Mistral is putting regional control and guaranteed capacity at the centre of its enterprise AI offer. Its new Regional Endpoints are generally available for Europe and the US, while a Priority Tier enters public preview with committed service levels, custom rate limits and an uptime SLA. Mistral is also expanding access to third-party open models and organising European buyers around long-term compute commitments, with a stated ambition to reach up to 1 GW of capacity by 2030. The practical shift is from choosing a model to specifying where inference runs, how reliably it runs and who controls the infrastructure. Read Mistral’s infrastructure announcement.
What else matters
NVIDIA targets the model-routing layer for always-on agents
NVIDIA launched Nemotron 3.5 Lightning, a 30-billion-parameter mixture-of-experts model for specialised, high-volume agent tasks, alongside NeMo Switchyard, an open-source router for directing requests across open, proprietary and NVIDIA models. The combination treats agent systems as model ensembles rather than one-model applications. Read NVIDIA’s announcement.
GitHub adds a cheaper vision-capable coding model
MAI-Code-1.1-Flash is rolling out across GitHub Copilot with native vision support, better instruction following and tool use, and a 73% lower list price than its predecessor. It is available for manual selection on paid plans and through auto-selection for Free and Student users; Business and Enterprise administrators must enable its policy. Read the GitHub changelog.
Europe’s AI Act enforcement moves into its operating phase
The European Commission says the AI Office and national authorities began enforcing the AI Act’s rules and new transparency requirements from 2 August. For providers and deployers, the immediate issue is operational: documentation, governance and disclosure controls now need to work in production, not just appear in compliance plans. Read the Commission’s enforcement notice.
Google’s device strategy makes Gemini a product layer
At its Made by Google 2026 event, Google showed the Pixel 11 range, Pixel Watch 5 and new Gemini-powered device features. The significance is distribution: Google is continuing to move model capabilities from standalone chat surfaces into the operating system and hardware people already use. Read the event coverage.
