SandboxAQ Launches AI-Powered Drug Discovery Tool Without Protein Structures — SandboxAQ

Event summary

SandboxAQ launched AQPotency, an AI model for drug discovery that predicts molecule-target interactions without requiring protein structures.
The tool costs $1 per 1,000 comparisons and provides confidence intervals for predictions, making it more actionable than traditional methods.
AQPotency is available through Claude via Model Context Protocol (MCP) and will soon be on Google Cloud's Marketplace.
The model has been validated in eight customer programs, including Parkinson's research at the University of Dundee and Columbia University.
SandboxAQ also launched AQCat Adsorption Spin, another Large Quantitative Model for catalyst discovery.

The big picture

SandboxAQ's AQPotency addresses a critical bottleneck in early-stage drug discovery by eliminating the need for protein structures, a requirement that has stalled many promising programs. This launch aligns with the broader trend of AI-driven tools transforming R&D efficiency in biopharma, potentially lowering costs and accelerating timelines for novel therapeutics. The company's strategic backing from high-profile investors underscores the market's confidence in its ability to deliver scalable solutions at the intersection of AI and quantum techniques.

What we're watching

Adoption Pace: How quickly biopharma companies will integrate AQPotency into their workflows, particularly for targets without solved structures.
Validation Impact: Whether the experimentally validated success in eight programs will translate to broader industry acceptance.
Competitive Response: How existing players in computational drug discovery will react to AQPotency's cost and speed advantages.