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

Drug discovery is one of the most costly failures in today’s industry. Finding a single viable molecule takes a decade and costs billions, and most of those who want to have it don’t make it. The first generation of AI has promised to fix this – many have made the problem less painful for researchers, who are already sophisticated enough to use the tools.

But SandboxAQ thinks the bottle isn’t colored. It’s a look.

The company has partnered with Anthropic to integrate its AI scientific models directly into Claude – putting powerful drug discovery and scientific tools behind an interactive interface that doesn’t require special computer hardware to operate.

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

One of the unique things that SandboxAQ does, however, is create large models, or LQMs. These proprietary models are “physics-based,” meaning they are built according to the laws of the universe rather than based on literature. They can calculate quantum chemistry and model molecular dynamics and microkinetics, the study of how reactions occur at the molecular level. This is important because it tells researchers how the selected molecules will behave before anyone enters the lab.

“Trained on real-world facts and scientific equations, LQMs are AI models designed for the multi-economy, a $50+ trillion sector that spans biopharma, financial services, energy, and advanced materials,” the company said in a press release that strongly suggests that Sandbox AQ is not building another chatbot or code assistant — it’s AI driving the economy.

Chai Discovery and Isomorphic Labs – Bets that are paid well on the best brands – focused on science. SandboxAQ focuses on who will use it.

“For the first time, we have the kind of limit (density) on the LLM limit that one can find in natural languages,” Nadia Harhen, SandboxAQ’s general manager of AI simulation, told TechCrunch. Previously, users of SandboxAQ’s LQM would have had to provide their own digital devices to run the models.

SandboxAQ’s customers tend to be statistical scientists, research scientists, or experimentalists. Usually, these people work in large pharmaceutical or industrial companies and are looking for new products that can be marketed.

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