Here is today's AI news. Story 1. OpenAI. Launched **GPT-6 Astra** on September 3, 2026, positioning it as a frontier model for autonomous computer use, software engineering, scientific research, and complex professional workflows.. OpenAI said on Sept. 3, 2026 that it began rolling out GPT-6 Astra, its newest flagship model for computer use, coding, research, and complex workflows. The company says Astra is its first model to hit the “Critical” cybersecurity tier. It’s launching first to select Daybreak customers, then to ChatGPT paid tiers and API users “in the coming days.” Story 2. OpenAI, Anthropic, Google, Microsoft and others. More than 100 companies issued a joint warning about AI-powered cyberattacks, while OpenAI separately committed **$1 billion** to cyber-defense efforts.. On Aug. 27, OpenAI, Anthropic, Google, Microsoft and more than 100 companies—including CrowdStrike, IBM, Visa and Cloudflare—warned that AI-enabled cyberattacks could become “far more widespread and sophisticated” within months. Their joint letter urges governments to fund defensive AI, share threat intelligence and protect hospitals, utilities and other critical infrastructure. Separately, OpenAI pledged $1 billion in subsidized tools, training and support for frontline defenders. Story 3. Anthropic. A U.S. federal court ruled that the Pentagon’s designation of Anthropic as a national-security supply-chain risk was unlawful retaliation, requiring the government to rescind related directives.. A federal judge ruled that the Pentagon unlawfully retaliated against Anthropic by labeling it a national-security “supply chain risk” after the company opposed unrestricted military use of Claude, including autonomous weapons and domestic surveillance. Judge Rita Lin found First and Fifth Amendment violations, calling the evidence of risk “slim” and the action “classic illegal First Amendment retaliation.” The ruling threatens government-wide AI blacklisting practices and reinforces vendors’ rights to set safety limits. Story 4. Nvidia. Agreed to acquire AI model and development platform **Hugging Face** for approximately **$13 billion**, expanding its control over the open-source AI ecosystem.. Nvidia agreed to buy Hugging Face for about $12.9 billion, according to Reuters, CNBC and AP. The deal would pull one of the most important open-source AI hubs under Nvidia’s control, deepening its reach from chips into software, models and developer tools. Jensen Huang said Hugging Face will “remain an open platform” and Nvidia will “scale” it globally. Story 5. Microsoft. Released **MAI-Transcribe-2**, a speech-recognition model reported to rank first on the FLEURS benchmark across 60 languages.. Microsoft released MAI‑Transcribe‑2 in public preview through Microsoft Foundry/Azure Speech. It claims No. 1 on FLEURS across 60 languages, at 5.2% average word-error rate. The model adds speaker diarization and word-level timestamps—useful for meetings, captions and call centers—and is priced at $0.10 per audio hour through 2026. That's all for today's AI news briefing.