SourceX Research · public sources
Public-source briefings on AI data licensing.
Quick answer
Each briefing summarizes published public sources and lists them at the end; none contains SourceX survey or deal data. The one SourceX research item is the open De-identification Benchmark: its method and synthetic test data are published, and it includes a sample report.
AI Data Price Index
Publicly reported prices, linked to their sources. Not SourceX offers or payouts.
Explore the price indexBenchmarks
Explore SourceX benchmarksThe SourceX De-identification Benchmark
How well we protect personal information while keeping business data useful. Includes a sample report.
Research by Noah Loul.20 briefings
- Published October 9, 2026 · Updated October 17, 2026 · Noah LoulA framework for thinking about the value of company data, grounded in public researchA framework for evaluating a company's data based on OECD research on data valuation, for businesses considering licensing their records.
- Published October 9, 2026 · Updated October 17, 2026 · Noah LoulAI content licensing deals, July 2023 to June 2024: a dated selectionA dated, sourced selection of AI content licensing deals announced or reported between July 2023 and June 2024, each linked to the announcement or report that confirms it.
- Published October 9, 2026 · Updated October 17, 2026 · Noah LoulCall recording consent laws: one-party vs all-party states for AI data licensingGeneral information, not legal advice: the federal one-party rule, the states most often listed as all-party, and why the consent rule matters before licensing call recordings for AI use.
- Published October 9, 2026 · Updated October 17, 2026 · Noah LoulCommon terms in public data licensing agreementsWhat public AI data licensing deals and open template licenses, like Creative Commons, Open Data Commons and the Linux Foundation's CDLA, typically cover.
- Published October 9, 2026 · Updated October 17, 2026 · Noah LoulCompany software sprawl: what public SaaS data shows about scattered recordsA review of public SaaS adoption research, such as Zylo's SaaS Management Index, on how many applications a typical company runs, a proxy for how scattered its records are.
- Published October 9, 2026 · Updated October 17, 2026 · Noah LoulConsent and privacy practices for enterprise data licensing: what regulators sayA review of FTC guidance and HHS HIPAA de-identification rules relevant to any business considering licensing its data to an AI developer. General information, not legal advice.
- Published October 9, 2026 · Updated October 17, 2026 · Noah LoulData licensing readiness: what public AI adoption data shows by firm sizeA qualitative look at AI and data readiness using public figures from the Census Bureau's Business Trends and Outlook Survey, which reports AI use by firm size; no single public dataset scores industries on licensing readiness.
- Published October 9, 2026 · Updated October 17, 2026 · Noah LoulEnterprise data readiness: a review of published industry surveysSourceX has no survey of its own. This page reviews published enterprise data readiness research from IBM and Precisely, including figures on the cost of poor data quality.
- Published October 9, 2026 · Updated October 17, 2026 · Noah LoulHow much operational data does a 200-person company generateA transparent, labeled estimate of software-usage and support-ticket data for a 200-employee company, built from public benchmarks.
- Published October 9, 2026 · Updated October 17, 2026 · Noah LoulState of enterprise data licensing: Q4 2026A roundup of publicly reported AI data licensing litigation and government filings through September 2026, with two earlier licensing deals for scale.
- Published October 9, 2026 · Updated October 17, 2026 · Noah LoulThe economics of expert work data for AI agentsWhat public surveys and pay data show about demand for AI agents and the human expertise used to train and check them.
- Published October 9, 2026 · Updated October 17, 2026 · Noah LoulWhat AI developers pay for in data: a review of public researchSourceX has no Acquisition Partner survey of its own. This page reviews public statements, filings and reporting about what AI developers have said they value and paid for in data.
- Published October 9, 2026 · Updated October 17, 2026 · Noah LoulWhat AI developers say they need: public signalsA review of public programs and research showing what kinds of data AI developers say they are seeking, and why.
- Published October 9, 2026 · Updated October 17, 2026 · Noah LoulWhat content owners have earned from AI licensing: public reportsReported AI data licensing revenue for Reddit, Shutterstock and News Corp, drawn from SEC filings, earnings statements and press reporting.
- Published October 9, 2026 · Updated October 17, 2026 · Noah LoulWhat published surveys say about executives and AI dataA review of public executive surveys from PwC, KPMG and the US Census Bureau on AI adoption, investment and data readiness, not a SourceX survey.
- Published October 9, 2026 · Updated October 17, 2026 · Noah LoulWhich business software holds the most AI-relevant recordsA qualitative look at which categories of business software hold records most useful to AI, with public customer-count facts.
- Published October 9, 2026 · Updated October 17, 2026 · Noah LoulWhich business systems companies use: a public data reviewA look at which CRM, support and SaaS systems businesses run, based on public vendor filings, customer counts and adoption research.
- Published October 9, 2026 · Updated October 9, 2026 · Noah LoulGlobal privacy laws and AI training data: a comparisonHow GDPR, UK GDPR, CCPA/CPRA, Brazil's LGPD, Canada's PIPEDA, Japan's APPI, China's PIPL and the EU AI Act treat personal data used to train AI, in plain language.
- Published October 9, 2026 · Updated October 9, 2026 · Noah LoulHow much time data teams spend on data preparation: a review of public surveysA review of published survey data, including Anaconda's State of Data Science report, on how much time data professionals spend preparing data before it can be used.
