Baseten distribution via Benchling

This partnership turns Benchling into Baseten’s wedge into biopharma R&D. Benchling already sits inside the daily system of record for experiment design, sample tracking, and scientific data across roughly 1,200 to 1,300 biotech customers, so adding Baseten powered inference lets model serving show up where computational scientists already work, instead of asking Baseten to win a separate platform evaluation led by central IT or procurement.1 2 3

• Benchling Inference was launched in May 2026 as Benchling’s way to give biopharma teams scalable GPU capacity, preloaded scientific models, and infrastructure that scales up for bursty workloads and back down when lab cycles go quiet. That makes Baseten the engine underneath a Benchling owned workflow, not a stand alone tool scientists must separately buy.2
• Benchling’s distribution matters because it already reached scale before this launch. Benchling was estimated at $210M ARR in May 2024 with about 1,200 customers, and its blog now describes the company as supporting more than 1,300 biotechnology companies. Baseten gets immediate access to a large installed base of labs that already trust Benchling with sensitive R&D workflows.1 3 4
• The user path is bottom up, not classic enterprise software sales. Benchling historically spread by getting scientists to adopt its lab software early, then expanding into larger R&D and IT deals. Putting inference behind notebooks, APIs, and Benchling embedded tools follows the same pattern, scientists use models inside familiar workflows first, then spend can grow from usage.3 5 6

The likely next step is a tighter stack where Benchling becomes the front end for scientific AI and Baseten remains the compute layer underneath. As more model driven biology work moves from one off experiments into routine lab workflows, distribution through incumbent research software should become a faster path to revenue than selling raw inference infrastructure lab by lab.2 3 6