Multiply Labs Raises $75M to Automate Cell Therapy Manufacturing With Robotic AI Systems

Multiply Labs, a San Francisco robotics company that automates the production of cell and gene therapies, raised $75 million in a Series B round on October 6, 2026—bringing its total funding above $100 million since its 2016 founding. The round was led by Dr. Patrick Soon-Shiong through NantWorks and drew in a strategically notable mix of backers: pharmaceutical giant AstraZeneca, industrial robotics leader Teradyne, and returning investors Lux Capital, Founders Fund, and Casdin Capital. The central argument behind the company, articulated by co-founder and CEO Fred Parietti, is that AI-driven drug discovery has created an asymmetric problem: "AI is designing more therapies than the industry will ever be able to manufacture."

AI Drug Discovery Is Outpacing the Factories Built to Produce It

The promise of AI in pharmaceuticals has focused almost entirely on the discovery layer—identifying viable drug targets, designing candidate molecules, and compressing the years-long screening timelines that defined twentieth-century drug development. That focus has produced results. AI-designed therapeutics are moving through clinical pipelines at rates previously impossible. But the downstream problem those pipelines create has received less attention.

Advanced biologics—particularly CAR-T cell therapies, T-cell receptor therapies, and gene therapies—require manufacturing processes that remain stubbornly manual. A typical autologous cell therapy starts with a blood draw from a patient, whose T cells are then isolated, genetically reprogrammed, expanded through several rounds of culture, quality-tested at multiple checkpoints, and reformulated into a dose for reinfusion. Each hand-off between instruments is a potential contamination event. Each manual pipetting step introduces variability. The skilled technicians required for GMP (Good Manufacturing Practice) work are scarce, and cleanroom square footage is expensive.

Industry analyses estimate that labor-intensive manual processes account for as much as half of cell therapy manufacturing costs. Batch failure rates caused by contamination can exceed ten percent in manual workflows. Those failure rates compound across a therapy's commercial launch: every failed batch is not only a financial loss but potentially a delayed or missed treatment for a patient already enrolled in a program.

This is the bottleneck Multiply Labs is selling into.

Robotic Clusters That Fit Around Existing Instruments Rather Than Replace Them

Multiply Labs' core product is a robotic cluster—an enclosed, GMP-compliant enclosure in which up to four robotic arms operate in parallel. The arms move materials between instruments, execute timed liquid-handling steps, and handle the repetitive transfer operations that constitute the majority of manual labor in biologics manufacturing.

The system's defining commercial characteristic is its integration-first philosophy. Rather than requiring pharmaceutical manufacturers to redesign their production lines or abandon validated processes, Multiply Labs' clusters are designed to work around existing instruments: Cytiva bioreactors, Thermo Fisher cell culture systems, and other GMP-certified equipment already in use. A manufacturer adopting the platform does not need a new facility, nor does it need to re-validate its products from the ground up—a process that can cost tens of millions of dollars and years of regulatory engagement.

This matters because pharmaceutical process validation is not primarily a technology problem. FDA-regulated manufacturers must demonstrate that their processes consistently produce a product meeting predetermined specifications. Any material change to the process—new equipment, altered temperature profiles, different handling procedures—can trigger a revalidation requirement under 21 CFR Part 211 and related guidance. A robotics system that automates without altering validated parameters sidesteps this regulatory friction in a way that purpose-built replacement systems cannot.

Multiply Labs' AI layer draws on a partnership with NVIDIA announced in January 2026. The company uses NVIDIA Isaac Sim to build digital twins of its robotic systems, enabling software-in-the-loop testing before physical deployment. NVIDIA Isaac GR00T, the company's open robot foundation model, is being explored for learning new biomanufacturing manipulation skills. FoundationPose and FoundationStereo handle object pose estimation and depth perception in the GMP environment, turning scientist demonstrations into robot training data without disrupting ongoing operations. The NVIDIA collaboration is described in forward-looking terms—these are R&D workstreams rather than deployed capabilities—but they signal the direction of the platform's technical development.

Peer-Reviewed Validation Exists, but the 74% Cost Claim Requires Context

Multiply Labs' most cited figures are a 74% reduction in cost per dose and up to 100 times the throughput of manual manufacturing. Both numbers originate from company disclosures. A peer-reviewed study published in Cytotherapy, the official journal of the International Society for Cell & Gene Therapy, provided partial scientific grounding: the study demonstrated that robotic T-cell expansion produced outcomes—cell yield, viability, and phenotype—statistically equivalent to manual processes.

Several important caveats apply. The Cytotherapy study evaluated T-cell expansion, one stage of one category of cell therapy. It does not encompass the full end-to-end manufacturing workflow, nor does it address antibody production, viral vector manufacturing, or mRNA therapies—the categories into which Multiply Labs is now expanding. The 74% cost figure, per a separate company statement, was derived from industry data estimates rather than measured results from deployed commercial systems. The 100x throughput figure refers to output per square foot of cleanroom space under specific conditions, not absolute batch productivity.

