IN Brief:
- The 2026–28 programme contains four actions and nine tasks linking atomic-scale manufacturing with integrated circuits.
- Priorities include atomic-scale etch, atomic layer deposition, ion implantation, process software, sensors, and precision equipment.
- Pilot, measurement, testing, and certification infrastructure is intended to bridge research and industrial qualification.
Beijing Economic-Technological Development Area has launched a 2026–28 action plan for atomic-scale manufacturing, targeting semiconductor process technology, equipment, software, metrology, pilot production, and industrial demonstration. The programme contains four broad actions and nine tasks intended to move research from laboratory development towards repeatable manufacturing.
The district describes atomic-scale manufacturing as precise control at the scale of atoms or atomic building blocks, with the longer-term aim of producing structures whose dimensions and composition approach theoretical material limits. In semiconductor manufacturing, that ambition maps onto processes where adding, removing, modifying, or measuring extremely small quantities of material already determines whether advanced device structures can be fabricated reliably.
Priority technologies include atomic-scale etching, atomic layer deposition, ion implantation, dynamic simulation, and process-design software. The programme also calls for development of ultra-high-vacuum components, high-precision sensors, thin-film deposition equipment, ultra-precise material removal, nanoscale repair, wafer thinning, and related manufacturing systems.
Several of those techniques are established parts of advanced semiconductor production, but device scaling and three-dimensional integration continue to tighten the tolerances around them. Atomic layer deposition is valued because surface-controlled reactions can build highly conformal films with very fine thickness control, while advanced etch processes have to remove material selectively without damaging neighbouring structures or distorting high-aspect-ratio features.
The difficulty is increasingly one of integration rather than any single process step. Logic, memory, interconnect, and packaging structures can pass through long sequences of deposition, etch, implantation, cleaning, annealing, lithography, and metrology. A small deviation introduced early in the flow may only become visible as a yield problem much later, which makes process control and feedback as important as the nominal resolution of the equipment.
Beijing E-Town’s programme addresses that gap by proposing pilot manufacturing, measurement, testing, and certification platforms alongside core technology research. These facilities are intended to support process verification, equipment testing, and product validation before new tools and techniques move into commercial fabs, while initial equipment deployments inside the district are intended to shorten the path from engineering prototype to industrial use.
That pilot stage is critical for semiconductor equipment. A tool can produce an impressive result on a small number of laboratory wafers and still be unsuitable for volume manufacturing if throughput, contamination, chamber matching, maintenance intervals, uniformity, or repeatability are inadequate. Production qualification is therefore an exercise in equipment reliability and statistical control as much as in achieving one exceptionally fine feature.
The plan also places software alongside the physical process. Dynamic simulation and process-design tools can reduce the number of expensive physical experiments, while higher-precision sensors and metrology can provide the feedback needed to compare modelled behaviour with fabricated structures. At atomic and near-atomic scales, however, measurement uncertainty itself becomes part of the engineering problem because process control cannot be tighter than the ability to characterise the result.
Beijing E-Town says the policy is intended to address fragmented research, insufficient pilot validation, limited application scenarios, and a weak industrial ecosystem around atomic-scale manufacturing. It proposes shared facilities, standards activity, intellectual-property mechanisms, investment support, talent recruitment, and collaboration between companies, universities, research institutes, and users.
The district reports more than 400 integrated-circuit-related companies with an industry scale above RMB130 billion, covering design, wafer manufacturing, packaging and test, equipment, and components. That existing base gives the programme potential equipment developers and production users in the same area, although each new process or tool will still have to pass the qualification requirements of the manufacturer adopting it.
The initiative sits beside Beijing E-Town’s separate AI4Chip programme but addresses a different engineering layer. AI4Chip concentrates on applying artificial intelligence to semiconductor design and manufacturing workflows; the atomic-scale plan focuses on the physical processes, equipment, software, and measurement systems used to manipulate material. The two may meet in simulation and process control, but they are not versions of the same policy.
By 2028, Beijing E-Town wants the district to become a nationally influential centre for atomic-scale manufacturing technology and industrial concentration. The harder benchmark will be found on production lines: whether locally developed etch, deposition, metrology, and process tools can demonstrate repeatability, uptime, defect control, and integration over enough wafers to turn nanometre-scale research into dependable manufacturing.
