January 30, 2024

Gen AI and quantum to design new drugs: Paris-based Aqemia completes €60m Series A

The French biotech combines generative AI with quantum-inspired algorithms to “invent” new molecules that can target certain diseases

Paris-based biotech Aqemia, which uses AI to fast-track drug discovery and design, has raised a €30m extension to its Series A, bringing total funding for the round to an all-equity €60m. 

The startup raised a first €30m in 2022 led by French VC Eurazeo and public bank Bpifrance, with historical investor Elaia also participating. The latest extension, led by French growth investor Wendel Growth, saw all three of Aqemia’s previous investors return.

Aqemia has developed an AI-powered technology that it says can accelerate the discovery and design of molecules that have the potential to become drugs that fight different diseases — a complex and cost-intensive process that traditionally takes many years. 


The startup is building its own drug discovery pipeline, meaning that it also plans to take on the pre-clinical and clinical testing of the molecules that it identifies as drug candidates.

Currently, three drug candidates produced by Aqemia’s technology to target different types of cancer are undergoing preclinical tests in animals. The company says that it took less than two years for these projects to reach the pre-clinical phase, and that it expects the candidates to progress to clinical stages in 2025.

Using conventional means, the discovery and preclinical phases of drug discovery take an average five to six years, according to industry estimates

Combining GenAI and quantum-inspired algorithms

The discovery process in drug design consists of identifying a biological target — such as a protein, enzyme or DNA — that signals a disease and then finding out which molecules might be good candidates to have an effect on that target.

Whereas scientists typically have to resort to time-consuming trial and error to identify potential drug candidates, Aqemia leverages generative AI — the same technology that powers chatbots like ChatGPT — to suggest which molecules might work best.

The AI model is trained on the laws of physics that govern how molecules are built in nature. 

When prompted to create a molecule that can address a certain target, the technology “invents” millions of possible candidates, of which a very small percentage of the generated options are likely to show promise.

This is why the company has developed quantum-inspired algorithms, which excel at simulation problems, to test the millions of candidates produced by the model in a short period of time. Those that have the most potential are fed back to the AI model.

With every feedback loop, the model’s answers to the prompt become more efficient and targeted — until the candidates are considered efficient enough to be built and then tested. 

“It’s as if ChatGPT didn’t have any literature to train on, but 10m literature teachers in front of it who can correct and provide feedback to the AI, so that it can invent another 10m sentences,” says Aqemia cofounder Maximilien Levesque.

Combining quantum-inspired algorithms with AI means that Aqemia’s technology can rapidly trial a high number of drug candidates, while not requiring any training data that’s specific to any one illness or condition.


Trialling new drug candidates

In addition to developing the technology to improve the identification of drug candidates, Aqemia also intends to take these candidates through pre-clinical and clinical trials.

It currently has three candidates in pre-clinical trials and expects more to come thanks to the recent fundraise, in a range of therapeutics including immunology, inflammation and the central nervous system.

“Our technology has produced three seeds,” says Levesque. “The objective of the new fundraise is to sow more and create a large pipeline of proprietary drugs.”

The startup also anticipates that at least one candidate will enter clinical trials in 2025. Given that pre-clinical and clinical trials can cost tens of millions of euros with current methods, Aqemia’s technology will need to significantly bring the cost down if it hopes to take a good number of compounds to market with VC funding.

To support the expansion of its activity, it intends to recruit around 40 new employees — a significant increase on the company’s current 50-strong workforce. 

Towards AI-designed drugs

A small part of Aqemia’s activity also comes from partnerships with pharmaceutical companies.

The funding extension comes just two months after Aqemia signed a $140m partnership with Sanofi, in which the startup will be tasked with identifying molecules that can address diseases that are of interest to the pharmaceutical giant.

In 2021, the startup also announced that it was collaborating with French pharmaceutical company Servier to accelerate the discovery of drug candidates for an undisclosed target in immuno-oncology.

AI has sparked the interest of pharmaceutical companies in recent years thanks to its promise to make it cheaper and faster to find new drugs — but the technology is yet to have proven itself. 

Late last year, US-based Insilico Medicine progressed an AI-designed drug candidate to the clinical phase, which was described as the industry’s first fully AI-based preclinical candidate. Results for the drug, which targets an incurable lung disease called idiopathic pulmonary fibrosis, are expected early 2025.

Levesque is realistic about the challenges ahead. “Right now, we can rapidly generate drug candidates — but the clinical trial phase is a completely different game,” he says. 

“The next step will be accelerating and increasing success rates in clinical trials. That’s the next big challenge.”

Daphné Leprince-Ringuet

Daphné Leprince-Ringuet is a reporter for Sifted based in Paris and covering French tech. You can find her on X and LinkedIn