TotalEnergies and Mistral announced on September 15, 2026, a three-year program worth more than €100 million to develop frontier artificial-intelligence models for oil and gas exploration, reservoir characterization, field development and engineering. The companies also plan to create a joint scientific laboratory for the work.
The project is aimed at helping TotalEnergies’ specialists interpret complex subsurface information and evaluate possible development paths. It is a major industrial AI initiative—but it is being presented as a development program, not as a finished product that customers can buy.
What TotalEnergies and Mistral plan to build
The planned models are intended to integrate, process and interpret large volumes of subsurface data. In practical terms, that could help geoscientists and reservoir engineers examine information about underground formations and compare different ways to develop an exploration opportunity.
The companies also describe systems that could generate multiple development scenarios, support engineering decisions and help optimize existing oil and gas projects. The stated scope includes the possibility of extending the productive lives of those projects.
| Dimension | Announced detail |
| Program | Development of frontier AI models for exploration and reservoir work |
| Duration | Three years |
| Investment | More than €100 million |
| Intended data work | Integrating, processing and interpreting subsurface information |
| Planned output | Multiple scenarios for developing exploration opportunities |
| Industrial use | Reservoir engineering, project optimization and possible life extension |
| Research structure | A joint scientific laboratory between TotalEnergies and Mistral |
The models are associated with agentic AI, a term generally used for systems that can coordinate several steps toward a task. Here, the announced applications focus on analysis, scenario generation and decision support. TotalEnergies has not described the project as an autonomous system for operating oil and gas facilities.
A laboratory joining geoscience and AI
The planned laboratory will bring together TotalEnergies’ subsurface specialists and Mistral’s researchers and engineers. TotalEnergies says the project will draw on nearly a century of geoscience knowledge and data, while Mistral will contribute its scientific and technological capabilities.
That combination matters because reservoir work is not simply a matter of feeding documents into a general-purpose chatbot. Subsurface data can include complex geological information, historical records and engineering constraints. The proposed models are meant to work within that specialized industrial context and help experts compare possible decisions.
TotalEnergies chairman and CEO Patrick Pouyanné said the company wants to combine its geoscience data and team expertise with Mistral’s capabilities to support analysis and decision-making. Mistral co-founder and CEO Arthur Mensch presented the partnership as an example of customizable AI for demanding industrial processes, with an emphasis on protecting enterprise intellectual property.
The companies also frame the project as part of a European digital ecosystem. That positioning reflects Mistral’s role as a French AI company and TotalEnergies’ interest in applying customizable systems to strategic business data.
What the program means for oil and gas operations
The intended use is closer to an advanced decision-support layer than to a machine taking over a drilling platform. The models could help teams organize large datasets, explore alternative development scenarios and assess ways to optimize existing projects.
That distinction is important. A system that helps experts analyze reservoirs and compare scenarios has a different role from one that directly controls pumps, valves or other field equipment. The announced applications concern information, modeling and engineering decisions.
The program also spans both new exploration opportunities and existing projects. Its scope therefore includes finding and characterizing potential resources, planning development and examining whether established projects could operate productively for longer. The announcement does not turn those intended uses into measured production gains or a demonstrated commercial deployment.
A second step in the TotalEnergies–Mistral relationship
The reservoir-focused program follows an earlier TotalEnergies and Mistral AI collaboration publicized on June 12, 2025, which covered broader applications of artificial intelligence across TotalEnergies’ activities.
The September 2026 announcement gives that relationship a more specific industrial target: frontier models built around subsurface data, exploration and reservoir engineering. The initial program is scheduled to run for three years, with the joint laboratory serving as its research structure.
