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

Drug discovery is one of the most costly problems in modern industry. Finding a single working molecule can take a decade and cost billions, and many candidates are still unsuccessful. The first generation of AI promised to fix that – most made the problem painful for researchers, who were already sophisticated enough to use the tools.

But SandboxAQ thinks the bottleneck isn’t models. Interface.

The company partnered with Anthropic to integrate its scientific AI models directly into Claude – putting powerful drug discovery and scientific tools behind a conversational interface that doesn’t require special computing infrastructure to run.

Founded nearly five years ago as an Alphabet spinout, SandboxAQ counts Eric Schmidt, the former CEO of Google, as its chairman. The company, which has raised more than $950 million from investors, has built many different business lines, including the cybersecurity business.

One of the most unique things SandboxAQ does, however, is generate large scale models, or LQMs. These identity models are “physics-based,” meaning they are built on the laws of the physical world rather than patterns in text. They can use quantum chemical calculations and simulate both molecular dynamics and microkinetics, the study of how chemical reactions occur at the molecular level. That’s important because it tells researchers how candidate molecules might behave before anyone sets foot in the lab.

“Trained with real-world lab data and scientific calculations, LQMs are AI models built for the economy of scale, a $50 billion+ industry that includes biopharma, financial services, energy, and advanced materials,” the company said in a press release that strongly suggests Sandbox AQ isn’t building another chatbot or coding assistant — which AI should chase down the economy.

Chai Discovery and Isomorphic Labs — both well-funded bets on better models — are focused on science. SandboxAQ focuses on who can actually use it.

“For the first time, we have a border [quantitative] a model at the LLM frontier that one can access in natural language,” Nadia Harhen, SandboxAQ’s general manager of AI simulation, told TechCrunch. Previously, SandboxAQ’s LQM users would have had to provide their own digital infrastructure to run the models.

SandboxAQ’s customers are often computer scientists, research scientists, or testers. Usually, these people work for large pharmaceutical or industrial companies and are looking for new things that can be marketed products.

“Our clients come to us because they’ve tried all the other software out there, and the complexity of their problem is that it didn’t work or didn’t bring them good results when that translation happened in the real world,” said Harhen.