Maynooth DNA Computer Runs 100-Bit Calculations
The Maynooth DNA computer runs 100-bit calculations by letting programmable strands assemble toward a lower-energy molecular state. The peer-reviewed system completed addition, multiplication and division in water, showing that a reusable computer can operate without the continuous electrical power required by silicon logic.
Researchers at Maynooth University reported the scaffolded DNA computer in Nature on September 16. The work is an experimental computing platform, not a replacement for processors: its fastest small arithmetic test took seconds, while a larger addition problem required hours.
How the Maynooth DNA Computer Works
The device consists of a long DNA scaffold and many shorter DNA strands mixed in salt water. Heating separates molecular structures; controlled cooling then allows complementary sequences to bind. The final arrangement of strands encodes the answer to the programmed calculation.
That process differs fundamentally from transistor logic. Silicon computers repeatedly consume electricity to switch billions of gates and maintain data flows. Maynooth's molecular system instead moves toward a thermodynamically favored structure, using the chemistry of DNA binding to perform the calculation after the mixture is prepared.
The researchers can change the program by selecting which short strands enter the mixture. A single scaffold architecture therefore supports different operations rather than hard-wiring one answer into one batch of molecules.
The published experiments demonstrated four main capabilities:
- Addition across small and multi-million-number tests
- Multiplication by three
- Division by two
- Eight-bit parity detection
Ten Programs and More Than 700 Computations
The team tested ten programs, including 100-bit computations, and completed more than 700 individual calculations across its experiments. Adding 10 and 3 produced a result in about 30 seconds. An addition involving numbers from roughly 11 million to 34 million took as long as 14 hours.
Those timings reveal both the achievement and the limitation. Molecular computers remain much slower than modern chips for ordinary arithmetic, but the Maynooth design is unusually fast and programmable within DNA computing. Its value lies in proving that complex molecular interactions can be organized into reliable, repeatable computation.
The system also completed as many as 25 calculations in sequence. In a separate durability test, researchers partially dried a sample, stored it for about 18 months and restored its operation by adding water. That resilience could matter for computing systems designed to coexist with molecular data storage.
The Energy Claim Needs a Careful Reading
Maynooth describes the computer as operating without a continuous electricity supply because the calculation is driven by molecular self-assembly. That does not mean the entire experiment consumes no energy. Preparing strands, heating and cooling the mixture, and reading the output still require laboratory equipment.
The distinction is important for assessing the technology. The work does not show a drop-in, zero-power alternative to a data center processor. It demonstrates a different physical model in which the computational step can be powered by a system's movement toward chemical equilibrium.
Researchers also supplied the sequence-design code, thermodynamic analysis, raw quantitative-PCR data and scale-up modeling through an open Zenodo archive. That record gives other laboratories a path to inspect the methods and attempt independent reproduction.
Where Molecular Computing Could Fit
DNA is attractive as a computing material because enormous numbers of strands can interact in a tiny volume, while the same medium can preserve information for long periods. A practical system could eventually combine dense archival storage with calculations performed near the stored data.
The Maynooth team also points to longer-term possibilities in smart materials and biological environments. Molecular computers might someday respond to conditions inside cells, where conventional electronics cannot easily operate, enabling tightly targeted sensing or disease detection. Those applications remain research goals rather than demonstrated products.
Scaling is the immediate scientific challenge. Larger programs require more carefully designed strands, reliable reaction kinetics and a dependable way to read results. The Nature paper shows that 100-bit computation is possible in this architecture; it does not yet establish manufacturing economics, general-purpose programmability or performance at industrial scale.
The project was led by Damien Woods with Tristan Stérin, Abeer Eshra, Constantine Evans and Janet Adio. It is supported in part by the European Innovation Council's DISCO project, which is exploring DNA infrastructure for storage and computation.
A New Route Beyond Silicon Logic
The strongest result is architectural rather than competitive. By making the answer the energetically preferred molecular structure, the researchers created a computer that can be reprogrammed, reused and stored without constantly driving every logical transition with electricity.
Silicon will remain vastly superior for phones, servers and AI accelerators. The Maynooth DNA computer instead broadens the definition of a programmable machine—and offers a credible experimental foundation for computation in places where molecules, not chips, are the native hardware.
