SandboxAQ brings its drug discovery models to Claude — no PhD in computing required

In the realm of drug discovery, the journey from concept to viable molecule is fraught with challenges, often stretching over a decade and costing billions. Despite advances, many candidates fail to succeed. A new wave of AI startups is emerging to tackle these obstacles, but SandboxAQ believes the real issue lies in accessibility, not just advanced models. Teaming up with Anthropic, SandboxAQ is integrating its sophisticated scientific AI models into Claude, a conversational interface designed to simplify the drug discovery process. This innovation allows users to leverage powerful tools in materials science without needing specialized computing knowledge or infrastructure. Launched about five years ago as a spinout from Alphabet, SandboxAQ has attracted over $950 million in funding and boasts Eric Schmidt, former CEO of Google, as its chairman. The company has diversified its offerings, including a cybersecurity division, but its standout feature is its development of large quantitative models (LQMs). These models are unique in that they are ‘physics-grounded’, meaning they are based on established physical laws rather than mere text patterns. The ability of these LQMs to perform quantum chemistry calculations and simulate molecular dynamics is crucial, as it provides researchers insights into how potential molecules will behave before any lab work begins. "Trained on actual lab data and scientific principles, LQMs represent AI models tailored for the quantitative economy, encompassing sectors worth over $50 trillion, including biopharma and advanced materials," the company stated, indicating that SandboxAQ is not just creating another chatbot but is instead targeting the transformative potential of AI. While competitors like Chai Discovery and Isomorphic Labs concentrate on enhancing scientific models, SandboxAQ is prioritizing user accessibility. "For the first time, we have a frontier quantitative model on a frontier LLM that can be accessed using natural language," explained Nadia Harhen, SandboxAQ’s general manager of AI simulation. Previously, users had to have their own digital infrastructure to utilize SandboxAQ’s LQMs. The company's clientele, which often includes computational and research scientists in pharmaceutical and industrial sectors, seek innovative materials for commercial application. "Our clients approach us after trying other software solutions that fell short due to the complexities of their challenges, especially when translating laboratory results to real-world applications," Harhen added.