What’s happening in data licensing
Quick answer, updated
We follow people working in company data licensing to understand what they are seeing, where their views differ, and what a possible seller should ask next.
Dated observations from public conversations on X. Source statements and our reading are kept separate. A market signal is a reason to investigate, not a price or an offer for your data.
Oct 19, 2026 to Oct 25, 2026
Our reading of this period
The voices in this selection focus on finding suitable sellers, understanding complete workflows and deciding how value is shared. Mercor discusses introductions, Trent Krupp and Hugo Leprince describe what buyers seek, and Ryan Lock and Korra raise questions about deal structures and process. Our reading: a useful starting point for a seller is a clear account of the work its records capture, followed by careful comparison of the proposed rights and payment terms. These posts do not establish market prices or the likelihood of a sale.
This is our first brief and a baseline for following the conversation. There is no comparable earlier collection from which to infer a market trend.
Mercor · Brendan Foody
Finding suitable sellers is part of the business
Our reading. Mercor's company account and CEO describe a referral route as well as data payouts. That makes seller sourcing visible in this selection; it does not show that all buyers face a shortage or that sellers can command higher prices.
What remains uncertain. Both posts come from the same organization. The referral ceiling and paid-out total are company claims, not independent evidence about market conditions.
For your company. An introduction can have its own payment terms. A seller can ask who is being paid for the introduction while separately assessing the offer for its records.
What we are watching. Whether other buyers publish comparable referral terms, and what those terms say about eligibility and when a fee is earned.
Read the discussion (2 posts)
“We’ve paid out $50M+ to businesses for their data. Now we’re paying for introductions too.”
Context. Mercor presents this as its own paid-out total. Its data page also describes a referral route, but that does not independently verify the amount paid or tell a seller what their records would earn.
Read @mercor’s post on X“We’ll pay you up to $100K for every company you refer that licenses its data to @Mercor.”
Context. The stated ceiling is a fee for an introduction, not a price offered for the referred company's data. This is a statement from Mercor's CEO, not an independent seller account.
Read @BrendanFoody’s post on XTrent Krupp · Hugo Leprince · Korra
Workflows and operating history shape the conversation
Our reading. Krupp's emphasis on workflows and Leprince's demand profile point toward business context as a useful way to describe a dataset. Korra's posts also emphasize company size, although the anecdotes do not establish a universal threshold.
What remains uncertain. Korra's figures lack a stated sample and period, and the smallest-company figure is attributed to an unnamed lab. They cannot establish acceptance rates for the wider market.
For your company. Before approaching buyers, describe the systems you use, the years of history available and how records connect a request to an outcome. Published buyer criteria can then be compared with those facts.
What we are watching. Whether buyers publish more specific workflow requirements, and whether documented examples show viable cases outside the company sizes discussed here.
Read the discussion (6 posts)
“Everyone knows Labs are buying data, but they don't understand what they are actually buying:”
Context. In the surrounding post, Trent Krupp describes the interest as workflows. That suggests a useful description for a seller to prepare: how a request moves through the company's tools to an outcome. It does not establish a price premium.
Read @TrentKrupp’s post on X“When I'm buying a dataset what I want:”
Context. The post's intake list asks for operating history, people, outcomes and the systems holding records. A company can prepare that inventory without sending files or granting account access.
Read @TrentKrupp’s post on X“Somehow just got huge demand for knowledge work, US based business data, +100 FTEs, +5 yrs in biz.”
Context. Hugo Leprince describes a specific demand signal he has received. It is a snapshot from one founder, not a standing requirement across buyers or evidence that every company matching it will receive an offer.
Read @hugoleprince’s post on X“An HVAC company with over 100 employees got around $350K to $400K for its records.”
Context. A broker's account of one HVAC company is a lead for further research. The excerpt gives no contract, dataset scope or independent confirmation, so the figure cannot price another company's records.
Read @GotDataa’s post on X“Anything over 100 employees is closing at about 90%. Out of 100 smaller files, maybe three close.”
Context. Korra reports different close rates by company size, but gives no sample, period or definition of a close in this excerpt. These are not a seller's probability of completing a deal.
Read @GotDataa’s post on X“A CEO of a lab told me their acceptance rate is 2% for anything less then 50 employees.”
Context. Korra reports an unnamed lab CEO's figure. The lab, sample and measurement period are absent from this excerpt, so it cannot establish a rule for smaller sellers or be attributed to a named buyer.