- Published October 9, 2026 · Updated October 9, 2026 · Noah LoulWhy datasets get rejected: a review of public research on licensing and data quality problemsSourceX has no rejection data of its own. This page reviews public academic research, including the Data Provenance Initiative, on why datasets fail licensing or quality checks.
SourceX's own research
SourceX De-identification Benchmark (SX-TXT-1.0) publishes an open method, synthetic corpus and scorer. It includes a sample report. SourceX’s survey and transaction research programs remain in development; no survey or deal findings have been published.
Which questions would research into the buyer value of licensed business records have to answer?
Five questions, each asking whether a licensed dataset changes what a model can do on a business task, and none answered yet. As the note above says, SourceX has published no findings of its own; this section reports no result. It sets out the questions and how each could be tested, so that any later study can be held to them.
Measuring a dataset by holding everything else fixed is established practice. DataComp-LM (DCLM) has participants filter a shared pool or bring their own data, train "with standardized training code and scale-specific hyperparameters", and score the model on 53 downstream tasks "to judge dataset quality". DataPerf, a community-led benchmark suite for datasets, includes a data-acquisition challenge whose buyer must decide what to purchase while it "can observe no sellers' datasets but some summary information from the sellers", which is roughly where an AI team stands before it licenses records.
The second question borrows that framing. How it works lists ten fit signals SourceX evaluates, from volume and historical depth to whether records show how work ended, and labels the illustration "Not a valuation". Whether ratings on those signals track the gain a dataset produces once it is used is a question a study can answer.
| Question | Comparison | Measure | Result that would count against value |
|---|---|---|---|
| Does the dataset improve the task it was licensed for? | One base model, recipe and compute budget, trained with the dataset or with a matched volume of public or synthetic data | Task score on a test built from later records | No gain beyond the spread between repeated runs |
| Do fit signals rated before delivery predict that gain? | Ratings recorded before delivery against the gain measured afterwards | Rank correlation with an interval | Ratings that fail to order datasets by measured gain |
| Which parts of a record carry the gain? | Full records against copies without outcome fields, linked steps or human feedback | Score change for each removed field group | No loss when outcomes or links are removed |
| Does the gain reach companies the model never saw? | Tests drawn from suppliers held out of training | Held-out score against in-sample score | Gains that disappear on held-out suppliers |
| How many records does a gain need? | Training on fixed fractions of the dataset | Score at each fraction | A flat curve, where more records add nothing |
How would a dataset-utility study be set up so an outsider could check it?
With the plan fixed before any model trains and the report written against that plan. SourceX's methodology page already sets a floor for its own research publications: methodology, sample size and research dates, definitions, limitations and sources, under standards that include "Limitations and conflicts of interest are disclosed" and "Corrections are dated and noted".
Rights draw the outer boundary. For closed deals, section 4 of SourceX's terms describes a license to use the de-identified dataset to "train, validate, test, benchmark, monitor and improve" models operated by SourceX and its customers, and the methodology page says survey and transaction data are aggregated and anonymized, so supplier and buyer identities would not appear in any report. Within that boundary, a utility study would follow six rules:
- Register first: question, datasets, baseline, test sets, metric and the smallest gain that would count, all dated before training starts.
- Freeze the recipe: one base model, training recipe and compute budget per comparison, so the data is the only thing that differs.
- Match the baseline: compare against the same volume of public or synthetic data, never against adding nothing.
- Split by time and by supplier: test on records dated after the training snapshot and on suppliers left out of training, so leakage cannot pass for value.
- Repeat runs: several seeds per arm with the spread reported, so a single lucky run never becomes a finding.
- Version the scoring: DCLM's maintainers note that bug fixes to baseline calculations in September 2025 mean earlier CORE and EXTENDED scores are not directly comparable to new results, so every score should carry the scoring version it was computed under.
What can a reader conclude while no buyer-value results exist?
Nothing about value yet. The questions above have no published answers, and an agenda is a plan, not evidence; it should not be read as a claim that SourceX's datasets improve any model.
Results, when any exist, will carry limits of their own. A gain describes the model, recipe and test it was measured on and may not carry over to another setup. Datasets a buyer agrees to have measured may be the strongest ones. A company measuring the value of what it licenses has a stake in the answer, which is why disclosed conflicts matter. Buyers' tasks can reveal what they are building, so some detail may be withheld, and every withheld detail narrows what a result can prove. A benchmark gain is not revenue, demand or evidence that any dataset will be licensed.
An investor with a specific question about these designs can raise it through the investor page; SourceX shares company and commercial metrics with investors directly, and none is published here.
Methods note. The questions, the comparison table and the design rules were written for this page on 2026-10-27 from SourceX's live research, methodology and How it works pages, its terms, and two public benchmark projects read from their repositories (DCLM at commit 361714b, DataPerf at commit 25f1477). No experiment was run for this page and none is reported; the table's last column states what would count against value, not an expectation of what a study would find.
This article is for information only and is not an offer to sell or a solicitation of an offer to buy securities. Any offering by SourceX is made only to verified accredited investors through its offering documents.
Sources for this section
- DataComp-LM (DCLM) has participants filter a data pool or bring their own data, train with standardized training code and scale-specific hyperparameters, and evaluate on 53 downstream tasks to judge dataset quality; its maintainers note that baseline fixes in September 2025 make earlier CORE and EXTENDED scores not directly comparable to new ones. Source
- DataPerf, a community-led benchmark suite for datasets, includes a data-acquisition benchmark in which a buyer decides which datasets to purchase while seeing only summary information from sellers, not their datasets. Source