Video and footage · Broadcast and news archives · Updated 30 Oct 2026
Can a broadcaster license its news and TV archive to AI companies, and what is it worth?
AIData.Broker represents the companies and people that hold broadcast archives when they sell to AI labs: we value what they hold against public prices, describe it without their name to every buyer that takes it at the same time and help negotiate the terms. Our fee is 12% of the first $250,000 a sale pays, and nothing if no sale closes.
News, documentary and programme archives, with their transcripts and logs. Decades of professionally shot, edited and described video and speech, which buyers of both video and speech pay for.
- No data upload
- About two minutes
- No email needed
- Prices on record
- 14
sales that happened, offered to holders, asking prices off the shelf, cost to produce
- Who holds it
- Local and national TV stations, Radio broadcasters, News agencies
- Who has paid, offered or listed
- Protege, Shaip, ELRA, Linguistic Data Consortium, Korea Media and Communications Commission
The short answer · figures from the market's records, last reviewed 30 Oct 2026
Can a broadcaster license its news and TV archive to AI companies, and what is it worth?
Yes, the programmes it owns outright. Protege reports that an EMEA broadcaster earned over $1M in 6 months licensing its scripted and cinematic library, and Korea's media regulator paid broadcaster-led consortia KRW 20 billion in 2025 to turn KBS, MBC and KT ENA archives into labelled training video. Broadcast speech sells in volume and cheaply per hour: a French broadcast news corpus is listed at €31,250 for 900 hours for commercial use by ELRA. Agency footage, acquired programmes and music inside the archive stay their owners'.
What buyers have paid, offered or asked for broadcast archives
The whole price boardEvery price in the market's records about broadcast archives, each with its source and the day it was read. Open a row for the source's own words.
14 public prices across all kinds of data
An EMEA broadcaster's scripted and cinematic library licensed for AI training through Protege
The intermediary's own case study. Protege held exclusive AI training rights to the catalogue, so buyers could not play rival sellers against each other. Segment corrected on 2026-10-17 from source:broadcaster to source:film: the licensed asset is a scripted and cinematic library; no figure changed. It is a contract total, not a per-hour price.
Video and footageSales that happenedOct 2025The payer's own page
$1M or more
earned in six months
Existing Indic-language audio a holder already has (podcasts, radio programmes, news broadcasts, talk shows; published or archived), licensed to Shaip for AI training only
The only buyer page read that states a per-hour price for audio a holder already has; the licence is for training only, the content is not published or redistributed and ownership stays with the holder. Media houses and radio channels are named among those it partners with, and the page pays 5% of the contract value to whoever brings in a holder. Undated page (footer 2026).
Audio and speechOffered to holdersOct 2026The payer's own page
Up to $2.11
per hour of audio
NEMLAR Broadcast News Speech Corpus: 40 hours of Arabic radio broadcast news (ELRA catalogue ELRA-S0219)
ELRA's non-member price for a company under the commercial-use (VAR) licence; an academic non-commercial licence is EUR 300. Discounts of 15% and 30% when bought with the other NEMLAR resources. Broadcast news, so mostly MSA.
Audio and speechAsking prices off the shelfOct 2026A catalogue or marketplace listing
ELRA catalogue, NEMLAR Broadcast News Speech Corpus (ELRA-S0219)
$4,477
per corpus of 40 hours
Published as 4000.00 € (non-member, commercial use)
$112 per hour of audio
NetDC Arabic Broadcast News Speech Corpus: about 22.5 hours of Standard Arabic radio news from Radio Orient (France), November 2001 to January 2002, about 90 speakers (ELRA catalogue ELRA-S0157)
ELRA's non-member price for a company under the commercial-use (VAR) licence; an academic non-commercial licence is EUR 200. Twin of LDC's VOA Arabic set ($700 for 10 hours, research only).
Audio and speechAsking prices off the shelfOct 2026A catalogue or marketplace listing
ELRA catalogue, NetDC Arabic BNSC (Broadcast News Speech Corpus) (ELRA-S0157)
$3,022
per corpus of 22.5 hours
Published as 2700.00 € (non-member, commercial use)
$134 per hour of audio
Arabic Broadcast News Speech: 10 hours of Voice of America Arabic satellite radio news recorded 2000 to 2001, 16 kHz (transcripts sold separately)
LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership. The reduced-licence fee is $350.00. Mostly Modern Standard Arabic news reading.
