The Pipeline · Issue One, October 2026
The Pipeline Is Narrower Than the Catalogue
77 compound names, one registry, one day. 59 return at least one registered study, and ten of them account for 73.3% of every record returned. A catalogue and a pipeline are not the same list.
Read with Feature II, which shows what the federal publication record is actually about, and with Feature IV, which sets these counts against measured attention.
The shape of the answer
On 2026-10-13 the public clinical trials registry was asked the same question 77 times: how many registered studies name this compound as an intervention? The question is narrow on purpose. It does not ask whether a study finished, what it reported, who funded it, or what the compound does. It asks how many records exist under a name, which is the one thing a registry can answer without anyone interpreting anything.
The answers come to 10,992 records. 59 of the 77 names return at least one. The remaining 18 return none at all.
Then the distribution does something the total hides. The ten highest names account for 8,058 of those 10,992 records, 73.3% of everything the registry returned for the whole list. The lower half of the same list, thirty-eight names, accounts for 102 records between them, which is 0.9%.
A list carrying 77 names is not carrying 77 development programmes.
Concentration is ordinary in drug development and nobody should be surprised by a long tail. What is worth saying plainly is what a tail of that shape does to a catalogue. A list carrying 77 names is not carrying 77 development programmes. On the registry’s own evidence it is carrying about ten, and then a very long sequence of names with one record, two records, or nothing.
The names the registry has never seen
18 names return zero registered studies: AOD9604, Capromorelin, Cortagen, Cortexin, Epitalon, FOXO4-DRI, GHRP-2, GHRP-6, Hexarelin, IGF-1 DES, IGF-1 LR3, KPV, Livagen, Noopept, Pinealon, Semax, Thymalin, Thymulin.
Several of those are among the most heavily traded names in the research-chemical channel. Two of them carry substantial measured search demand while returning nothing here at all: KPV at 17,460 monthly searches and Epitalon at 12,220. Feature IV takes that gap apart properly.
A zero here is the absence of a registration, nothing more and nothing less.
What a zero means has to be stated carefully, because it is easy to overread. It records that no study registration exists under that exact name in this registry on this date. It does not record that no research exists. This project’s own corpus indexes 902 published study rows across 77 compounds, and several of the 18 carry published records in it. Laboratory work, animal work and published human observation can all happen without a trial registration, and a great deal of it does. A zero here is the absence of a registration, nothing more and nothing less.
The commercial point survives the caveat. A registration is the artefact an acquirer, a lender or an insurer can actually examine. Its absence is not evidence of nothing happening. It is the absence of the particular document those parties ask for.
What a count is not
Four limits apply to every figure above, and they matter enough to put before the argument rather than after it.
A record is a registration, not a trial that ran. The registry holds withdrawn, terminated, completed and not-yet-started studies in one count. The narrower measure is current recruitment, and on the same date 1,245 records across 48 names were recruiting.
A name that is also an endogenous human substance matches studies in which it is measured rather than developed. Oxytocin returns 932 records, and a large share of those are physiological studies in which the substance is a probe or an outcome, not a product in development. The same caution applies to ghrelin, glutathione and somatostatin. The top of this table is not uniformly a pipeline.
One entry is a class term rather than a product. GLP-1 returns 2,155 records, the largest figure in the set, because the corpus carries GLP-1 as an entry in its own right. It is reported here as the corpus holds it and it should be read as a class signal.
The query matches a name. Spelling variants, code names and sponsor-internal designations that differ from the corpus name are not captured, which will understate some counts. Every query URL is published row by row in The Pipeline so that any individual figure can be re-run rather than taken on trust.
The desk’s reading
What follows is this desk’s reading of those counts. It is an argument, not a finding in the data, and it is separated from the data for that reason.
Any statement that begins “the peptide sector is” has averaged two populations that have nothing in common except the word peptide, and the average describes neither.
The first consequence is that sector-level diligence in this category is close to worthless. The spread between the top and the bottom of one 77-name list is three orders of magnitude. Any statement that begins “the peptide sector is” has averaged two populations that have nothing in common except the word peptide, and the average describes neither.
The second is that live recruitment, not cumulative registration, is the figure that moves. Of the 1,245 recruiting records, 265 sit under two single-product names, semaglutide at 168 and tirzepatide at 97. A reader watching the category for change is watching a very short list.
The third is the one this issue keeps returning to. The names with no registration are not thereby worthless, but they cannot be underwritten on registry evidence, because there is none to examine. Whatever supports them sits somewhere else: in published literature, in a compounding position, or in demand. Feature III looks at what the documentary record will carry, and Feature IV at what demand is actually telling anyone.
| Compound name as queried | Registered records | Of which recruiting |
|---|---|---|
| GLP-1 | 2,155 | 308 |
| Gonadorelin | 1,164 | 118 |
| Oxytocin | 932 | 79 |
| Somatropin | 818 | 45 |
| Semaglutide | 781 | 168 |
| Leuprolide | 521 | 57 |
| Liraglutide | 514 | 25 |
| Vasopressin | 433 | 41 |
| Exenatide | 378 | 12 |
| Somatostatin | 362 | 56 |
| Octreotide | 327 | 29 |
| Tirzepatide | 298 | 97 |
Sources, and the queries that produced every figure above
- ClinicalTrials.gov Data API, version 2, studies endpoint. U.S. National Library of Medicine, National Institutes of Health. Retrieved 2026-10-13. One query per corpus compound name. totalCount is the number of study records the registry returns for the term on the retrieval date. The full result set, one row per compound with the URL that produced it, is published in data/quarterly/clinicaltrials.json.https://clinicaltrials.gov/api/v2/studies?query.intr=<compound name>&countTotal=true&pageSize=1
- Peptide research corpus: 77 compound records and 77 evidence profiles. This project’s validated dataset, data/compounds and data/evidence. Counts are taken directly from the dataset files. A derived tier on an evidence profile is labelled INFERRED in the data and carries its own basis string, which is reproduced wherever a tier is printed.
- Google Ads search-volume instrument, United States, held as data/keywords/volumes-us.json. Measured for this project on 2026-10-12 and stored with its own method note. Retrieved 2026-10-12. One of three separate demand instruments in the dataset. The dataset’s own note states that the three must never share a denominator, so only this one is used here and it is named at every figure. A null in this instrument is an absence of advertiser data, never a measured zero.