EMET, BenchSci’s agentic research environment, will run on Google Cloud as its primary infrastructure—establishing BenchSci as a key partner and provider of scientific reasoning and agents
Solution combines BenchSci’s proprietary biological knowledge graph, closed-access data, ontologies, agents, and scientific reasoning with Google’s frontier life sciences AI models, AlphaGenome and AlphaFold 3, in a single environment for preclinical R&D
TORONTO, August 26, 2026–(BUSINESS WIRE)–BenchSci, a leading provider of AI software for biopharma research and development, today announced a strategic, multi-year initiative establishing Google Cloud as the primary infrastructure platform for EMET, BenchSci’s agentic research environment for preclinical drug discovery.
The combined solution gives pharmaceutical scientists unified access to BenchSci’s proprietary biological knowledge graph—858 million nodes, 2.2 billion relationship edges, and legal access to 16 million closed-access scientific papers—alongside Google’s frontier life sciences AI models, including AlphaGenome and AlphaFold 3, within a single, purpose-built environment for preclinical R&D.
Seventy percent of drugs fail in the clinic, not due to lack of data, but because that data has never converged. Scientists spend roughly half of their time manually stitching together paywalled literature, siloed databases, disconnected tools, and internal data locked in spreadsheets. While other technical disciplines solved this infrastructure challenge, biology has historically lacked a unified foundation.
EMET addresses this gap by giving every scientist a series of agentic assistants—unifying data, models, software packages, workflows, and scientific reasoning into a single environment purpose-built for preclinical R&D. By running EMET on Google Cloud, BenchSci customers can now gain integrated access to the most advanced life sciences AI models available alongside BenchSci’s proprietary scientific foundation.
The majority of the top 20 pharmaceutical companies by revenue are existing BenchSci customers. Across its customer base, 80% of users report 60% efficiency gains, with some research teams achieving up to 50x time savings on complex workflows, according to BenchSci data.
“Every other technical discipline built a unified stack, and productivity exploded. Biology never did—until now,” said Liran Belenzon, CEO and co-founder of BenchSci. “Running on Google Cloud, with AlphaGenome and AlphaFold integrated natively alongside BenchSci’s knowledge graph, EMET gives scientists the complete picture of the biology before a drug ever enters the clinic. That’s how we close the 70% failure gap.”
