The privately held AI drug discovery company says a new deal with synthesis shop onepot links its virtual design platform to 2.7 billion ready to make compounds, while its chief scientist takes a co editor credit on a major medicinal chemistry reference.
Receptor.AI, a privately held small-molecule AI drug discovery company, is making a concrete claim about timing: computationally designed hits can reach wet labs as assay-ready compounds in about five business days, instead of the months that often pass between virtual design and bench validation.
The number comes from a new collaboration with onepot, an automated synthesis company whose CORE chemical space covers roughly 2.7 billion synthesis-ready compounds (Receptor.AI and onepot partner announcement). The deal links onepot's automated synthesis machinery to Receptor.AI's REAXENSE chemistry-enablement platform, so a structure prioritized in software can in principle be queued for synthesis and delivered to a lab within a workweek. The bundled announcement was picked up in TipRanks' private-companies coverage (TipRanks recap).
That targets the slowest handoff in early drug discovery. The standard Design-Make-Test-Analyze (DMTA) loop, in which a team designs a molecule computationally, synthesizes it, tests it, and feeds the results back into the model, loses most of its time at the synthesis step, where AI-discovery shops without in-house chemistry either contract out or wait on a partner queue. A five-business-day turnaround, if it holds at scale, would compress a real bottleneck rather than decorate a slide. The 2.7 billion figure is platform-stated, not independently benchmarked, and the throughput metric reflects the combined capacity the partners are claiming rather than measured output from a finished campaign.
The onepot deal is the operational news. The other half of the announcement is reputational. Receptor.AI's chief scientific officer, Dr. Daniele Andreotti, was named co-editor of Volume 4 of Comprehensive Medicinal Chemistry IV, a reference series focused on cancer and G-protein coupled receptor (GPCR) drug discovery. GPCRs form one of the largest families of drug targets and include the receptors for everything from adrenaline to dopamine. Per company materials, Andreotti has more than 30 years of medicinal chemistry experience and has contributed to 15 or more preclinical candidates across central nervous system, anti-infective, and metabolic disease programs. An editorship co-credit on an established reference volume is a credibility signal, not a pipeline milestone, and should be read as such.
The onepot agreement is Receptor.AI's third publicly disclosed partnership of 2026. In March, the company and PRISM BioLab announced a collaboration covering protein-protein interaction targets, a notoriously hard class of drug targets because the binding surfaces between proteins are usually flat and featureless, and PRISM's α-helix mimetic chemistry platform, which mimics the helical shapes proteins use to grip each other (PR Newswire). Separately, Receptor.AI is working with Sareum to discover blood-brain-barrier-permeable inhibitors of TYK2 and JAK1, two kinase enzymes implicated in neuroinflammation, where crossing into the brain is the hard part (Sareum collaboration announcement).
Both collaborations sit on top of the REAXENSE platform, which Receptor.AI folded into its stack earlier in the year to move from pure virtual screening toward full-stack discovery that includes chemistry, biology, and preclinical development (Reaxense incorporation announcement).
The partnerships and the credentialing effort converge on a single argument. Receptor.AI is no longer pitching "AI drug discovery" as a black box and is trying to back it with named chemistry, named partners, and a named chief scientist. The BIO International Convention 2026 in San Diego, where the company plans to meet prospective pharma and biotech partners around discovery collaborations, candidate optimization, and pre-Investigational New Drug (pre-IND) or clinical-stage in-licensing deals, extends that surface area into face-to-face dealmaking (Receptor.AI on LinkedIn).
The hard question is what would falsify the five-day claim. Several things would. Compounds that fail synthesis feasibility checks and fall back to traditional make-on-demand timelines. Assay-ready material that arrives but does not perform as the virtual screen predicted. A steady drumbeat of partnership announcements with no disclosed candidates advancing. The 2.7 billion synthesis-ready compounds cited by the company are a count of previously digitized chemistry, not novel matter, and the throughput metric is platform-stated rather than independently benchmarked. Until a candidate from this stack is disclosed, the test is whether the timing survives contact with real medicinal chemistry campaigns rather than headline-scale press releases.