Alphabet Spinout SandboxAQ, Backed by Nvidia, Unveils Synthetic Molecule Megaset to Revolutionize Drug Discovery

SandboxAQ, an AI startup spun out of Google parent Alphabet and backed by Nvidia, has announced that it has released a massive dataset of 5.2 million synthetic 3D molecules to accelerate drug discovery. These molecules don’t exist in nature—they were generated using Nvidia’s chips and grounded in real-world experimental data to simulate how drugs bind to proteins, a critical step in developing effective treatments.

This synthetic dataset, called the Structurally Augmented IC50 Repository (SAIR), is publicly available and designed to train AI models that can predict drug-protein interactions far faster than traditional lab methods. The goal? To virtually replicate lab results with high accuracy, potentially compressing months of research into a single AI-driven prediction.

It’s a bold move that blends physics-based modeling with machine learning—an approach that could reshape how we think about early-stage pharmaceutical R&D. And with SandboxAQ planning to monetize its own trained models, it’s also a glimpse into the future of AI-powered biotech platforms.

While the dataset is public, SandboxAQ plans to monetize its proprietary AI models trained on it—essentially offering virtual labs as a service.

Headquartered in Palo Alto, California, right in the heart of Silicon Valley, SandboxAQ is a fascinating fusion of quantum tech and AI muscle. It emerged from Alphabet’s moonshot factory and is backed by Nvidia, with nearly $1 billion in venture capital fueling its ambitions.

SandboxAQ's signature tech is Large Quantitative Models, the AI systems grounded in the laws of physics, chemistry, and biology. These aren’t just data-driven models; they simulate the real world with scientific rigor. Think of it as a tireless digital researcher. This tool autonomously explores millions of chemical pathways, helping discover novel molecules far beyond human capacity.

Beyond drug discovery, SandboxAQ is tackling – 1) Cybersecurity: Detecting cryptographic vulnerabilities; 2) Navigation: Enhancing precision in GPS-denied environments. 3) Medical Diagnostics: Using AI to analyze cardiac signals. 4)Materials Science: Predicting atomic-level properties for breakthrough materials.

SandboxAQ was founded by Jack Hidary, a tech entrepreneur with a deep background in quantum computing and neuroscience. He previously led a quantum initiative at Alphabet before spinning it out as SandboxAQ in 2022. Hidary studied philosophy and neuroscience at Columbia University and has authored a well-regarded book on quantum computing.

While Hidary is the primary founder and CEO, the company’s early development was also shaped by influential figures like Eric Schmidt, former CEO of Google, who serves as Chairman of the Board.