Harvard Silicon Chip Writes 64 DNA Sequences at Once, Clearing a Major Gene Therapy Hurdle
Researchers at Harvard University published a study in Nature Electronics on July 8, 2026, describing a silicon chip that synthesizes 64 distinct DNA sequences simultaneously using electrical currents and water-based enzymes. The advance sets a new benchmark for enzymatic DNA synthesis and could eventually reduce the cost and environmental footprint of producing the custom genetic material that modern medicine increasingly depends on.
The problem with how DNA is made today
Most synthetic DNA is produced through phosphoramidite chemistry, a well-established process that can manufacture millions of sequences in parallel. The catch is that it relies on toxic organic solvents and requires centralized industrial facilities. Enzymatic synthesis has been proposed as a cleaner, safer alternative because it works in water, similar to how living cells build DNA. Until now, though, it could only produce about a dozen sequences at a time. The Harvard chip managed 64 distinct sequences in a single run, each up to 39 nucleotides in length.
That gap matters for medicine. Synthetic DNA is essential for diagnostics, genome engineering, and cancer research. Gene therapy, which delivers corrective genetic instructions directly to cells, requires custom sequences tailored to individual patients. Treatments currently on the market carry price tags from several hundred thousand dollars to more than $3 million per course of treatment. Reducing the manufacturing cost of synthetic DNA would not automatically lower what patients pay, but it removes one significant barrier in the supply chain.
How the chip actually works
The chip was not originally designed to synthesize DNA. Jeffrey Abbott, a former PhD student in the lab of Donhee Ham at Harvard's John A. Paulson School of Engineering and Applied Sciences, built the silicon electronics to record electrical activity from large populations of neurons. At some point, the team started wondering whether the same current-control technology could be used to drive chemistry instead of biology. It worked.
Each of the 64 synthesis sites on the chip contains two concentric ring electrodes surrounding DNA anchored at the center. When a site is activated, the inner electrode generates protons that lower the local pH, creating the acidic conditions that allow a water-based enzyme to add a nucleotide to the growing DNA strand. The outer electrode simultaneously pulls protons away from neighboring sites, confining the reaction to that specific location. Cycling through this process, the chip builds 64 separate sequences independently and in parallel.
The team also demonstrated a separate application: using the 64 synthesized sequences to encode a 169-byte text message. DNA-based data storage is a growing field, and the chip's precision could support that direction alongside its medical applications.
Where the limitation actually sits
To find out how far the platform could scale, the team tried placing synthesis sites closer together to increase the number of simultaneous sequences. That experiment failed. But the failure was informative. The chip itself was accurately confining the low-pH zone to each intended location. The problem was the chemistry used during deprotection.
DNA synthesis requires removing a temporary blocking group after each nucleotide is added before the next one can attach. Low pH triggers this removal indirectly, by generating intermediate molecules that do the actual work. Those intermediates drift, and at closer site spacing they crossed into neighboring synthesis locations, blurring the boundary between reactions.
"The chip did what we asked it to do: it localized low pH at selected sites," said Han Sae Jung, co-first author of the study. "The limitation came from the deprotection chemistry, not from the silicon." That finding points directly to the next step for the field: developing a more direct, localized acid-driven deprotection chemistry that can keep pace with what the chip's electronics can already do.
What the timeline looks like
The current platform cannot yet support sequences long enough for most gene therapy or vaccine applications, and clinical testing has not begun. The researchers and independent analysts place the timeline to medical benefit in years, not months. What the Harvard chip provides right now is a clear proof of concept and a roadmap: enzymatic synthesis in water is viable at meaningful scale, and the bottleneck is chemistry, not electronics.
The work was a collaboration between Harvard, the Broad Institute, DNA Script, and POSTECH. The paper is titled "Parallel enzymatic DNA synthesis using a semiconductor chip" and was published in Nature Electronics with DOI 10.1038/s41928-026-01662-9. Harvard's Office of Technology Development has filed intellectual property related to the platform.