This week, we feature insights from Radical Partner Sanjana Basu’s article on the transformative impact of AI in the sciences. As AI continues to penetrate all areas of our lives, Sanjana explores how applying generative AI to fields like biology, chemistry, and physics holds the potential for exponential value creation.
AI-accelerated scientific discovery has entered the mainstream.
In 2024, Demis Hassabis and John Jumper won the Nobel Prize in Chemistry for their groundbreaking work on protein structure prediction. Radical portfolio company Nabla Bio announced JAM, their generative model for antibody design, demonstrating clear evidence of AI’s ability to create novel antibodies with drug-like properties. Another Radical portfolio company, Orbital Materials, released Orb, the fastest and most accurate AI model for simulating advanced materials, beating models from Google and Microsoft. We have also seen progress toward AI transforming the scientific method itself — Sakana AI published their work on an AI Scientist automating hypothesis generation, experiment design, and data analysis, while Stanford research showed how LLMs aid researchers by suggesting novel ideas and acting as creative collaborators in generating research directions.
Early stage investment opportunities in AI for science are keeping pace with this accelerated rate of breakthroughs, driven by three primary factors:
Scarce Talent: While leading labs like Google DeepMind have paved the way with breakthroughs like AlphaFold and GNoME, a growing number of technically differentiated teams, prioritizing domain expertise, are leaving leading big tech labs and research institutes, to launch companies in this space. Top talent equally conversant in AI and scientific disciplines is rare and creates inherent value; and multidisciplinary teams are able to productize and commercialize innovations within accelerated timeframes. Notable examples of best in class teams building companies in this category are Orbital Materials, founded by DeepMind alumnus Jonathan Godwin; Nabla Bio, founded by Surge Biswas, author of the 2019 UniRep paper that pioneered the application of language models to protein engineering and Frances Anastassacos, Ph.D. in Biological and Biomedical Sciences from Harvard; and Latent Labs, founded by Simon Kohl who led AlphaFold’s protein design team.
Increased Industry Appetite: Industry appetite for AI solutions is maturing, particularly in life sciences. Major pharmaceutical companies are rapidly expanding their AI teams, and have existing arrangements to work with AI companies, enabling testing and scaling of the best solutions. Examples of these agreements include Genesis Therapeutics’ $670 million collaboration with Eli Lilly and Aspect Biosystem’s $2.6B deal with Novo Nordisk. Energy and industrial sectors are following suit, as evidenced by Occidental Petroleum’s purchase of carbon capture company Carbon Engineering for $1.1 billion.
Exponential Value Creation: AI applied to the sciences is poised to create some of the largest value categories in the future, with some analysts pegging the bio-economy between $4 trillion to $30 trillion globally. These are trillion-dollar industries waiting to be created, and the work has already begun.
Despite their complexity, capital requirements, and inherent risks, the timing could not be better for building companies that combine human and artificial intelligence to solve previously intractable problems.
Sanjana writes a regular blog. Read her article in full.
AI News This Week
-
Vancouver’s Aspect Biosystems raises $115 million to develop 3-D printed human tissue technology (The Globe and Mail)
Radical portfolio company Aspect Biosystems secured US$115 million in series B funding, led by Dimension with participation from Radical Ventures and other investors. The company develops 3D printers that create synthetic tissues from living cells and hydrogel polymers to replace or enhance organ functions. The round follows Aspect’s US$2.6 billion 2023 partnership with Novo Nordisk to develop treatments for diabetes and obesity. The company plans to use the capital to advance multiple programs into clinical trials and to execute its government-backed US$200 million manufacturing facility project in Vancouver.
-
Canadian VCs try to fill an AI compute gap for startups (The Logic)
Radical Ventures launched a dedicated compute cluster in the fall as part of the Radical AI Founders Masterclass program with the goal of ensuring founders and entrepreneurial researchers have the infrastructure to bring their designs to life. In partnership with Google Cloud, Radical AI Founders participants can apply for access to dedicated clusters of Nvidia GPUs and Google TPUs. Through the Radical AI Founders Masterclass program, Radical is offering up to US$250,000 in Google Cloud credits to early-stage AI startups, supporting researchers looking to commercialize their work.
-
What will viruses do next? AI is helping scientists predict their evolution (Nature)
Researchers are making significant strides in using AI to predict viral mutations, particularly for SARS-CoV-2 and influenza. Tools like EVEscape from Harvard Medical School and CoVFit from the University of Tokyo successfully predict short-term viral changes and test potential future variants. The models combine massive viral sequence databases (nearly 17 million for SARS-CoV-2) with experimental data to forecast which mutations might succeed. Researchers are working to forecast major evolutionary leaps to enable pre-emptive vaccine and antiviral treatment development.
-
At the intersection of A.I. and spirituality (The New York Times)
As AI becomes increasingly woven into the fabric of society, religious leaders are integrating it to expand their reach and enhance spiritual practices. Rabbi Josh Fixler pioneered “Rabbi Bot,” an AI trained on his sermons that successfully writes and delivers content in his voice, demonstrating AI’s potential in religious spaces. The technology is helping leaders translate sermons in real-time for international audiences and providing instant access to vast theological resources.
-
Research: Towards understanding systems trade-offs in retrieval-augmented generation model inference (Cornell/NVIDIA/Penn State)
As enterprises increasingly adopt RAG to enhance LLMs with private data without costly retraining, researchers seek to understand the trade-offs between performance, memory usage, and accuracy in RAG systems. This study reveals retrieval operations account for 41% of processing time, with memory-efficient algorithms using 2.3× less RAM but achieving lower accuracy (0.65 vs 0.95 recall). When data stores grow from 1 to 100 million entries, throughput decreases by up to 20×, demonstrating key scalability challenges for production deployment.
Radical Reads is edited by Ebin Tomy (Analyst, Velocity Program, Radical Ventures).