Remanence Aims to Make Post-Training Accessible

Remanence aims to convert corporate AI agents' data into specialized models. Its CEO, Pierre-Louis Cedoz, aspires to provide them with proprietary intelligence that evolves with use, potentially creating a sustainable advantage.
Remanence anticipates increased computing power
For Pierre-Louis Cedoz, co-founder and CEO of Remanence, post-training has a power he claims to have directly observed, yet it remains largely inaccessible to most companies. He presents his company as a means to overcome this barrier by giving them control over specialized models built with their agents' data.
The company has just exited its confidential phase, or "stealth mode." It plans to dedicate its initial funding to developing its computing capabilities as well as recruiting to refine its learning environments. It also intends to establish an office in London and a branch in San Francisco, with the stated ambition of being at the forefront of the next advancements in AI.
Le brief IA que les pros lisent chaque soir
Les 7 actus IA du jour, décryptées en 5 min. Gratuit.
Inclus dès l'inscription : notre sélection des meilleurs guides & comparatifs IA.
Choisissez votre rythme
Gratuit · Pas de spam · Désabonnement en 1 clic
Agents' traces fuel training
The solution connects to the models already used by the company: it converts their execution traces into customized training environments. Additional traces can then feed into new training sessions. These could enhance accuracy, track changes in work processes, and reduce costs such as latency.
The founders aim to customize open-weight models for the specific needs of each company. Their goal is to enable them to perform complex tasks without the increase in the number of tokens consumed causing a spike in AI costs.
EQT Ventures and FirstMark lead the funding
The initial funding amounts to $30 million, led by EQT Ventures and FirstMark. Zero Prime, Emblem, Drysdale, Global Founders Capital, and Yellow also contributed. Among the individual investors are Philippe Corrot (Mirakl), Olivier Pomel (Datadog), Rimain Niccoli (Pigment), and Mehdi Ghissassi (AI71).
At the origin of the project, Pierre-Louis Cedoz, former CTO of H Company who moved to Google DeepMind, teamed up with Hamza Benchekroun, previously at InstaDeep, and Axel Moyal, formerly at Upstart. All three worked at H Company, a Parisian startup with a challenging start, now led by Gautier Cloix.
Brief IA — L'actualité IA en français
L'essentiel de l'actualité de l'intelligence artificielle, décrypté et expliqué chaque jour.