May 21, 2024

Drug discovery biotech LabGenius raises a £35m Series B

It’s robotic platform is using machine learning to create new molecules to tackle cancer

Mimi Billing

4 min read

UK drug discovery startup LabGenius, which uses machine learning to find and design antibody therapies to treat cancer, has raised a £35m Series B led by the VC arm of pharma company Merck, M Ventures, with participation from Octopus Ventures, LG Corp, Atomico, Kindred Capital, Lux Capital and Obvious Ventures. The investment brings LabGenius’s total funding to date to £58m.

The company will use the capital raised to make its machine learning-driven discovery platform more comprehensive, to facilitate more strategic partnerships and invest in its drug development for solid tumours.

From conservative testing to machine learning free from human bias

LabGenius was founded in 2012 and has developed a smart robotic platform capable of designing, conducting and, critically, learning from its experiments in drug discovery. The drug discovery startup is focused on creating antibody therapies — which trigger an immune response in the body to target cancer cells — to treat solid tumours.


Healthy cells surrounding a tumour have the same surface markers as it, which means that when using conventional treatment methods to try to destroy cancer cells on-target, off-tumour killing of healthy cells occurs. LabGenius says its platform uses an active learning method to create molecules that limit the damage to healthy cells.

The development of antibodies has become more complex over time – going from antibodies only binding to a target cell to the latest antibody therapeutics which combine multiple different binding domains into one molecule to create these complex multi-specific antibodies, according to CEO and founder James Field.

Picture of drug discovery startup LabGenius' founder and CEO James Field.
LabGenius' founder and CEO James Field.

GenAI has often been in the limelight when it comes to drug development and drug discovery, but for LabGenius the technology is only useful in the very early stages of the process.

“This is where I think some aspects of GenAI are really overhyped,” says Field. “Gen AI is typically used to describe the methodologies where you're taking vast amounts of public data and using that to create novel antibody designs. That's fine if the problem that you're looking to address is something like, ‘Does my antibody bind to a target?’ But if you want to address some of these more complex questions, you can't use those GenAI methods because the data just doesn't exist,” Field says.

With the help of its robotic and machine learning platform, Field says LabGenius can create a greater number of molecules for a specific task than conventional methods, and the automated discovery engine enables the rapid identification of high-performing antibodies with non-intuitive designs.

“If I can only test a few dozen of these molecules in these complex assays, I'll be very conservative,” says Field. “We've been able to take that whole experimental testing process and really scale. We combine that with machine learning and it massively opens up the design space and tests designs that you would never think of testing using conventional methods.”

Two legs to stand on

LabGenius has split the company into two focus areas: one on the pipeline described above for finding treatments for cancer, and one to co-develop new drugs with large pharmaceutical companies for revenue.

A couple of years ago, LabGenius initiated a multi-year collaboration with French pharma company Sanofi, in the area of inflammation.

“As a startup, the cash is one element [behind these collaborations], but there are also some other key pieces, like being able to benchmark how well your platform is performing and access areas and operating spaces that you may not conventionally be able to do,” Field says.

The 55-strong team at LabGenius is investing in extending its platform capabilities to facilitate more and broader strategic partnerships.


It is important for drug discovery startups to own their lead assets for as long as possible, according to Field.

“What's important is that, at least for the next few years, we maintain control of the lead assets. Ultimately, these machine learning-driven discovery platforms are only any good if you can create molecules that can be progressed,” he says.

“There's always the anxiety for any company, that if you partner up on your lead molecule too early, it may be the case that a pharma company just decides to pop it on the shelf for strategic reasons. And we've seen great molecules being shelved.”

Mimi Billing

Mimi Billing is Sifted's Europe editor. She covers the Nordics and healthtech, and can be found on X and LinkedIn