Here is today's AI news. Story 1. Mistral AI. Mistral reportedly raised €3 billion in a Samsung-led Series D, valuing the French AI lab at more than €21 billion—the largest equity raise reported for a European tech company.. Mistral AI raised €3 billion in a Series D led by Samsung Electronics, the EU-backed Scaleup Europe Fund and PSG Equity, valuing the French lab above €21 billion—nearly double its €11.7 billion valuation in 2025. Mistral called it Europe’s largest private-tech equity raise. The funding will expand AI-model research and computing infrastructure, reinforcing Europe’s push for sovereign AI. Story 2. Meta. Meta debuted Muse, a consumer personal AI agent designed to handle tasks such as email, travel bookings, bill negotiation, and purchases, subject to user approval.. Meta launched Muse, a U.S.-only personal AI agent available through iOS, Android, the web and WhatsApp. Powered by its Muse Spark models, it can open a browser, fill forms, send emails, book travel, negotiate bills and shop, pausing for approval before sensitive actions. Muse runs in a dedicated cloud virtual machine and uses Stripe Link’s disposable card numbers. Its launch intensifies competition to make AI agents autonomous—but raises privacy and trust concerns as they access users’ apps and data. Story 3. Microsoft. Microsoft introduced MAI-Transcribe-2, a speech-recognition model offered at $0.10 per audio hour through year-end.. Microsoft launched MAI-Transcribe-2 on Sept. 3 as a public-preview speech-to-text model in Azure Speech and Microsoft Foundry. It supports 60 languages, speaker diarization, word-level timestamps, automatic language detection and code-switching. Microsoft says it records a 5.2% average word-error rate on FLEURS and processes audio up to 10 times faster than OpenAI’s GPT-Transcribe. At $0.10 per audio hour through Dec. 31, 2026—about 72% below its predecessor’s $0.36 rate—the model could sharply reduce transcription costs, though Microsoft has not announced post-promotion pricing. Story 4. OpenAI. OpenAI launched faster GPT Image 2.5 variants, Flare and Sunburst, with reporting describing the image-generation release as a 50% speed improvement.. OpenAI launched GPT-Image-2.5 with two API variants: Flare, the default for fast, high-quality generation, and Sunburst, which prioritizes precision for detailed edits. OpenAI says Flare delivers higher-quality images than GPT-Image-2 at up to 50% lower latency. The update improves natural lighting, textures, reference-photo fidelity and multi-turn editing, potentially accelerating creative workflows, rapid prototyping and high-volume visual production. Story 5. Anthropic. Anthropic reportedly signed compute agreements totaling as much as $517 billion, securing large-scale capacity across AWS, Google, Microsoft Azure, SpaceX, and specialized cloud providers.. Anthropic has reportedly lined up up to $517 billion in compute, cloud and data-center agreements over the next decade, according to *The Information*. The deals cover at least 14.8 gigawatts, including roughly 11 GW from AWS and Google, plus Microsoft Azure, SpaceX, Lambda, Nscale and Fluidstack. The figure is a potential ceiling—not money already spent—highlighting the enormous infrastructure race behind frontier AI. That's all for today's AI news briefing.