Remanence raises $30M seed to help firms own their AI models
What's the deal? Paris-based RemanenceDealroom has a profile for this one. Try Dealroom → has raised $30 million in seed funding to build a platform that turns companies' AI-agent activity into specialised models they own. EQT VenturesDealroom has a profile for this one. Try Dealroom → and FirstMark co-led the round, with Zero PrimeDealroom has a profile for this one. Try Dealroom →, EmblemDealroom has a profile for this one. Try Dealroom →, Drysdale, Global Founders Capital, and Yellow participating.
Angel backers include Olivier PomelDealroom has a profile for this one. Try Dealroom → (Datadog), Philippe CorrotDealroom has a profile for this one. Try Dealroom → (Mirakl), Romain NiccoliDealroom has a profile for this one. Try Dealroom → (Pigment), Victor Riparbelli and Steffen Tjerrild (Synthesia), and Mehdi Ghissassi (AI71Dealroom has a profile for this one. Try Dealroom →).
What's the endgame? Remanence plugs into a company's existing AI models and turns their execution traces into tailored training environments, then post-trains open-weight models through reinforcement learning. As new traces accumulate, models can be retrained to improve accuracy and cut latency and cost. Companies keep the model weights, and data can stay inside their own infrastructure.
The product: The platform spans four steps — trace analysis, environment synthesis, long-horizon reinforcement learning, and inference, run fully managed or in a client's own cloud. Remanence is already working with design partners in legal, customer support, and enterprise software across Europe and the US.
Who's behind it? Remanence was founded by Pierre-Louis Cedoz, Hamza Benchekroun, and Axel Moyal, who previously worked together at H Company, France's computer-use AI lab. There they led the post-training of the open-weight Holo models. The team also brings experience from Google DeepMind, Meta, and InstaDeep.
Why now? AI spend keeps rising as businesses deploy general-purpose agents on more complex tasks, pushing up token usage. A November Reuters report found 59% of surveyed large-company executives plan to increase their use of cheaper open-weight models as costs exceed budgets. "AI companies spend more on intelligence every quarter and own none of it," Cedoz said.
What's the money for? The funding will support research and compute infrastructure, including a dedicated GPU cluster, plus work on Remanence's reinforcement-learning environments, evaluation systems, and post-training. The company will hire primarily in Paris and establish an early presence in London and San Francisco.
The signal: Remanence is betting that enterprises will pay to improve models on their own workflows while keeping control of data and weights. The space is drawing comparable investment — PluraiDealroom has a profile for this one. Try Dealroom → recently announced $13 million in seed funding for tools that simulate agent workflows and post-train task-specific models. The outcome will hinge on whether the "own, don't rent" approach delivers reliable results and real savings in production.
Read more: remanencelabs.ai
Image credit: Remanence