Breaking
Multiply Labs reports 74% cost reduction for robotic cell therapy manufacturing NVIDIA + Universal Robots power the Multiply Labs cluster $85M partnership with Retro Biosciences announced May 2024 Series A: $20M led by Casdin Capital, Bee Partners and Y Combinator Pharma customers include AstraZeneca, Legend Biotech, Kyverna Therapeutics Founded 2016 in San Francisco by Fred Parietti and Alice Melocchi Multiply Labs reports 74% cost reduction for robotic cell therapy manufacturing NVIDIA + Universal Robots power the Multiply Labs cluster $85M partnership with Retro Biosciences announced May 2024 Series A: $20M led by Casdin Capital, Bee Partners and Y Combinator Pharma customers include AstraZeneca, Legend Biotech, Kyverna Therapeutics Founded 2016 in San Francisco by Fred Parietti and Alice Melocchi
Multiply Labs logo
YesPress / Company File No. 067

Multiply Labs

A San Francisco robotics company that decided cleanrooms should run themselves - and is rebuilding cell therapy manufacturing from the actuator up.

Frame: two co-founders, a fleet of robotic arms, and a very stubborn industry.
HQ: 1760 Cesar Chavez, SF Founded: 2016 Team: ~67 Stage: Series A+

Walk through a cell therapy cleanroom in 2026 and you will not see a single pipette held by a human hand. You will see a Universal Robots arm pivoting between bioreactors at 1760 Cesar Chavez Street in San Francisco, lifting a sealed cassette, sliding it into a sterile chamber, and signing the electronic batch record itself. The technicians are upstairs, in front of monitors. The patient whose cells are inside is somewhere across the country. The company that orchestrated all of this calls itself Multiply Labs - and would like a word about your factory.

01 / WHO THEY ARE NOWThe cleanroom went quiet

Multiply Labs is a robotics company that builds factories. Specifically, it builds robotic clusters - closed, sterile, software-controlled cells that take cell and gene therapy manufacturing end-to-end. The arms come from Universal Robots. The simulation comes from NVIDIA Isaac and Omniverse. The compliance comes from FDA 21 CFR Part 11. The customers, increasingly, are the largest pharmaceutical companies in the world.

It is a small company - around 67 people, based in a Dogpatch warehouse - doing very large things. AstraZeneca has signed on. Legend Biotech, the maker of one of the most consequential cancer cell therapies on the market, is a customer. Kyverna Therapeutics, an autoimmune cell therapy company, is on the roster. So is Retro Biosciences, the Sam Altman-backed longevity outfit that, in May 2024, committed to $85 million of robotic manufacturing capacity.

The fully robotic biologics factory is inevitable. Multiply Labs - company manifesto, on the front page

Caption: confident, declarative, refreshingly unaccompanied by hedging.

02 / THE PROBLEM THEY SAW$500,000 a dose, and a labor shortage

Cell therapies are arguably the most extraordinary drugs ever invented. They cure - actually cure - certain leukemias and lymphomas. A single dose of CAR-T can put a metastatic blood cancer into multi-year remission. The catch, as anyone who has ever read a pharmacy benefits memo knows, is that they cost around half a million dollars per patient.

There is a reason. Cell therapies are personal. Each dose is made from a single patient's own cells, shipped on ice to a manufacturing site, expanded by hand by trained operators in a Class B cleanroom, and shipped back. It is artisanal pharmaceutical work. It is also, by any honest accounting, unscalable. There are not enough trained operators in the world to make CAR-T for every eligible patient, and the ones who exist work in the most expensive real estate the industry has.

Multiply Labs noticed something the rest of pharma had been quietly avoiding. The bottleneck in advanced therapy is not biology. It is people in cleanrooms.

The bottleneck in cell therapy isn't chemistry. It's headcount, square footage, and gowning protocol. Fred Parietti, paraphrased from public interviews

03 / THE FOUNDERS' BETTwo PhDs walk into a YC batch

Fred Parietti is an MIT-trained roboticist. His co-founder Alice Melocchi is a pharmaceutical scientist with a PhD in industrial pharmacy from Milan. The pairing is the entire thesis of the company in two résumés: someone who knows how robots actually move, and someone who knows how pills - and later, living cells - actually behave under regulatory scrutiny.

They started Multiply Labs in 2016 with what now reads as a charming opening act: they built robots that assembled personalized vitamins. Customers could specify a stack and the system would 3D-print individual capsules, fill them, seal them, and ship them. It worked. It did not, in retrospect, solve a problem that anyone with a half-million-dollar treatment plan was losing sleep over.

The pivot was the correct one. The same robotic primitives - dispense, transfer, seal, log - turned out to map almost one-to-one onto the unit operations of cell therapy manufacturing. The hardware had been ready for years. It was waiting for the industry to admit it had a problem.

Co-founder

Fred Parietti

CEO. MIT robotics PhD. Speaks Italian and Python with equal fluency. The one who wrote the manifesto.

Co-founder

Alice Melocchi

Chief Scientific Officer. PhD in industrial pharmacy. The one who knows what 21 CFR Part 11 actually requires before you sign the email.

Habitat

1760 Cesar Chavez

A warehouse in San Francisco's Dogpatch district that does not look, from the outside, like the future of pharma. That is rather the point.

