Mistral AI said on August 11, 2026 that it is adding regional inference controls, third-party open-model support, and a European compute-capacity program to support what it calls sovereign AI. The company said enterprises and countries need control over which models they use, where intelligence runs, how compute capacity scales, and how the value created by AI is retained.

Mistral Regional Endpoints are now generally available and let customers choose whether inference runs in Europe or the US. The company said inference and associated processing take place in the selected region, subject to limited, safeguarded transfers to sub-processors outside that region as described in its Trust Center. Mistral also introduced Mistral Priority Tier in public preview, offering committed service levels for mission-critical workloads, custom rate limits, and an uptime SLA. Mistral said it is the only European AI lab offering both a choice of processing region and a committed, SLA-backed service level.

Mistral said many customers already run its models inside their own data centers and cloud environments, but production AI workloads require capacity that remains available, resilient, and aligned with regional controls. The company framed the inference layer around two requirements: keeping data and processing in-region, and ensuring dependable access to capacity when demand peaks. Mistral said its models will also remain available through partners.

Mistral said AI sovereignty also requires control over which intelligence runs on infrastructure, because current AI systems often combine multiple capabilities such as extended reasoning, high-volume production, specialist tools, and custom models trained on company knowledge. The company pointed to its frontier models, specialist products including Mistral OCR and Voxtral, and custom models. Mistral said open weights let customers see inside a model, adapt it, and retain the intelligence they build with it.

Mistral said its platform will support third-party open models, starting with Z.ai’s GLM-5.2. The company said GLM-5.2 and future open models will run on the same infrastructure, regional controls, and service commitments as Mistral models, allowing customers to broaden model choice without fragmenting where AI runs. Matan Grinberg, CEO and cofounder of Factory, said Mistral lets Factory run open models under strict regional controls and service commitments while maintaining data-residency and compliance requirements.

Mistral also announced a coalition for long-term AI capacity in Europe. The company said it is bringing together an anchor group of enterprises whose multi-year commitments can support European infrastructure at a scale no participant could secure alone. European Compute Units, or ECUs, will convert those commitments into access to Mistral-built infrastructure over multiple years, with capacity usable across products available on Mistral Compute as customer needs change.

Executives from Amadeus, ASML, Capgemini, Caisse des Dépôts, and CMA CGM backed the program in statements included by Mistral. Luis Maroto of Amadeus cited capacity, deployment control, and operating continuity for enterprises. Christophe Fouquet of ASML described European AI capacity as important to Europe’s next generation. Aiman Ezzat of Capgemini said the program gives clients compute, partnerships, and confidence for critical workloads. Olivier Sichel of Caisse des Dépôts said Mistral Compute offers a European neocloud and referred to the IRN, or Digital Resilience Index. Rodolphe Saadé of CMA CGM said Mistral deployment is already under way among thousands of employees and across geographies, including customer care.

Mistral said Europe is the starting point for a framework based on operational control, choice of intelligence, and assured access to compute. The company said the same challenge exists elsewhere, as organizations and governments seek frontier AI for mission-critical needs without giving up control over infrastructure and the intelligence loop.

Mistral AI said on August 11, 2026 that it is adding regional inference controls, third-party open-model support, and a European compute-capacity program to support what it calls sovereign AI. The company said enterprises and countries need control over which models they use, where intelligence runs, how compute capacity scales, and how the value created by AI is retained.

Mistral Regional Endpoints are now generally available and let customers choose whether inference runs in Europe or the US. The company said inference and associated processing take place in the selected region, subject to limited, safeguarded transfers to sub-processors outside that region as described in its Trust Center. Mistral also introduced Mistral Priority Tier in public preview, offering committed service levels for mission-critical workloads, custom rate limits, and an uptime SLA. Mistral said it is the only European AI lab offering both a choice of processing region and a committed, SLA-backed service level.

Mistral said many customers already run its models inside their own data centers and cloud environments, but production AI workloads require capacity that remains available, resilient, and aligned with regional controls. The company framed the inference layer around two requirements: keeping data and processing in-region, and ensuring dependable access to capacity when demand peaks. Mistral said its models will also remain available through partners.

Mistral said AI sovereignty also requires control over which intelligence runs on infrastructure, because current AI systems often combine multiple capabilities such as extended reasoning, high-volume production, specialist tools, and custom models trained on company knowledge. The company pointed to its frontier models, specialist products including Mistral OCR and Voxtral, and custom models. Mistral said open weights let customers see inside a model, adapt it, and retain the intelligence they build with it.

Mistral said its platform will support third-party open models, starting with Z.ai’s GLM-5.2. The company said GLM-5.2 and future open models will run on the same infrastructure, regional controls, and service commitments as Mistral models, allowing customers to broaden model choice without fragmenting where AI runs. Matan Grinberg, CEO and cofounder of Factory, said Mistral lets Factory run open models under strict regional controls and service commitments while maintaining data-residency and compliance requirements.

Mistral also announced a coalition for long-term AI capacity in Europe. The company said it is bringing together an anchor group of enterprises whose multi-year commitments can support European infrastructure at a scale no participant could secure alone. European Compute Units, or ECUs, will convert those commitments into access to Mistral-built infrastructure over multiple years, with capacity usable across products available on Mistral Compute as customer needs change.

Executives from Amadeus, ASML, Capgemini, Caisse des Dépôts, and CMA CGM backed the program in statements included by Mistral. Luis Maroto of Amadeus cited capacity, deployment control, and operating continuity for enterprises. Christophe Fouquet of ASML described European AI capacity as important to Europe’s next generation. Aiman Ezzat of Capgemini said the program gives clients compute, partnerships, and confidence for critical workloads. Olivier Sichel of Caisse des Dépôts said Mistral Compute offers a European neocloud and referred to the IRN, or Digital Resilience Index. Rodolphe Saadé of CMA CGM said Mistral deployment is already under way among thousands of employees and across geographies, including customer care.

Mistral said Europe is the starting point for a framework based on operational control, choice of intelligence, and assured access to compute. The company said the same challenge exists elsewhere, as organizations and governments seek frontier AI for mission-critical needs without giving up control over infrastructure and the intelligence loop.