This distinction matters because the company's current deployments are still in clinical-stage collaborations rather than commercial-scale operations. Moving from a validated clinical pilot to a certified commercial manufacturing run involves substantial additional regulatory work—process performance qualification, continued process verification, and batch record documentation requirements that increase in scope and scrutiny at commercial scale. Multiply Labs describes this transition to commercial-scale production as a primary use of the Series B proceeds.

AstraZeneca as Customer and Investor: A Structural Signal

The presence of AstraZeneca in this funding round carries meaning beyond capital. In January 2026, nine months before the Series B closed, AstraZeneca signed a separate agreement with Multiply Labs to evaluate GMP-ready robotic systems for commercial-scale cell therapy manufacturing. That evaluation agreement—described as an assessment of end-to-end automation with industry-standard instruments—placed AstraZeneca in the position of prospective customer before it became a financial backer.

Strategic investors in pharma typically do not co-invest alongside evaluation agreements unless they see a credible path to procurement or partnership expansion. AstraZeneca's cell therapy pipeline—centered on oncology—represents a direct commercial use case for the technology. Its investment is effectively a hedge on its own manufacturing needs.

Teradyne's participation is equally revealing. Teradyne is the parent company of Universal Robots, the Danish company that manufactures the collaborative robotic arms widely deployed in industrial automation. Its presence as a new investor suggests interest in the pharma-robotics market beyond what Universal Robots currently addresses—a possible technology supply relationship as Multiply Labs scales, or a strategic preview of demand for robotic components in GMP environments.

Other notable backers include Retro Biosciences, which signed a commercial and supply agreement with Multiply Labs in 2024 valued at up to $85 million to automate cell therapy manufacturing for age-related diseases. That deal provides the clearest public signal of commercial revenue trajectory, though its full terms remain undisclosed.

A Crowded Race With Better-Funded Rivals and Different Architectures

Multiply Labs is not building the only robotic solution for cell therapy manufacturing. The competitive landscape includes companies with substantially larger capital bases and different architectural philosophies.

Cellares, its most directly comparable competitor, had raised approximately $612 million by mid-2026—including a $257 million Series D closed in January 2026, led by BlackRock and Eclipse. Cellares' Cell Shuttle platform is designed as a closed, end-to-end manufacturing system that handles the full sequence of operations: enrichment, selection, transduction, activation, expansion, and formulation. The company claims up to 80% labor reduction and a 90% smaller facility footprint versus conventional manufacturing. Like Multiply Labs, Cellares received FDA Advanced Manufacturing Technology designation in 2025, a designation that facilitates expedited regulatory review of manufacturing processes.

Ori Biotech takes a different approach with its IRO platform, emphasizing a closed-system architecture and digitization of manufacturing data across the process. Ori had raised approximately $281 million through 2026 and also holds FDA AMT designation.

The architectural differences between these companies represent a genuine strategic fork. Cellares and Ori are building proprietary end-to-end systems—the manufacturing process runs inside their hardware, on their software, producing data flows in their ecosystem. Multiply Labs is building modular automation that slots into manufacturers' existing workflows. The end-to-end approach potentially delivers more comprehensive automation and tighter data integration. The modular approach reduces adoption friction and regulatory re-validation costs. Which architecture dominates commercially will depend on how drug manufacturers weigh implementation speed against total automation coverage—and how the FDA's AMT framework evolves to accommodate different automation philosophies.

Lonza and Charles River, established contract development and manufacturing organizations (CDMOs), represent a different kind of competition: not robotic competitors but institutional manufacturing capacity that cell therapy developers can access without capital investment. The risk for Multiply Labs and its peers is a market in which CDMOs invest in their own automation rather than purchasing third-party robotic platforms.

Commercial Scale Is the Next Threshold to Validate

Multiply Labs exits this funding round with meaningful but not conclusive evidence of commercial traction. Its clinical-stage partnerships with Legend Biotech, Kyverna Therapeutics, and Retro Biosciences demonstrate that drug manufacturers are willing to test the technology. The AstraZeneca evaluation agreement indicates strategic interest at the enterprise level. The Cytotherapy publication establishes independent scientific validation for one manufacturing stage.

What the public record does not yet confirm: the company's revenue run rate, the number of robotic clusters in active GMP deployment, customer renewal behavior, or the economics of deploying clusters against the cost of a CDMO relationship. The announced products expand into antibodies, viral vectors, and mRNA—categories that differ substantially from cell therapy in process complexity and regulatory classification. Whether the integration-first architecture that reduces adoption friction in cell therapy translates cleanly to those categories remains unproven.

The $75 million is explicitly positioned to fund the clinical-to-commercial transition. That transition will test whether a pilot-stage robotic platform can achieve the consistency, documentation rigor, and batch reliability that commercial biologics manufacturing demands. The industry is watching whether physical AI, in this context, means something more than well-engineered lab automation—or whether it becomes the infrastructure layer for an AI-accelerated generation of therapies that existing factories cannot produce fast enough to keep up.