Mistral AI provides open and portable generative AI solutions, offering customizable and deployable models like Mistral Nemo and Mistral Large for developers and businesses. The company emphasizes openness, portability, and efficient AI applications, supporting a wide range of industries with serverless APIs and compatibility with major cloud services.
Mistral AI released the weights for Mistral Large 4, with access through its API, Hugging Face, OpenRouter and NVIDIA. API pricing starts at $0.70 per million input tokens.
NVIDIA cloud partners are expanding computing capacity across Europe and nearby regions while supporting a new architecture for controlling AI agents. BBVA plans to evaluate the architecture for agent security and governance.
NVIDIA plans 13 additional minority stakes in European technology companies as part of $2 billion in new regional investments.
European tech companies raised €27.8 billion across 732 deals in the third quarter of 2026, with AI and cloud companies securing nearly half of the total.
Mistral AI and NVIDIA are developing AI workflows for engineering simulation, design, 3D scene generation, and automated tool use.
Voxtral 3 combines transcription, speech generation, reasoning and tool use for voice agents across more than 50 languages.
Crédit Agricole and Mistral AI will develop a conversational banking assistant, AI agents for internal processes, and cybersecurity tools.
Deutsche Telekom is using AudioCodes VoiceAI Connect to record, transcribe and summarize business calls through its Voice AI Notes Mobile service.
Mistral AI has released Timestral 1.0 and a smaller 123M variant in public preview for zero shot time series forecasting.
CERN and Mistral AI will develop a foundation model trained on billions of simulated particle collisions from the Large Hadron Collider.
Mistral AI has acquired design studio New Material and will create a Solutions Design team focused on customer deployments.
Orbitworks and Loft Orbital launched the first satellite in a planned 50 spacecraft AI constellation. The satellite uses NVIDIA hardware to process imagery and run Mistral models in orbit.
Ecosia is replacing Mistral AI with open weight models hosted through Melious, including Qwen, GLM and Kimi. The search engine says the change has cut operating costs in half while improving performance.
Mistral AI released Mistral Large 4, a multimodal model with 1 trillion parameters. The company plans to publish its model weights after three weeks of safety testing.
Mistral is opening a Munich hub focused on industrial AI and physical simulations, with BMW, Siemens Energy and the Technical University of Munich as initial partners.
ClosedQuorum consults Gemini, DeepSeek, Qwen, and Mistral models to choose actions after compromising Windows systems.
Mistral AI and European officials argue that coordinated limits on model development could protect dominant US companies while Europe works to close its technology gap.
TotalEnergies and Mistral AI will invest more than €100 million over three years to develop AI models for oil and gas exploration and reservoir engineering.
Mozilla is using Mistral models for the beta version of Firefox Smart Window in France and North America. Support in the United Kingdom and Germany will follow later in 2026.
Mistral AI will integrate its models with Cloudera's hybrid data platform, supporting training and inference across private clouds, public clouds, local infrastructure, and isolated environments.
Daily report covering major AI developments and industry news, with both top stories and complete market updates