SandboxAQ has announced the acquisition of Good Chemistry, a prominent computational chemistry firm that uses quantum computing, artificial intelligence (AI), and other cutting-edge technologies to accelerate the development of new drugs and materials.
The deal, worth approximately $75 million in cash and stock, will advance SandboxAQ’s computational chemistry and simulation capabilities by bringing in top-tier talent, tried-and-true technology, and invaluable industry knowledge gained from collaborating with pioneers in advanced materials and the life sciences.
Let’s get physics
SandboxAQ spun out of Google Alphabet two years ago, where team members had been incubating several technologies at the intersection of AI and quantum physics. Nadia Harhen, general manager of simulation and optimization at SandboxAQ, told GEN Edge that a few of these technologies could bear upon drug discovery, material science, and financial services.
“What we’re doing is building up computationally, atom by atom, running calculations, and then further training AI models on those calculations,” said Harhen. “We’re using physics to ground the reality of what nature wants to happen when a drug aligns with its target so that the AI is not hallucinating. It’s very much grounded in the physics of reality. That is what we’re doing in terms of drug discovery.”
Regarding drug development, SandboxAQ incorporates many elements with different types of machine learning data structures to discover new targets to comb over the corpus of failed clinical trials.
Harhen’s group specializes in the calculations that enable them to efficiently comb over that search base to deliver to a client, shaving years off their discovery and development time.
“What we’ve done is computationally develop methods to efficiently explore the chemical space and then exploit very specific areas because the chemical space is so large to explore, there’s so many combinations of what you could explore—it exceeds the atoms in the universe,” said Harhen. “No matter how much compute you have, that’s just a hard thing to solve.”
Good Chemistry
Good Chemistry, which spun off from 1Qbit in 2021, is SandboxAQ’s second acquisition and first in the AI simulation sector. As part of the acquisition, SandboxAQ will integrate Good Chemistry’s software—QEMIST Cloud and Tangelo—into its growing enterprise software portfolio.
Tangelo is a platform that bridges the gap between today’s classical world and the quantum computing world that’s backend agnostic. It enables SandboxAQ to “future-proof” their work to be compatible with a quantum computer—if it were ever to become a reality.
QEMIST Cloud is a product that efficiently searches the chemical space. Harhen said that QEMIST Cloud is already being deployed in material science and chemical settings, with customers using it daily, and SandboxAQ presents a robust platform to leverage for drug discovery efforts.
Partnerships for drug discovery
SandboxAQ does not have its own drug discovery or development pipeline. Instead, its computational tools reduce the time, cost, and risk of identifying promising pharmaceutical compounds by providing faster, more accurate predictions for lead generation and hitting targets. They can also accelerate drug development and clinical trials via in silico testing that measures the effect of the intervention.
“We want our customers to keep their drug IP,” said Harhen. “That is a very valuable part of the transaction to understand and develop new candidate molecules that they wouldn’t have ever found without us helping to tweak, modify, and optimize against their various criteria and delivering that back to them.”
When companies come to SandboxAQ, Harhen said, they typically bring hard problems because they’ve tried multiple computational tools without success, leaving them with “the rejects of what hasn’t worked” that are still highly impactful problems.
“We see hundreds and hundreds of clinical trial failures, and they’re labelled undruggable because no one’s been able to drug them yet because it’s complicated—it’s too complex to drug,” said Harhen. “That’s what we build specialized tooling to tackle, and that specialized tooling we’re going to make available on this platform.”
This news follows an announcement in November 2023 that Nvidia and SandboxAQ partnered to help companies simulate large volumes of chemical reactions to develop new materials for medicine and batteries. The drug discovery and development focus of that partnership is protein-ligand binding computations for large-scale undruggable targets in neurodegenerative disease and novel solutions for pose and toxicity prediction.
No longer will it take just ML and AI to play in the drug discovery and development sandbox—it’s going to include molecular simulation too.
