A public map for the next conversation
These nine starting nodes connect public evidence with a practical verification question. They are not an exhaustive map or a ranking of districts. The proposed uses and experiments are editorial hypotheses; a source establishes only the policy, institution or development it actually describes.
Need → scarce input → functional node → counterpart → bounded experiment. Use this map after comparing China with the best alternative. The outcome can be a partnership, a temporary project, a small node or no China operation.
Haidian
Use when the scarce input is: Robot evaluation, research collaboration and engineering iteration.
Public evidence: A September 1 report records the robotics accelerator launch and activation of evaluation, shared processing and application-scenario platforms.
Evidence status: Launched platforms; project access to verify.
Counterpart to identify: Robotics park / accelerator operator and the relevant research or engineering team.
Smallest useful test: Agree one evaluation or processing task. Confirm technical scope, scheduling, fees, results ownership and access for the actual team.
E-Town / Yizhuang
Use when the scarce input is: Embodied robotics, application partners and industrialization.
Public evidence: The published robotics measures cover real-world data, secondary development, public testing and pilot services, component supply and manufacturing. Policy support is distinct from an available service booking.
Evidence status: Published policy; service and project eligibility to verify.
Counterpart to identify: Robotics and intelligent manufacturing bureau, platform operator or industrial application partner.
Smallest useful test: Identify one task and a willing scene owner, then confirm the robot, integration work, test evidence and commercial terms needed to run it.
Changping
Use when the scarce input is: Life sciences, neurotechnology and medical-device translation.
Public evidence: Changping has published dedicated brain-science and brain-computer-interface measures alongside its medical-health policy framework. These provide a sector-specific starting point for counterpart research.
Evidence status: Published measures; clinical and facility access are project-specific.
Counterpart to identify: Life-science park or platform, research group and any relevant clinical institution.
Smallest useful test: Define whether the next step is research, preclinical validation or clinical work. Seek the appropriate partner and protocol; an industry cluster is not permission to conduct a study.
Huairou
Use when the scarce input is: Scientific instruments, facility-linked research and proof of concept.
Public evidence: The August 30 Shineng Peninsula announcement describes a major-facility results-translation cluster. Its construction and intake programme should be distinguished from the availability of each laboratory or instrument.
Evidence status: Announced translation carrier / build-out; access by facility.
Counterpart to identify: Scientific-facility user office, instrument team or translation-carrier operator.
Smallest useful test: Specify the measurement, sample, instrument and output needed. Verify the facility schedule and collaboration terms before designing a recurring research node.
Pinggu
Use when the scarce input is: Agricultural validation, food technology and synthetic biology.
Public evidence: Official reporting on Pinggu’s 2026 development links agricultural innovation with food and nutrition industry development. Agricultural Zhongguancun is a starting geography for investigating the relevant research, production and field counterpart.
Evidence status: Reported industry development; individual host readiness to verify.
Counterpart to identify: Agricultural research team, food-industry platform or a suitable farm / field operator.
Smallest useful test: For agtech, agree a crop, season, field protocol and data output. For food or biological processes, verify the lab or pilot specifications separately from ordinary office availability.
Yanqing
Use when the scarce input is: UAV development and low-altitude testing.
Public evidence: The May 28 official announcement reports structural completion of the Zhongguancun Yanqing UAV industrial park. It is evidence of a developing carrier, not confirmation that every facility is operational or open to outside teams.
Evidence status: Construction milestone; commissioning and test access to verify.
Counterpart to identify: UAV park operator and the responsible flight-service / test-site interface.
Smallest useful test: Confirm the aircraft, test objective, site, flight permissions and available dates. Distinguish existing test services from space still being delivered.
Mentougou
Use when the scarce input is: Model adaptation and AI deployment support.
Public evidence: Jingxi Zhigu’s large-model ecosystem service station provides an identifiable institutional interface for AI deployment questions. Evaluate the service and the actual computing environment separately.
Evidence status: Reported service station; workload suitability and terms to verify.
Counterpart to identify: Jingxi Zhigu service station, compute provider or application partner.
Smallest useful test: Test the actual workload, hardware compatibility, service level, total cost and data arrangements before treating compute availability as an operating advantage.
Fangshan
Use when the scarce input is: Energy and materials scale-up.
Public evidence: A June 10 district announcement identifies storage-cell, hydrogen-engine and high-performance membrane projects in Beijing’s initial pilot-platform cultivation list. Selection for cultivation does not establish service availability for a new project.
Evidence status: Named pilot-platform pipeline; readiness to verify.
Counterpart to identify: Relevant pilot-platform operator, industrial company or university technology-transfer team.
Smallest useful test: Match the process and scale to one platform. Confirm equipment, sample acceptance, throughput, cost and the path from a successful pilot to production.
Miyun
Use when the scarce input is: Measurement, control and specialist equipment.
Public evidence: Official reporting identifies the Zhongguancun Miyun measurement-and-control equipment industry cluster. It supports a supplier and engineering search, rather than a claim that a particular testing service is already available.
Evidence status: Recognized cluster; supplier and service capability to verify.
Counterpart to identify: Miyun park service team, equipment company or testing provider.
Smallest useful test: Use an actual specification to check calibration, testing or component fit with a named supplier. Evaluate environmental field access separately; a nearby application context is not an open test site.
One company can use more than one ecosystem
A robotics team might investigate research and evaluation in Haidian, an industrial application partner in E-Town, and prototype suppliers in Shenzhen or Dongguan. A scientific-instrument team might investigate a Huairou research user and a separate equipment supplier. These are candidate architectures, not client case studies.
Compare the full loop: prototype → test → evidence → next revision. Add another node only when its scarce input outweighs the coordination cost. Do not choose a district merely because it carries the same industry label as your company.
Turn an interesting node into an executable experiment
Record the owner, scope, budget, deadline, output, success threshold and exit condition. Establish host willingness, technical fit, service terms and the permissions needed for the actual work. Public programme language and direct agreement answer different questions.
For an invited technical project, reply to Leo’s original email with a non-confidential milestone, bottleneck and existing evidence. Initial qualification follows the project-development route described in How We Work; a Fit Check or Sprint purchase is not required.
Where the other Beijing guides fit
AI / OPC Policy Intelligence is the specialist policy branch. The District Decision Desk compares outer-district operating environments. The Village Atlas records the conditional village-compatible branch. A field station or research retreat still needs a suitable host and site-specific verification; a lifestyle listing alone does not establish technical suitability.