London-based startup Basecamp Research has recently secured $60 million in funding and is developing a remarkable AI assistant for biology. This project aims not only to answer any questions related to biology and biodiversity in the natural world but also promises new insights that humans cannot independently obtain, potentially leading to revolutionary breakthroughs in the biotech field.
Glen Gowers, co-founder and CEO of Basecamp Research, points out that there is a significant data gap in current biology model training. Even the world's top pharmaceutical companies cannot fully encompass the complexity of the natural world with their trained models. This highlights the importance and potential impact of Basecamp Research's project.
Image source note: The image was generated by AI, provided by the image licensing service Midjourney
The company is experiencing strong growth. According to Gowers, Basecamp Research has established partnerships with over 100 organizations in 25 countries to expand its raw information database. About 15 of these organizations are using its AI technology to develop new products. For example, Procter & Gamble is using these models to design enzymes for detergents that can remove stains at low temperatures; Colorifix is working on developing more sustainable fabric dye formulations.
More notably, Basecamp Research claims that its foundational model, BaseFold, outperforms DeepMind's AlphaFold2 model, which recently won the Nobel Prize in Chemistry, in accurately predicting the structures of large, complex proteins and interactions with small molecules. This assertion certainly bolsters the credibility of Basecamp Research's technical prowess.
Basecamp Research adopts an ambitious approach to building biological AI: starting from scratch. Founders Glen Gowers and Oliver Vince both hold PhDs in biology, and their collaboration began during their undergraduate studies at the University of Oxford. The company name "Basecamp Research" stems from their experience living on an ice sheet, where they used homemade hardware for DNA sequencing.
Vince mentioned that they pioneered the first mobile DNA sequencing lab and modified some components of the hardware into smaller units for data collection by startups. This innovative spirit and practical experience have laid a solid technical foundation for Basecamp Research.
Although the field of biology has accumulated a wealth of research results and data, much of it is outdated, unstructured, or inconsistent. Therefore, Basecamp Research is building models by collecting original data firsthand. Their goal is to develop an AI system that can gain deeper insights into biology than any human, thanks to its ability to process and analyze vast amounts of data.
Gowers explained that they combine exploration activities (such as expeditions to hot springs and volcanoes around the world) with AI programs focused on training large language models, essentially creating a "ChatGPT of the natural world." The company has also established what is possibly the largest computing cluster dedicated to researching the natural world.
Similar to ChatGPT's ability to recall and construct natural language responses, Basecamp Research's AI also possesses this capability. The difference is that, given our understanding of biodiversity is limited to about 1%, humans currently cannot even ask the right questions. As investor and former CEO of Google's Verily Life Sciences, Andy Conrad, put it, Basecamp Research's platform can "answer questions that the biopharmaceutical industry hasn't even realized it needs to ask."
Gowers further explained that their platform not only understands text or speech but also understands the language of DNA and biology, thus surpassing human capabilities in the field of biological design. Traditionally, humans have limitations in understanding DNA, but these language models, if given enough data, can excel in this area.
This Series B funding round was led by European company Singular, and Basecamp Research also announced a multi-year collaboration with Dr. David R. Liu and the Broad Institute of Harvard-MIT. The company plans to use the funds to continue expanding partnerships with other biomedical and research organizations and to collect more data to expand its models.
Basecamp Research's future plans include assisting organizations in drug discovery and other significant challenges involving understanding and better utilizing the natural world. Currently, Dr. Liu's lab is researching "new fusion proteins and other large molecules" for creating gene therapies and using Basecamp Research's dataset to develop these molecules.
It is worth noting that Basecamp Research currently focuses more on B2B business rather than developing products for the general public. This strategy seems to be adopted by other companies building large "scientific" models, such as Jua, which initially targeted organizations needing better insights into weather patterns.
Although Basecamp Research did not disclose the specific valuation, it indicated that this Series B round was an up round. The company has so far raised $85 million, with previous investors including Hummingbird, True Ventures, and strategic investor Valo. According to PitchBook data, the company was valued at $71 million in 2022.
This round of financing also attracted participation from notable investors such as S32, redalpine, Roche's deputy chairman André Hoffmann, Royal Philips' chairman and former DSM CEO Feike Sijbesma, and former Unilever CEO Paul Polman.