Youdu Capital · Primary-Market Financing
In an update about portfolio company BioGeometry published on 2026-10-28, LinkX Capital (星连资本) described a new financing round. The article listed the construction of an automated, high-throughput laboratory among the uses of the financing. This cites a public statement of intended spending; it does not establish that construction is complete or assess the effectiveness of any drug, model or research process.
In its investment announcement about Warp dated 2026-10-22, Lightspeed said that Warp records every step of an experimental run, including failed steps. Published more than seven days ago, the material is used as background within the fourteen-day window and only relays the investor's description. Overseas information is background only. Youdu Capital, acting as an FA, serves only domestic financing for Chinese technology companies and provides no cross-border financial services.
As automated experiments and data feedback attract attention, founders should ask another question: did a new result help the team select a solution, or did it provide a new test using information that played no part in that selection? The following is general management analysis. Its hypothetical scenarios make no allegation about the companies above and do not imply first-hand testing of their systems. The issue is whether the same material has been assigned two conflicting roles, rather than how many experiments were run.
The first causal step occurs while a solution is being developed. Suppose a team examines results from a set of samples, uses them to adjust a model, process parameters or candidate solutions, and then retains the best-performing version. Those results have already influenced the selection. Even if the team never fed them directly into a training program, it has used their information. Moving the files to a different directory cannot remove that influence; manual selection is also part of developing a solution.
The second causal step appears when progress is reported. If the final solution is demonstrated using the same material, its apparent performance may partly reflect adaptation to characteristics of already-known samples. That does not necessarily add reliable evidence about new situations. If the founder reports independent validation as complete, subsequent resource planning may skip the retesting that should have been retained, carrying unanswered questions into delivery preparations or financing commitments.
This does not make exploratory data worthless. Repeated experiments, comparisons and revisions are normal research activities. The distinction is between helping to identify a candidate solution and testing that candidate using material that did not participate in selection. The former can be recorded as exploratory progress; better-looking results alone do not turn it into the latter. Nor does using a different presenter automatically establish the latter.
Management records can begin by documenting how material was used: who saw the results, at which stage, which decisions changed because of them, and whether the current solution was fixed before or after those results were examined. Use existing records and mark uncertainties for checking. The concern is whether information entered the decision process. It does not require obtaining a partner's unauthorized data or revealing trade secrets to every prospective capital provider.
Before retesting, technical personnel should define the question, solution version and evaluation criteria, and then arrange material that did not participate in the earlier selection. The specific design requires relevant professional expertise; there is no universal sample proportion for every company. If retest results are subsequently used to revise the solution, record that their role has changed. The same results cannot continue to be counted as independent validation.
Different material does not automatically mean independent information. Copying the same original records into two files, or assigning new identifiers to samples already examined, does not make the team forget known results. Conversely, independent retesting does not require buying equipment from scratch every time. What needs checking is which information participated in selection, so that evidence is not counted repeatedly according to the number of files.
Financing communications can present completed selection work, planned retesting and unanswered questions together. The next stage of spending then corresponds to a specific kind of new evidence, rather than another presentation of the best existing result. If retesting is not yet feasible, retain its unvalidated status and explain the resources and dependencies required. Do not report an anticipated outcome as a capability already established.
Over the next three to six months, a company preparing to move a research result into larger-scale delivery could select one active solution and review how information was used to develop it. If all current performance results come from samples repeatedly examined, discuss independent retesting with the technical lead before updating financing milestones. If retesting already meets relevant professional requirements, preserve its process records without creating unnecessary duplicate work.
If performance on new material differs from the exploratory results, first check the consistency of records, conditions and the solution, and have qualified professionals assess the next steps. A difference should not be automatically interpreted as technical failure. Nor should evaluation criteria be changed repeatedly to preserve an earlier result. The observation to make in this scenario is whether the team can explain what new information a result adds, rather than whether it can keep refreshing the same scorecard.
As an FA serving non-listed companies, Youdu can help align research progress with domestic financing materials. It does not replace technical evaluation, an institution's investment judgment or professional certification. This article provides no specific investment, legal or tax advice and promises no financing or research outcome. Accurately distinguishing exploration from retesting first helps a company understand what it has learned and what it still needs to validate.
Sources 星连资本：百奥几何被投动态 (2026-10-28) https://ad.36kr.com/p/4045277817411209 Lightspeed: Why We’re Backing Warp (2026-10-22) https://lsvp.com/stories/why-were-backing-warp/
Content note: This article was drafted with AI assistance and reviewed through an independent AI process, using the listed public sources. It contains no original interviews, and no first-hand testing of the products, services or methods discussed was conducted. Checking public sources is not a field investigation or first-hand experience.