TotalEnergies, Mistral Invest US$108 Million in Frontier AI

TotalEnergies and Mistral have launched a US$108 million joint program to build agentic AI models for subsurface geoscience, reflecting a broader shift as energy majors deploy high-performance computing, predictive maintenance, and digital twins across upstream and downstream operations. This transition toward proprietary digital infrastructure allows global operators to optimize capital deployment, accelerate reservoir scenario planning, and capture structural cost savings while securing critical technical data. For energy stakeholders and service providers, integrating real-time operational analytics is becoming essential to strengthen infrastructure reliability and reduce production costs.

___

France-based energy major TotalEnergies and France-based artificial intelligence (AI) developer Mistral have launched a three-year joint research program representing an investment exceeding €100 million (US$108 million) to develop frontier AI models for subsurface geosciences. 

The collaboration aims to assist TotalEnergies' experts in the exploration, characterization, and development of oil and gas reservoirs by integrating advanced AI capabilities into core industrial workflows.

According to TotalEnergies, the partnership focuses on leveraging agentic AI frameworks to process nearly 10 petaflops of subsurface data alongside TotalEnergies' accumulated domain expertise in reservoir engineering. The resulting models are designed to integrate, analyze, and interpret large volumes of geological data to generate development scenarios for both new exploration opportunities and the life-extension of existing producing fields.

As part of the agreement, the two French companies will establish a joint scientific laboratory. The facility will combine TotalEnergies' subsurface engineering teams with Mistral's software architecture to build customized models tailored to complex subsurface technical requirements. The initiative also aims to preserve proprietary data security within a European digital infrastructure framework.

"Exploration and reservoir engineering are among the areas where artificial intelligence can create the greatest value for our activities. By combining a century of geoscience data, the expertise of our teams and Mistral's capabilities, we aim to develop a new generation of tools capable of supporting our experts in analyzing the most complex data and in their decision-making," stated Patrick Pouyanné, Chair and CEO, TotalEnergies, addin that this partnership illustrates the company’s commitment to contributing to the development of a high-performing European digital ecosystem.

The program marks an expansion in the deployment of generative and agentic AI models across heavy industry applications, focusing on data-intensive scientific processes rather than general commercial administration. "We are delighted to take our collaboration with TotalEnergies to the next level. This project demonstrates the ability of our models to support complex industrial processes and adapt to the most demanding scientific and energy challenges. It also highlights the strength of Europe's industrial ecosystem and the importance for large enterprises of adopting state-of-the-art, customizable AI solutions that protect and respect their intellectual property," stated Arthur Mensch, Co-founder and CEO of Mistral.

How Supermajors Are Implementing AI in Operations

The deployment of AI across the upstream and downstream energy sectors extends beyond subsurface modelling, as global operators integrate machine learning, high-performance computing, and digital twin technology across their asset portfolios. UK-based energy major BP embeds AI technologies throughout its US operations to optimize drilling trajectories, enhance refinery throughput, and improve operational safety. High-performance computing facilities in Houston allow BP to reduce the evaluation time for optimal drilling paths in the Gulf of Mexico and the Permian Basin from months to days, utilizing geological data to identify high-margin recovery trajectories.

Operators also pair AI with cloud-based virtual models of offshore facilities, commonly referred to as digital twins, to enable remote maintenance planning and system monitoring. At its US onshore division, BPX Energy, machine-learning applications generate automated operational reports to direct frontline activities and manage production costs. Furthermore, predictive machine-learning tools analyze real-time data to anticipate equipment failures on offshore production platforms, while downstream facilities deploy data analytics software to accelerate pre-job safety evaluations and optimize maintenance budgets.

Energy Companies Can Save over US$225 Billion With AI

Research conducted by Honeywell Technologies in collaboration with the MIT Center for Sustainability Science and Strategy indicates that digital technologies and artificial intelligence could generate global annual savings of up to US$225 billion in production costs for traditional petroleum-based fuels. 

In an MBN Expert Contributor piece, Jose Simon, Vice President and General Manager, Process Automation, Honeywell, notes that beyond initial sub-surface exploration, operators deploy real-time digital solutions in industrial control rooms to anticipate equipment failures and production losses. Commercial applications co-developed with TotalEnergies and Chevron demonstrate that integrating historical and real-time operational data into machine-learning control assistants allows operators to predict unsafe conditions and alarm incidents 5 to 10 minutes in advance, reducing cognitive load on personnel and reinforcing infrastructure reliability.