The Manufacturing Gap: Why Cell Therapy Needs a Semiconductor Moment
The biologics industry solved scale through automation and process standardization. Cell therapy hasn't. Here's what it will take — and why it matters more than any single therapy.
The semiconductor industry didn't become what it is today because of better transistors. It became what it is because of better factories.
When Intel and TSMC invested in photolithography, clean rooms, and yield optimization, they weren't just making chips faster — they were making the *manufacturing of chips* a solved problem. Reproducibility became a feature, not an aspiration.
Cell therapy is at the same inflection point, and most of the field doesn't know it yet.
The Reproducibility Problem
Every GMP batch of iPSC-derived cells we produce starts the same way: with a biological input that varies. Donor cells vary. Reprogramming efficiency varies. Colony morphology varies. Even with standardized protocols, trained operators, and controlled environments, a cell therapy product made on Tuesday is not identical to the one made on Thursday.
This isn't a failure of science. It's a failure of infrastructure.
What Automation Actually Solves
When people hear "automated cell manufacturing," they imagine robots replacing scientists. That's the wrong frame. What automation actually does is *eliminate the decisions that shouldn't be decisions*.
Passage timing. Media exchange intervals. Confluence assessment. These are not scientific judgments — they are process execution tasks that humans perform inconsistently and robots perform identically. When we deployed our automated iPSC reprogramming system, batch-to-batch variability in colony morphology scores dropped significantly. Not because we got better at science. Because we removed human variability from a task that was never meant to require human judgment.
The AI Layer
The next unlock isn't more automation. It's closed-loop optimization.
We're building systems where machine learning models trained on thousands of manufacturing runs can predict, in real time, which batches are trending toward failure — before any release assay would catch them. Digital twins of our bioreactor processes allow us to run in silico process development before touching a single cell.
This is the semiconductor moment for cell therapy. The factories are being built. The question is whether the field moves fast enough to fill them.
Views expressed in this post are solely those of Dhruv Sareen in his personal and academic capacity and do not reflect the positions of any affiliated institution or organization. Full disclaimer
← Back to The Frontier