Atomic AI

Atomic AI was founded in 2021 in South San Francisco, California by Raphael Townshend, a Stanford-trained researcher whose graduate work on predicting RNA three-dimensional structure with geometric deep learning was published on the cover of Science (2021) and covered by Stanford News and the NIH Director's Blog. Rather than continue the research in academia, Townshend chose to commercialize it, founding Atomic AI to apply AI/deep-learning foundation models to structural biology, with an initial focus on RNA — a molecule whose structure had historically been far harder to resolve computationally than that of proteins (the domain AlphaFold had tackled).

The company's core technology, ATOM-1, is a large language/foundation model trained on proprietary chemical mapping data, designed to predict RNA structure and function more accurately than prior methods. ATOM-1 is embedded in a 'virtuous cycle' with purpose-built, in-house wet-lab (chemical mapping) assays: model predictions guide experiments, and experimental results are fed back to continually improve the model and expand the company's proprietary 3D RNA structural database — which Atomic AI describes as aiming to be the world's largest. The platform is intended to support both RNA-targeted small-molecule drug discovery and RNA-based therapeutics, targeting diseases considered 'undruggable' by conventional means.

Atomic AI raised a $7 million Seed round in 2022 from 8VC, Greylock, Factory, AIX Ventures, and AME Cloud Ventures. It emerged from stealth on January 25, 2023 with a $35 million Series A led by Playground Global, with participation from its seed investors plus angel/individual investors Nat Friedman and Patrick Hsu, and Not Boring Capital (Packy McCormick's fund, which published a widely read investment-thesis essay on the company the same day). Crunchbase records a third, financial-terms-undisclosed funding event/round beyond the Seed and Series A (with Modi Ventures cited among ~15 total investors), which is not corroborated by public press releases and should be treated as unconfirmed.

Business model

Biotechnology/AI drug-discovery platform company; generates revenue via pharma partnerships and out-licensing of RNA-targeted small molecule and RNA-based therapeutic programs developed on its ATOM-1 platform, alongside continued equity financing to fund internal pipeline and platform R&D.

Pharma/biotech R&D partnerships and licensing deals (business development contact: bd@atomic.ai); pre-revenue/pre-commercial on any owned therapeutics pipeline as of last confirmed reporting.