Basecamp Research has raised $140 million to move its biological foundation models further into drug development, giving one of Europe’s better-funded AI-biotech companies the capital to test whether models trained on proprietary biological data can produce medicines that survive the much harder journey towards the clinic.

Basecamp Research, headquartered in London with laboratories in Cambridge, Massachusetts, said the oversubscribed Series C was led by S32. Other investors include NVIDIA, Anthropic’s Anthology Fund, the NATO Innovation Fund, Redalpine, Singular, the European Tech Collective, The Rockefeller Foundation, and several specialist life-sciences investors.

The company will use the financing to train a new generation of its EDEN biological foundation models, expand pharmaceutical partnerships, and advance an internal therapeutics pipeline beginning with in vivo cell therapy. Richard Pearce, previously at Biogen, has also joined as chief business officer as Basecamp develops a commercial operation around technology previously demonstrated mainly through models, datasets, partnerships, and preclinical research.

EDEN is trained on Basecamp’s proprietary biological information and is intended to work across several design problems rather than one narrow scientific task. The company’s Trillion Gene Atlas draws on biological-data partnerships across more than 30 countries, creating a training resource that Basecamp argues exposes its models to a broader range of evolutionary information than public genomic datasets alone.

The company’s strategy places biological data alongside model architecture as a core asset. Public genomic repositories contain enormous quantities of useful information, but they are unevenly distributed across organisms, geographies, and experiments, while repeated or closely related sequences can make a dataset larger without necessarily adding equivalent biological diversity.

Basecamp is attempting to use a broader evolutionary training set to design therapeutic candidates directly from information about a disease. Its stated model capabilities include cell and gene therapies, enzymes, and peptides, while the new financing shifts the focus from demonstrating those capabilities towards building medicines around them.

The financing changes the evidential threshold

A foundation-model benchmark can establish that a system is scientifically interesting, but a therapeutics company ultimately has to show that a candidate can be manufactured, delivered into patients, tolerated at a useful dose, and produce a clinically meaningful effect. The Series C therefore changes the kind of evidence Basecamp will increasingly be expected to provide.

Its first internal focus is in vivo cell therapy, where the company wants to reprogramme a patient’s cells inside the body rather than removing them, modifying them externally, and returning them through a specialist manufacturing process. Existing cell therapies can be clinically powerful, although some approaches involve complex individualised manufacturing, specialist treatment centres, and substantial cost.

Basecamp says EDEN can design long and complex DNA sequences alongside large serine recombinases capable of integrating genetic material into the genome. The company believes pairing those elements could support more sophisticated cell therapies delivered directly inside the body.

That is still an ambition rather than a clinical result. Basecamp says it has demonstrated preclinical performance across several modalities and disease areas, but preclinical evidence does not establish that a treatment will be safe or effective in people.

Drug development introduces constraints that do not appear in a model evaluation. Toxicity, biodistribution, immune response, dosing, manufacturing consistency, stability, patient selection, trial design, and interactions with other treatments can all determine whether a scientifically plausible candidate becomes a medicine.

Biological data becomes infrastructure

The funding also illustrates an important difference between biological AI and many enterprise applications. The model may be valuable, but access to differentiated training data can become an equally significant competitive asset because biological information is expensive to collect, sequence, curate, and legally govern.

Basecamp’s Trillion Gene Atlas is being built with partners including NVIDIA, Anthropic, PacBio, and Ultima Genomics and draws on access and benefit-sharing partnerships across all seven continents. That creates a biological-data supply chain alongside the AI programme rather than a one-off training corpus assembled from public sources.

There is a commercial trade-off in that approach. Proprietary biological information can differentiate a model, although building and maintaining the underlying collection network requires field partnerships, sequencing, storage, curation, legal arrangements, compute, and continuing scientific validation.

The expense may be justified if greater diversity allows EDEN to produce therapeutics competitors cannot design from common datasets alone, but that advantage has to survive later stages of development. A better-generated candidate still enters the same world of manufacturing, regulation, trials, and clinical evidence as a molecule discovered through more conventional methods.

AI moves closer to biotechnology economics

The investor group also shows the convergence between large technology companies, strategic capital, and biotechnology. NVIDIA supplies computing technology relevant to Basecamp’s work, Anthropic is involved through both investment and scientific collaboration, and the NATO Innovation Fund brings a strategic European investor into the round alongside specialist life-sciences capital.

Those relationships can give Basecamp access to compute, models, distribution, and commercial networks beyond those of a conventional early-stage biotechnology business. They also increase the expectation that the technology moves beyond a research platform and into programmes capable of producing economic value through partnerships or proprietary medicines.

The company says the financing will advance its pipeline towards clinical development, but it has not announced a human trial beginning as part of the Series C. That distinction should remain clear because “towards clinical development” covers substantial work before a patient is dosed.

The next useful milestones will therefore be specific rather than architectural: nomination of development candidates, manufacturing progress, regulatory filings, trial approvals, and eventually human safety and efficacy data.

Basecamp’s latest round gives it the money to pursue those thresholds at a much larger scale. EDEN may help the company search biological design space and generate candidates more quickly, but the clinic remains deliberately resistant to shortcuts. A medicine designed by AI still has to behave safely and effectively inside a human body.