A short history of getting robots into the cleanroom

Milestones / 2016 - 2025
2016
Founded in San Francisco. First product: personalized vitamin capsules made by robots.
2017
Y Combinator W17 batch. The pharma-as-software pitch begins to sharpen.
2021
$20M Series A. The pivot to industrial cell therapy manufacturing is now explicit.
2023
Series A extension, $11.1M. Cluster v2 ships to first pharma partners.
2024
$85M Retro Biosciences partnership. NVIDIA Isaac / Omniverse collaboration deepens.
2025
Public results: 74% cost reduction and up to 100x throughput on cell therapy runs.

04 / THE PRODUCTOne cluster, many instruments, no humans

The Multiply Labs Robotic Cluster does not look revolutionary at first glance. It looks like a row of laboratory instruments that you have already seen - Cytiva, Thermo Fisher, Fedegari, Wilson Wolf, Akron Bio, GenScript - because, by deliberate design, it is. The company's wager is that pharma will not throw out its qualified instruments and start over. Pharma will, however, gladly let a robotic arm move samples between them, if the regulatory paperwork checks out.

So the cluster is, in essence, a sterile enclosure with collaborative robotic arms, a software layer that orchestrates every transfer, and an audit trail that satisfies FDA inspectors. It is the world's most boring science fiction. It is also why the platform works.

The software side is where the bet really lives. Every action - every cassette opened, every reagent dispensed, every minute a sample spent at a given temperature - is captured as an electronic record. That is not a feature. That is the only way regulated manufacturing has ever been allowed to operate. Multiply Labs simply has the receipts a robot can generate, which a human technician at hour seven of a shift may not.

What the cluster actually changes

Multiply Labs published results vs. manual baseline / 2025
Throughput
+100x
Cost / dose
-74%
Operator FTE
significantly reduced
Batch records
21 CFR Part 11 compliant
Source: Multiply Labs press materials, 2025. Numbers as reported.

Caption: every chart in pharma deserves a footnote. This one is no exception.

05 / THE PROOFThe customers who let robots near their drug

It is one thing to publish a 74% cost reduction in a press release. It is another to convince a Big Pharma quality team to let a robotic arm anywhere near a clinical batch. Multiply Labs has done both.

AstraZeneca announced a collaboration to manufacture cell therapies on the platform. Legend Biotech, the maker of Carvykti - one of the most clinically important CAR-T therapies of the decade - is publicly working with Multiply Labs. Kyverna Therapeutics, which is using CAR-T to treat autoimmune disease, is in the mix. So is Retro Biosciences, whose $85M commitment in 2024 was the loudest signal yet that the longevity field believes in robotic manufacturing.

On the technology side, NVIDIA picked the company for a flagship case study on Isaac and Omniverse, the chipmaker's robotics simulation stack. Universal Robots, the Danish cobot maker that has quietly become the picks-and-shovels supplier for almost every modern robotic lab, is the chosen arm.

If you can simulate the cleanroom, you can debug the cleanroom. That is news to most of pharma. YesPress, on why the NVIDIA collaboration matters

06 / THE MISSIONMake the cure cheaper than the disease

The company's stated mission is unfashionably plain: "Cost and production capacity should no longer stand between a patient and their cure." It is the kind of sentence that an investor deck used to talk you out of writing, because it sounded too earnest. The numbers, increasingly, give it cover. A 74% cost reduction is not a marketing slogan. It is, for an oncology payer, the difference between "we cover this for the sickest patients" and "we cover this for everyone who qualifies."

There is also the more uncomfortable version of the mission, the one Multiply Labs does not put on the front page. The global supply of trained cleanroom operators is finite. The number of patients eligible for cell therapy is growing every year. Without automation, the gap widens forever. Multiply Labs is, in the most literal sense, a labor-shortage company dressed as a biotech.

07 / WHY IT MATTERS TOMORROWThe factory as software

If you squint, the long-term vision is something stranger than "robots in a cleanroom." It is a world in which a cell therapy is not a thing a factory makes; it is a thing a factory's software is. A new therapy gets developed. It gets translated into a recipe of unit operations. That recipe runs on identical robotic clusters in San Francisco, Boston, and eventually somewhere outside Bangalore. The biology is bespoke. The execution is copy-paste.

This is not a uniquely Multiply Labs idea - it is roughly the dream of every advanced therapy CMO on the planet - but Multiply Labs is one of the very few companies that has shipped working hardware and signed real customers in pursuit of it. The competition is mostly slideware. The cluster, at this point, is not.

The cure becomes infrastructure. The factory becomes software. The cleanroom becomes a polite hum. The Multiply Labs proposition, in three sentences

08 / RETURN TO THE CLEANROOMThe hum, the arm, the receipt

Back in the Dogpatch warehouse, the cluster is mid-run. The arm pivots. A cassette moves from one instrument to another. A timestamp is written. Somewhere in the building, an engineer watches a dashboard and decides whether to get more coffee. The patient whose cells are inside will, in a few weeks, receive a dose of a therapy that might genuinely save their life. The dose cost a fraction of what it would have five years ago. Nobody in the room had to gown up.

Multiply Labs did not invent the cure. It is doing something less glamorous and arguably more important: it is making sure the cure can actually be made, at the price and the scale that turn a clinical miracle into a routine prescription. That is the company. That is the bet. The cleanroom went quiet, and the math finally started to work.