Basecamp Research Raises $140M Series C to Push AI-Designed In Vivo Cell Therapies Toward the Clinic
核心洞察
Basecamp Research (搜索) closed a $140 million Series C led by S32 (搜索), with Nvidia (搜索)'s NVentures and Anthropic (搜索)'s Anthology Fund among the investors.
The London-based company will use the proceeds to train new generations of its EDEN (搜索) biological foundation models and advance AI-designed therapeutics.
In a preprint, EDEN (搜索)-designed DNA-inserting proteins achieved a 63.2% functional hit rate, and half of 20 proteins tested inserted a CD19 (搜索) CAR into human T cells.
Basecamp Research (搜索) has raised $140 million in a Series C round to train a new generation of its biological foundation AI models and advance a pipeline of AI-designed therapeutics, beginning with in vivo cell therapy. The London-based company, founded in 2019, said the round follows strong preclinical results across multiple disease areas and brings its total raised to $225 million.
The round was led by S32 (搜索), the California venture firm founded by Bill Maris. Nvidia (搜索)'s venture arm NVentures returned as an investor, while Anthropic (搜索)'s Anthology Fund, a partnership between Anthropic and US venture firm Menlo Ventures, joined as a new investor. Other new backers include Catalio Capital Management, the NATO Innovation Fund, The Rockefeller Foundation, European Tech Collective, Inception Fund, PostScriptum, Singular, True Ventures and the UK government's Sovereign AI Fund. Returning investors include Firebrand, River Capital, King Philanthropies and Redalpine. André Hoffmann, Roche's vice chairman, was among additional investors from the biopharma and other industries.
Reprogramming Cells Inside the Body
Basecamp is applying its EDEN (搜索) models, which it describes as biological foundation AI models, to in vivo cell therapy, an approach that reprograms a patient's cells inside the body. The company says current cell therapies are limited both in the complexity of what they can deliver and by a manufacturing process that costs hundreds of thousands of dollars per patient.
By pairing the models' ability to design long and complex DNA sequences with specialised enzymes that can precisely deliver them, Basecamp aims to make cell therapies more sophisticated, more customisable and simpler to administer. According to the company, EDEN (搜索) designs DNA sequences and recombinase proteins intended to insert large genetic payloads into cells without making double-strand breaks. If the in vivo approach works, it could avoid some of the manufacturing steps required to produce conventional CAR-T (搜索) therapies.
Preprint Data on Design Hit Rates
In a preprint article, Basecamp reported that EDEN (搜索) could design DNA-inserting proteins from a target sequence just 30 DNA letters long, including sequences absent from its training data. The authors reported a 63.2% functional hit rate, meaning roughly six in 10 designs tested showed activity in laboratory assays.
The team also tested whether the designs worked in human cells. Half of the 20 proteins tested successfully inserted a CD19 (搜索) CAR construct into human T cells, supporting the technology's potential for engineering cancer-fighting cells. The model's capabilities extended beyond gene insertion: 32 of 33 AI-designed antimicrobial peptides showed activity against bacteria in laboratory tests, including multidrug-resistant pathogens.
The Trillion Gene Atlas and Its Limits
The core dataset behind EDEN (搜索) comes from the Trillion Gene Atlas, which Basecamp calls the world's largest proprietary biological AI training dataset. The company builds it by working with biodiversity partners to collect and sequence biological samples across more than 30 countries and all seven continents. NVIDIA (搜索), Anthropic (搜索), PacBio and Ultima Genomics are among the partners in the Trillion Gene Atlas project.
The name describes the project's ambition rather than a completed trillion-gene dataset. Basecamp's website currently reports more than 100 billion novel genes.
Pipeline and Origins
Basecamp lists three in vivo CAR-T (搜索) programs on its website: one for blood cancers (搜索) and one for an undisclosed autoimmune indication are in lead optimization, while a solid-tumor program is in discovery. It also lists a gene-insertion program for phenylketonuria (搜索), antimicrobial peptides and a diabetes (搜索)-focused peptide program. The company says its pipeline of AI-designed therapeutics is advancing towards clinical development.
Basecamp was founded by Glen Gowers and Oliver Vince. Its roots date back to the co-founders' PhD studies in synthetic biology and bioengineering and a 2019 expedition to Iceland's Vatnajökull ice cap, where they set up the first off-grid DNA sequencing operation.
Gowers said: "We believe the future of medicine lies in reprogramming the body to repair itself. We design the models and the medicines to teach it how. AI-based approaches promise to change what's possible for patients who currently have few alternatives. This funding brings this technology closer to those who need it most."
Joséphine Kant, head of ventures at Sovereign AI, pointed to the founders' history of off-grid sequencing in the Arctic and said they "could build this company anywhere in the world, but they're unapologetically proud to be doing it here in the UK."