropic after months of friction, including a Pentagon retaliation a federal judge later ruled illegal. Anthropic co-founder Tom Brown, who helped rebuild the relationship, shared the stage with Lutnick, while Sam Altman and Jensen Huang also attended the event, according to Axios, Forbes and CNBC.
Anthropic Signs $35 Billion Deal With Nvidia-Backed Lambda for Compute Capacity
Anthropic signed a $35 billion deal with Lambda, an Nvidia-backed cloud provider, for compute at a Hut 8 data center in Texas, first reported by The Information and confirmed by Bloomberg and Forbes. The contract adds to other multibillion dollar commitments from the company, with Nscale, Fluidstack and SpaceX, totaling more than $175 billion in reserved cloud capacity amid Claude's accelerating growth.
Nvidia Invests $3.5 Billion in MediaTek to Accelerate Custom AI Chips
Nvidia invested $3.5 billion in MediaTek convertible bonds, a deal revealed in late August 2026 that has the Taiwanese chipmaker adopt Nvidia's NVLink Fusion for custom AI chips. MediaTek shares jumped about 10% after the announcement, according to CNBC, TechCrunch and Forbes. The move shows Nvidia charging a toll on custom silicon rivals instead of fighting them.
Brazil's ANPD Builds Its Own AI Center and Opens Public Consultation for the 2027-2028 Agenda
ANPD, Brazil's data protection authority, revealed at a panel during the 17th Seminar on Data and Privacy Protection on September 2, 2026 that it is building its own AI center and testing a regulatory sandbox on the transparency of automated decisions. A day earlier, the agency opened a public consultation, open until October 16, for its 2027-2028 regulatory agenda. The news was reported by ConvergenciaDigital, Capital Digital, O Povo and Brasil em Folhas.
EU Classifies ChatGPT as a Very Large Online Search Engine Under the DSA
The European Commission designated ChatGPT as a Very Large Online Search Engine under the Digital Services Act (DSA) on August 31, 2026, the first time an AI chatbot has received that classification. OpenAI reported about 159.1 million average monthly users of ChatGPT's search function in the EU, well above the 45 million threshold that triggers the rule. The company has four months to meet systemic risk assessment and independent audit obligations, according to the European Commission, Search Engine Journal and Business Standard.
WhatsApp Business API Will Start Charging for Service Messages in October
Starting October 1, 2026, Meta will charge for service messages and utility templates inside the 24 hour window on the WhatsApp Business Platform, closing one of the last free traffic categories on the API. The reference rate cited for Brazil is about $0.0068 per message, with no volume discount, according to Meta's documentation and reporting by Canaltech and SendPulse. The change directly affects agencies and companies automating customer service with chatbots and AI on the channel.
Meta launches Muse Spark 1.3 and sells a jump in coding and agents
Meta put Muse Spark 1.3 in Muse Code and the API at 1.2 pricing. Zuckerberg calls it the house's biggest jump. Max reasoning has not shipped, and the Opus versus Sol table is Meta's own.