Quentir Medicine Monitor
Evidence-based insights for quantum medicine. Published by Quentir Systems LLC · November 5, 2026.
On the afternoon of 28 October 2026, representatives of fifteen pharmaceutical companies from Shaanxi province toured a quantum computing center in Xi'an that is built for drug research. The host, QureGenAI (医图生科 in Chinese), used the visit to present its case for quantum AI drug design to companies that make peptides, traditional Chinese medicines and generic drugs.
The Xi'an Advanced Manufacturing Association organized the visit, and the Chang'an Xiandao Life Sciences Industry Innovation Center co-hosted it. According to the event report published on ifeng.com on 5 November 2026, the companies attending included the Shaanxi Pharmaceutical Research Institute, Huikang Biotech, Jinhua Pharmaceutical and Lijun Pharmaceutical. Officials from the Xi'an industry and information technology bureau also attended.
The center itself is a joint project of the Chang'an Xiandao innovation center and Xi'an Xiandao Yitu Quantum Technology, a company linked to QureGenAI. The report describes it as housing the first 15-qubit superconducting quantum computer in northwest China, coupled to classical computers in a hybrid cluster for biomedical work. The company also calls its offering the world's first end-to-end quantum computing solution aimed at industrial drug design and discovery. That claim comes from the company and the report repeats it without any comparison to other platforms.
What QureGenAI's CEO Li Xiaoran told the Shaanxi companies on 28 October
Li Xiaoran, chief executive of QureGenAI and chief scientist of the center, listed four areas where the company applies its methods: design of new drugs, protein structure prediction, optimization of fermentation processes, and activity screening of natural products. He cited two examples from cancer research, a prodrug design for pancreatic cancer and drug development against the KRAS target, a protein that is mutated in many pancreatic, colorectal and lung tumors.
Li argued that quantum computing has an advantage that classical methods cannot match in two specific steps of chemistry: calculations on strongly correlated electronic systems, and searches for the transition states of chemical reactions. Both steps are hard for standard computational chemistry, so the choice of target is sensible. The report gives no data for either example, however. It names no molecule, no accuracy figure and no comparison with a classical calculation on the same problem.
The same applies to the largest number in the report. Li said the team had applied combined quantum and AI methods to traditional medicine design, enzyme engineering and aqueous two-phase extraction, and that peptide batch production had reached an efficiency of several dozen times that of traditional processes. No study, patent or customer is cited for that figure, so a reader cannot yet check how it was measured or how much of it came from the quantum part of the workflow.
Quantum pillar: computing. Technology readiness: TRL 3 of 9. This is Quentir's provisional reading of the event report: the company describes drug-design work on a small real quantum computer, and the report cites no study, data or outside validation for it.
What the drug makers asked for, and what the Xi'an association requested
The companies' own requests were concrete. Huikang Biotech described its AI virtual screening of a physical library of 175,000 peptides and asked for better ways to optimize peptide structures. Jinhua Pharmaceutical asked for computational support in optimizing classical formulas of traditional medicine and in further development of an artificial tiger bone powder. The Shaanxi Pharmaceutical Research Institute listed target discovery, molecular design, activity screening and prediction of drug-likeness.
The most useful remark came from the association side. A representative said the platform should state clearly which large models, datasets and AI agents it uses, so that companies can judge it with more confidence. A hospital pharmacy or a drug company considering such a service needs exactly that information before any pilot. Zhao Yuxiao, who heads the consumer goods division of the Xi'an industry bureau, proposed a list matching company needs to available quantum AI services, as part of the city's plan for its biomedical industry.
What the open TyxonQ code and the 2022 resource estimates by Blunt and colleagues show
One part of the company's work can be checked today. QureGenAI publishes TyxonQ, its quantum programming framework, on GitHub under the Apache 2.0 license. The repository describes a quantum chemistry layer for drug design with an interface modeled on the PySCF chemistry package, and the repository describes running the same code on a simulator or on real hardware. Anyone can inspect this code, and the repository describes the simulator and hardware interfaces. Evidence of a better drug candidate would have to come from separate published studies.
The size of the machine matters for any chemistry claim. In a 2022 perspective in the Journal of Chemical Theory and Computation, Nick Blunt and colleagues estimated the quantum resources needed to simulate part of a protein bound to the cancer drug ibrutinib. They concluded that calculations of that kind, with active spaces of around 50 orbitals and electrons, require an error-corrected quantum computer. Their estimate covers that class of full quantum chemistry calculation, and hybrid methods that use a quantum computer for smaller pieces are a separate case. A 15-qubit machine without error correction sits far below the level they described, so its near-term value for chemistry of that kind lies in testing workflows and small model systems.
Xi'an is one of several sites in Asia now pairing quantum hardware with drug programs. Yonsei University switched on a 120-qubit IBM processor in Songdo on 4 November with its own time targets for target protein search, and a Korean grant program funded quantum drug design that plans to synthesize the molecules it designs. Synthesis and testing give a reader the strongest check on any computational claim.
What a hospital or drug company should ask QureGenAI before a pilot
A buyer should ask for three things. The first is one worked example with the molecule disclosed, the quantum part of the calculation separated from the AI part, and the same problem solved with a strong classical method. The second is the measurement behind the peptide production figure. The third is the list of models and datasets that the Xi'an association itself requested. The companies at the meeting said they want joint development and shared platforms with the center, and published results from such projects would move the work to a higher readiness level.
Sources
Primary source: the Huaxia Zaobao event report on ifeng.com of 5 November 2026 on the 28 October 2026 visit of 15 Shaanxi drug makers to QureGenAI in Xi'an. Also drawn on: QureGenAI's TyxonQ repository on GitHub, the 2022 perspective by Nick S. Blunt, Joan Camps, Ophelia Crawford, Róbert Izsák and colleagues, and earlier Quentir coverage.