The Pulse
TotalEnergies and Mistral Put More Than €100M Into Reservoir AI
TotalEnergies and Mistral are launching a three-year, more than €100 million program to build AI models for oil and gas reservoir exploration and engineering. The joint laboratory will combine TotalEnergies’ geoscience data with Mistral’s m
AI.info Team ·
More than €100 million will fund a three-year effort by TotalEnergies and Mistral to build artificial intelligence models for oil and gas reservoir exploration, characterization and engineering. The companies announced the program in Paris on September 15, 2026, framing it as a move from broad enterprise AI experimentation toward models built around one of the energy industry’s most data-intensive technical disciplines.
The partnership will create a joint scientific laboratory. TotalEnergies will contribute subsurface expertise and nearly a century of accumulated geoscience and reservoir-engineering knowledge, while Mistral will provide model-development and AI capabilities. The companies say the work will focus on proprietary tools designed for TotalEnergies’ operations rather than a general-purpose assistant.
A €100 Million Program for Reservoir Decisions
TotalEnergies says the models will integrate, process and interpret large volumes of subsurface information. Their intended output is not a single prediction, but multiple development scenarios for exploration prospects and existing projects.
Those scenarios could support decisions about how to characterize a reservoir, how to develop an opportunity and how to extend the productive life of an existing asset. The announcement does not provide performance benchmarks, target fields or a timetable for putting the resulting models into production.
The company describes the data foundation as nearly a century of geoscience experience combined with what it calls “next to 10 petaflops of data.” The wording appears in TotalEnergies’ announcement; the release does not further define the measurement or identify the specific datasets included.
Agentic AI Meets Subsurface Engineering
TotalEnergies and Mistral say the program will draw particularly on agentic AI. In the companies’ description, the technology will help work across large and varied datasets rather than simply generate text or summarize documents.
Reservoir engineering combines geological interpretation, production history, subsurface modeling and development planning. A useful system would need to connect those forms of evidence while preserving the assumptions and constraints that experts use to assess an exploration opportunity or an existing field.
Patrick Pouyanné, Chairman and Chief Executive Officer of TotalEnergies, said exploration and reservoir engineering are areas where AI can create significant value for the company. “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,” he said.
A Joint Laboratory Builds on a 2025 Alliance
The announcement expands a relationship the companies disclosed in June 2025, when they agreed to establish an AI innovation lab covering TotalEnergies’ broader multi-energy strategy. That earlier work included potential applications in low-carbon energy, industrial performance and customer services.
The new program narrows the focus to exploration and production. TotalEnergies’ announcement says the laboratory will combine its subsurface teams with Mistral’s scientific and technical staff to develop solutions tailored to the company’s business needs.
Arthur Mensch, co-founder and CEO of Mistral, said the project would test whether the company’s models can support complex industrial processes and demanding scientific and energy work. He also emphasized customizable systems that protect the intellectual property of enterprise customers.
Data Control Is Part of the Deal
TotalEnergies presents the partnership as part of a European digital strategy. The company says it wants to create proprietary models within a European ecosystem, using its strategic subsurface data while helping Mistral expand its position as a European AI company.
That arrangement matters because geological and reservoir data can represent decades of exploration spending, technical interpretation and operational knowledge. The release does not specify where the models will run, which cloud or computing infrastructure will support them, or how access to the resulting systems will be divided between the two companies.
Mistral’s current energy offering lists reservoir and subsurface simulation among its target applications, alongside engineering-document analysis, predictive maintenance and energy-market workflows. The TotalEnergies project gives that sector strategy a named industrial partner and a defined technical domain.
The Measure of Success Will Be Technical
The announcement sets out an ambition, not a completed product. TotalEnergies and Mistral do not disclose the models’ architecture, training methods, benchmark results or the first reservoir projects where the system will be evaluated.
For the partnership to justify its scale, the laboratory will need to show that its models improve the quality or speed of expert analysis without weakening geological judgment, operational controls or ownership of sensitive data. The first concrete test will be whether the system can produce development scenarios that TotalEnergies’ geoscientists consider useful enough to influence actual reservoir decisions.