Use university support to test a problem
As checked on 10 October 2026, HKU's official DeepTech100+ page describes a one-year joint HKU and HKSTP incubation programme and six-week market-validation training. It lists the Cohort 8 application deadline as 29 November 2026 at 23:59 and limits pitch decks to 15 slides.
These details suggest useful preparation, not eligibility for every external team or automatic investment after participation. Confirm applicant affiliation, intellectual property, company and subsequent programme requirements against the full official guidance. This article is about getting more from market validation, not an application eligibility checklist.
Separate a scientific result from a reason to buy
Improved accuracy, speed or energy consumption can support a research finding. A customer's willingness to pay also depends on workflow, budget ownership, switching effort and alternatives. Combining these two kinds of evidence encourages an unsupported leap from better technology to inevitable demand.
The following table is our editorial framework, not HKU's assessment criteria.
| Team's claim | Evidence that may already exist | Question still to test |
|---|---|---|
| We measure faster | Controlled comparison under specified conditions | Is measurement the customer's delay, or is the bottleneck elsewhere? |
| We reduce manual work | Laboratory task-automation records | Do installation, review and maintenance add new labour? |
| We serve many industries | Several technically plausible applications | Which accessible user has the most pressing problem? |
| Customers are interested | Meetings, messages and demonstration feedback | What resources will they commit to further evaluation? |
Exploratory interest is not sales revenue. Nor should research records be changed to suit a commercial story. Scientific and commercial evidence should complement each other, not substitute for each other.
Interview three different roles
Imagine a fictional industrial sensing team with documented laboratory performance but an unsettled commercial application. This is not a report about an HKU, HKSTP or Jimhang Capital project.
The team could speak separately with an operator, a process owner and a budget owner. Ask the operator how the last exception was handled. Ask the process owner about downtime, review and handover effort. Ask the budget owner what would justify procurement and how decisions are made. Do not ask everyone only whether they like the product.
Keep the question, participant's role, date and intended meaning in interview records. Distinguish direct observations from the team's interpretations. Do not publish names, companies or site photographs without permission. AI may help organise material authorised for that processing, but must not turn a polite response into a purchasing commitment.
Run small experiments that can change the plan
One possible sequence is to test whether the problem recurs, observe the existing workflow with minimal disruption, and then discuss a bounded evaluation. This is an editorial suggestion, not the DeepTech100+ curriculum or a sequence applicants must complete before applying.
Before each round, write down which result would change your judgment. If the real bottleneck is not measurement speed, demonstrating a faster sensor again does not answer the commercial question. If only one unusually helpful contact recognises the need, there is not yet evidence that an entire industry will adopt the product.
Keep rejected hypotheses. Ending work on a poor-fit application does not make the research worthless; it may preserve time for a better use case. A pitch deck should not contain only the interview fragments that support the original plan.
Connect commercial support with a separate rights review
HKU's research-startup guide distinguishes commercialisation assessment, IP protection, venture building and licensing, and describes mentoring and startup-clinic resources. That division matters: access to business advice does not establish technology-use rights or complete approval for external activities.
Where university technology is involved, ask the relevant university office what the company can use, within what scope, and which steps require separate approval. This article does not determine ownership of any technology. Our analysis of academic-founder role boundaries addresses a different question from market validation and can be read alongside it.
Make the Greater Bay Area connection specific
Before seeking hardware engineering or manufacturing support, define the immediate question. Do you need a manufacturability discussion, a sample, or an engineering evaluation with a stated measurement objective? Without that scope, mentors, factories and investors may be responding to different problems.
Jimhang Capital, founded by Justin Zhan, publicly describes a role in connecting founders with Greater Bay Area supply-chain resources. Teams can send a non-confidential description of the application, market hypothesis and most important unresolved question to justin@jimscapital.cn to discuss possible resource connections. Any support is scoped separately.
This is not an HKU- or HKSTP-authorised application channel. It does not imply an unverified partnership, assessment role or investment relationship. The useful starting point is a shared, unresolved problem around which university research, industrial engineering and capital discussions can be organised.
Editorial analysis as of 10 October 2026, not official HKU guidance, a personal interview or investment advice. Application dates may change; IP rights, university approvals and investment arrangements require separate verification.