The research and analysis market has a measurement problem, and it is worth being precise about what the problem is. Buyers cannot distinguish a claim that was verified from a claim that was repeated. Both arrive in the same font, in the same document, with the same confidence. Since the buyer cannot tell them apart, the seller has no commercial reason to spend money on the difference. The result is an industry where rigour is expensive to produce and invisible to price, therefore under-supplied.
Serinyx built Quorum to make rigour visible, and today we published three reports to demonstrate it. This note is written for investors, and it argues that the reports are worth reading for what they reveal about the machine rather than for their subject matter.
What was published
The Quorum Showcase Series opens with three reports on general industry questions. The first, The Demo Economy, examines the gap between announced humanoid-robot adoption and the audited record. The second, Europe's Most Valuable Mine Is a Scrapheap, follows the industrial economics of battery recycling under European recycled-content law. The third, Hydrogen's Dark Horse Is Turquoise, models methane pyrolysis against its competing production routes.
Three subjects, three unrelated industries, three different kinds of evidence. That was the point of the selection. A pipeline that only works on one shape of problem is a consultancy with a good analyst on staff.
The four things we would look at
It is a pipeline, not a project
The three reports were produced through the same process, in parallel, from three separate research corpora, and each carries the same apparatus. Each states the size of the evidence base it summarises. Each carries a numbered reference list in which every link was fetched and verified before publication. Each records what it could not establish alongside what it could. Each is accompanied by a model of its subject in which every input is bounded by published evidence.
The reason this matters commercially is that the marginal report is cheap and the first one was not. What we built is reusable across subject matter, and the three pilots were the test of that claim rather than an assertion of it.
The compression ratio is the unit economics
Each report is a summary of something much larger, and the ratios are stated on the closing page of each.
The Demo Economy runs to roughly 7,000 words. It summarises one part of a study whose written chapters alone run to approximately 256,000 words, over a working corpus of about 50.1 million words in 29,111 files. Before any chapter was written, that study asked 263 atomic research questions, collected 3,027 separate model reports against them, and ran 4,250 audits of those reports. The published report is about one word in seven thousand of that corpus.
Europe's Most Valuable Mine Is a Scrapheap summarises one subtopic in a research programme of 12 fields and 72 subtopics. The battery-recycling subtree alone holds approximately 913,000 words across 285 documents and datasets, which makes the report about one word in a hundred and thirty of the research behind it.
The third draws on approximately 2.08 million words across 1,051 documents and 16 research domains, which is about one word in two hundred and seventy.
We would read those ratios as a statement about capacity. The pipeline synthesises evidence at a volume that no human research team can review, and it does so while keeping every published figure traceable to a source. Traceability at that volume is the hard part, and it is what the three pilots were built to demonstrate.
The defensibility is in the discipline, not the models
An obvious objection is that anyone can point a large language model at a subject and generate a report. That objection is correct, and it is why the interesting engineering sits elsewhere.
Quorum collects across many models and sources, read cold and without shared context, then adjudicates their disagreements rather than averaging them. Averaging is what produces confident nonsense, since it hides the fact that the sources disagreed. Adjudication requires a rule for what counts as corroboration, and The Demo Economy is a demonstration of that rule doing real work. The most-cited Western deployment metrics in humanoid robotics appear twelve times across the reviewed record. They collapse to a single origin, which is the vendor's own pilot dataset. A system that counts appearances would call that consensus. A system that counts independent origins calls it one source quoted twelve times.
The same discipline produces the harder outputs. The Demo Economy finds 292 disclosed award projects for 2025, one fully documented acceptance chain, no verified payment remittances, and no verified repeat purchases anywhere. It then declines to over-claim, since publishing those deeper stages is not generally mandated for ordinary goods procurement in China, therefore the zeros are ceilings on what can be known rather than statements about what happened. Refusing an available headline is not a natural behaviour for a research product, and it is the behaviour that makes the output worth paying for.
The accompanying models are the audit made usable
Each report is built on a working model of its subject. Every input in those models is bounded by the extremes of published estimates rather than by round numbers chosen to make a demonstration look good, and the model states where its own conclusion would fail.
The hydrogen report is the clearest example. Methane pyrolysis costs 5.24 euros a kilogram before any carbon revenue, 2.44 euros with the solid carbon sold at 700 euros a tonne, and 5.39 euros if that carbon has to be buried. Separately from cost, a methane leak rate above 0.63 per cent invalidates the climate case, and both published figures fail that threshold. A configuration can therefore be the cheapest option in the model and still not qualify. A model that can produce that answer is a model that can lose, which is the property that makes the rest of it credible.
What the showcase does not show
We want to be direct, since a demonstration invites the wrong conclusion about scope.
These reports answer questions we selected, at the depth a general reader will follow in twenty minutes. They cannot use client data or anything under confidentiality, and sources that cannot be publicly linked are listed without a URL and say so. They compress a body of research into a fixed length, therefore most of the evidence and nearly all of the working analysis stays behind the summary. The accompanying models expose a few dozen assumptions where the underlying studies carry many more.
None of that is a limit of the pipeline. It is the shape of a twenty-minute document, chosen so that a reader outside the industry can follow the argument end to end.
Where this goes
The commissioned engagement is the product. A client brings the question and agrees the brief. What they receive is the study underneath a report like these, at the scale the question requires, with their own data inside the evidence base. The audit trail is built to survive a board or an investment committee. The deliverable follows the decision rather than a template. It might be a diligence pack, a regulatory analysis, a pricing model, a patent landscape, an index methodology, or a live model the client's team keeps using after the engagement ends.
The three reports are free to read in full, and we would rather be assessed on them than on a description of our method.