AI News from September 8, 2026

OpenAI says GPT-5.6 Sol and Codex autonomously calibrated MIT qubits, cutting routine lab work while exposing AI agents’ limits with noisy data.

HPE Zerto is deploying an on-premises AI troubleshooting system with Amazon Bedrock, linking live recovery data to guarded, actionable guidance.

Mistral AI has raised €3 billion at a valuation above €21 billion, funding European compute, infrastructure and a sovereignty-focused global strategy.

NVIDIA says Vera Rubin NVL72 can deliver up to 30x more agentic inference throughput per megawatt than GB300, pending independent benchmark review.

NVIDIA’s Jetson deployment guide shows how quantization and speculative decoding can bring compact reasoning models to local edge AI workloads.

Reports say Samsung is leading new Mistral funding, but conflicting figures and missing primary documents leave the round’s size and terms unverified.

Danijar Hafner is developing world-model agents for robots that can predict outcomes and adapt to unfamiliar homes, objects, and environments.

Samsung is reported to lead a €3 billion Mistral AI round, lifting the French model maker’s valuation above $24 billion and sharpening Europe’s AI race.

Baidu’s inclusion in Shanghai and Shenzhen Stock Connect broadens mainland investor access, putting its AI strategy under closer market scrutiny.

OpenAI says GPT-6 Astra is its first broadly deployed model to reach Critical cybersecurity capability under its Preparedness Framework, raising safety stakes.