Read @GotDataa’s post on XTrent Krupp · Ryan Lock · Hugo Leprince · Korra
The payment model is only part of the deal
Our reading. Krupp describes upfront cash with broad rights, while Lock describes interest in revenue share. Together they raise a practical seller question: which future uses and payments does the proposed agreement actually cover? Leprince's resale thesis and Korra's concern about duplicate approaches add questions about the route to the end user.
What remains uncertain. Krupp separately distinguishes a buyer's exclusive rights from a broker's exclusive representation. Those are different restrictions. These posts do not compare matched offers or establish which payment model pays sellers more, and a reported resale multiple is not a typical return.
For your company. Compare the payment trigger alongside permitted uses, duration, exclusivity and reporting. Have counsel review the agreement before treating shorthand in a post as the terms you would sign.
What we are watching. Published examples of contract scope, revenue definitions, reporting and payment triggers that let a seller compare structures on the same basis.
Read the discussion (5 posts)
“RL Labs: Cash upfront, they own the data, exclusively and in perpetuity.”
Context. Krupp describes a deal pattern involving upfront cash and broad rights. The signed agreement determines scope, ownership, duration and exclusivity; this shorthand cannot stand in for those terms.
Read @TrentKrupp’s post on X“Seeing a growing trend around a Rev Share model for sellers to participate in Envs/task revenue instead of a contracted payout.”
Context. Ryan Lock reports interest in revenue-linked compensation. The excerpt does not establish how common that structure is or say that it is Polyshares' standard offer. The calculation and payment conditions matter.
Read @RyanLockAI’s post on X“The biggest flip of all time, buy the data, build top RL env, sell it 5-10x more.”
Context. Leprince describes a possible resale multiple after data is developed into an RL environment. The excerpt does not document a completed resale, the development costs or a seller's entitlement to that revenue.
Read @hugoleprince’s post on X“If you show the same dataset to several buyers, the lab ends up seeing it multiple times, and it loses its value.”
Context. Korra argues that overlapping approaches to labs can reduce value. The excerpt supplies no evidence of the effect, and a broker has an interest in how introductions are managed. It is a question to clarify, not a reason to assume exclusivity is necessary.
Read @GotDataa’s post on X“Labs dont really care about the former, but the later is super annoying to deal with if the Lab gets the same dataset twice from different sources.”
Context. Krupp distinguishes exclusivity for the lab buying the data from exclusivity for the broker marketing it. In his view, labs care about duplicate approaches from brokers. That is his interpretation of lab preferences, not a rule or evidence that a seller must accept exclusive representation.
Read @TrentKrupp’s post on XKorra
The route from interest to a decision can be unclear
Our reading. Korra's comments highlight a distinction between finding interest and obtaining a decision. This is one broker's reported experience, but it gives sellers concrete process questions to ask before committing time or granting exclusivity.
What remains uncertain. Both observations come from one company, and neither excerpt establishes a typical timeline, fee stack or experience across the market.
For your company. Ask who is handling each step, which counterparties may receive an approved brief and what happens if no decision arrives. A clear process helps a seller understand its commitments.
What we are watching. More first-hand accounts of actual deal stages, and clearer published explanations of who contracts with and pays the seller.
Read the discussion (2 posts)
“Four or five parties. Each takes a cut and adds weeks.”
Context. Korra describes several parties and fees in a deal chain. The excerpt does not identify those parties, so it cannot establish a relationship between named companies or how much a seller would keep.
Read @GotDataa’s post on X“Labs are horrendous at getting back to us on files we send.”
Context. This is Korra's account of response times on files it submits. It documents the broker's reported frustration, not a market-wide timeline or a promise about the next seller's experience.
Read @GotDataa’s post on XThe voices behind this brief
15 selected posts from 6 accounts. This is a reading of a selected conversation, not a survey of the market. Figures and experiences are the authors’ claims; our interpretation is identified above.
This selection mainly reflects buyers and intermediaries. The collection log reports searching 37 watchlist accounts for original posts; searches for four others and wider topics were partial. The log includes corrected excerpts, but original posts were unavailable for an independent editorial recheck.
People and accounts in this brief
- Mercor · Company account, Mercor@mercor on X
- Brendan Foody · CEO, Mercor@BrendanFoody on X
- Trent Krupp · Co-founder and former CRO, Troveo@TrentKrupp on X
- Hugo Leprince · Founder, Maryah@hugoleprince on X
- Korra · Company account, Korra (broker)@GotDataa on X
- Ryan Lock · Co-founder, Polyshares@RyanLockAI on X
Related buyer records
Compare their published criteria and terms alongside these observations.
What is your company’s data worth?
Answer a few questions about the company. We reply with an estimated range, the buyers that fit and the contract terms to watch. Free for sellers, and we never ask for data, credentials or exports.