Artificial intelligence has been a big theme in the world of health and medical research, and specifically in the area of drug discovery. Today, another hopeful in the space is announcing a funding round to expand its own contribution to the field. Causaly, a London startup that has built an AI platform to help researchers accelerate the development and testing of drugs, has raised $60 million, a Series B that will be going toward R&D and to continue building out its team.
ICONIQ Growth — the growth-stage fund affiliated with the iconic investment firm of the same name — is leading the round, with previous backers Index Ventures, Marathon Venture Capital, EBRD, Pentech Ventures and Visionaries Club also participating. The company has now raised $93 million in total and is not disclosing valuation.
Causaly is just over six years old, and Yiannis Kiachopoulos, the CEO who co-founded the company with CTO Artur Saudabayev, said that it already works with 12 of the world’s biggest pharmaceutical companies and some of the biggest names in medical research, including Gilead, Novo Nordisk, Regeneron, the Food and Drug Administration and the National Institute of Environmental Health Sciences.
These organizations use its cloud-based platform to work across the different stages that go into developing drugs: identifying interesting targets for research and development, determining biomarkers that are specific to those targets and aiding in pathophysiology to better understand a disease in order to determine what might be fixed with the right pharmaceuticals and other therapeutics.
Kiachopoulos estimated that the use of Causaly’s platform can reduce the 10-15 years that it might typically take to take an idea from target to the end of trials, down to around “several” years — a major reduction in the budget that needs to be dedicated to the process.
Just as importantly, its platform — which enables faster modeling and computations based on different chemical permutations and how they work in different environments — aims to reduce the number of false starts and dead ends that characterize the process of drug discovery.
“For each drug to make it to the market there are nine that failed,” said Kiachopoulos, working out to a 90% failure rate. Each of those drugs typically costs between $1 billion and $2 billion to develop, according to research from the National Institutes of Health in the U.S. “This gives us a real chance to accelerate and provide patient and societal benefits.”
The immense inefficiency in the biomedical research system is the classic kind of big data problem that suits AI — which can not only crunch large, multifaceted calculations in real time, but be applied to read images to better understand results on cells and more — and that is one reason it’s been a popular field not just among AI startups, but investors, too. Just yesterday, Recursion — an AI-based drug discovery startup that has raised hundreds of millions of dollars in funding — announced its latest investment, a $50 injection from Nvidia that came with an important strategic partnership: Recursion would use Nvidia’s cloud platform to train its models on giant datasets.
That deal underscores the immense amount of money that is being pumped into the AI drug discovery space — overall there have been billions put into startups in the field — but interestingly it also highlights something else.
I asked Kiachopoulos if compute power was an issue for his startup as well, given that this is indeed one of the big themes among AI startups right now, biomedical or otherwise, and his answer was a surprising “no.”
“Only a very small fraction will go into compute resources,” he said. This was partly due to how Causaly was built, and partly because of its role in the ecosystem. “Six years ago, when we were starting the company, there were no large language models, so what we have built is not compute-power hungry. We were building natural language querying before Chat GPT, and so we didn’t need large language models now.”
He did say that it’s working on incorporating more of this into future products, but that this was not going to have a noticeable impact on its compute needs.
“With LLM it can get easier to query AIs. That is true and we are working on that. But you don’t need to train an LLM from scratch so we can take and fine tune what there is, and fine tuning is a lot less of a drain on compute resources.”
The other detail that this highlights is that Causaly itself is not in the business of drug discovery: It’s providing tools to others who are. This is also something that differentiates Causaly from other startups in the field.
“Our solution helps biomedical teams, but we are not developing our own therapeutics,” he said. “We are a SaaS-based platform, training our scientists to get the most out of our AI. We have very strong partnerships and not competing, nor do we have plans to.”
With this round Caroline Xie, a general partner at ICONIQ Growth, is joining the startup’s board.
“The sciences are at a turning point driven by the adoption of AI, and we believe Causaly is a leader in delivering this power to scientists in a highly trusted and verifiable manner,” she said in a statement. “Causaly stands out to us as a uniquely powerful and user-oriented platform applying AI to drive significant productivity gains and commercial impact for many major pharmaceutical companies today. We are thrilled to support the entire Causaly team in their mission to revolutionize the way scientists find, visualize, and collaborate on scientific evidence across pharma, life sciences, and beyond.”
“Causaly gives scientists the power to solve the world’s biggest challenges like never before. It is one of the clearest real-life applications of AI today,” added Carlos Gonzalez-Cadenas, a partner at Index Ventures. “Already rolled out by some of the world’s largest pharmaceutical companies, Causaly is actively accelerating biomedical research now. We’ve been truly impressed with the level of adoption by leading research organizations, who continue to rapidly expand spend on Causaly, underlying the impact the technology is already having on R&D.”
Updated to correct the total amount raised to date and the time reduction (from six to “several” years).
Comment