Audio and speechAsking prices off the shelfOct 2026A catalogue or marketplace listing
$700
per corpus of 10 hours
$70 per hour of audio
ALLIES corpus: 900 hours of French broadcast news for diarization evaluation (ELRA catalogue ELRA-S0486)
ELRA's non-member price for a company under the commercial-use (VAR) licence. Broadcast speech in volume, so cheap per hour. A non-commercial-use licence is EUR 2,500 for an academic and EUR 9,375 for a company.
Audio and speechAsking prices off the shelfOct 2026A catalogue or marketplace listing
$35K
per corpus of 900 hours
Published as 31250.00 € (non-member, commercial use)
$39 per hour of audio
ESTER 2: about 100 hours of transcribed French broadcast news (ELRA catalogue ELRA-S0338)
ELRA's non-member price for a company under the commercial-use (VAR) licence. Commercial use costs several times the non-commercial price. A non-commercial-use licence is EUR 2,000 for an academic and EUR 7,500 for a company.
Audio and speechAsking prices off the shelfOct 2026A catalogue or marketplace listing
$28K
per corpus of 100 hours
Published as 25000.00 € (non-member, commercial use)
$280 per hour of audio
TRAD Pashto Broadcast News: 108 hours of transcribed TV and radio news from more than 1,000 speakers (ELRA catalogue ELRA-S0381)
ELRA's non-member price for a company under the commercial-use (VAR) licence. A low-resource language, priced well above broadcast speech in big languages. A non-commercial-use licence is EUR 3,500 for an academic and EUR 28,000 for a company.
Audio and speechAsking prices off the shelfOct 2026A catalogue or marketplace listing
ELRA catalogue, TRAD Pashto Broadcast News Speech Corpus (ELRA-S0381)
$31.4K
per corpus of 108 hours
Published as 28000.00 € (non-member, commercial use)
$291 per hour of audio
AIDA Ukrainian Broadcast and Telephone Speech: about 156 hours with 1.2 million words of transcripts
LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership. The reduced-licence fee is $1,000.00.
Audio and speechAsking prices off the shelfOct 2026A catalogue or marketplace listing
LDC catalog, AIDA Ukrainian Broadcast and Telephone Speech Audio and Transcripts (LDC2023S01)
$2,000
per corpus of about 156 hours
$13 per hour of audio
GALE Phase 2 Arabic Broadcast News Speech Part 1: about 165 hours of Arabic broadcast news
LDC's one-time non-member licence fee, shown under View Fees; LDC licences to non-members exclude commercial use, which needs a for-profit membership. The reduced-licence fee is $1,000.00.
Audio and speechAsking prices off the shelfOct 2026A catalogue or marketplace listing
LDC catalog, GALE Phase 2 Arabic Broadcast News Speech Part 1 (LDC2014S07)
$2,000
per corpus of about 165 hours
$12 per hour of audio
Rights-cleared archive video (sports federations', broadcasters' and other rights holders' footage) licensed to AI developers through the Veritone Data Refinery, priced per hour by application and metadata
The price range the broker charges AI buyers, as its CEO stated it; the rights holder receives a revenue share of it (share not stated), so the holder's take per hour is lower. The thousand-fold range turns on tagging and metadata, which is itself a pricing lever.
Video and footageAsking prices off the shelfJan 2026Press report
$1.00 to $1,000
per hour of video
Korea's 2025 broadcast-video AI training data programme: about 40,000 hours selected from over 2 million hours of KBS, MBC, MBC Chungbuk and KT ENA archive (news, documentary, drama), cleaned and labelled into 23,113 finished hours and about 4.6 million data sets
The commission's own budget, run with the Korea Radio Promotion Association, which paid broadcaster-led consortia (KBS, MBC, MBC Chungbuk, KT ENA) to turn their own archives into labelled training data; it prices selection, cleaning and labelling together with access, not a licence of raw footage to a model developer. The content mixes news, documentary and drama, so it fits no single source segment.
Video and footageCost to produceMay 2026The payer's own page
$13.8M
per programme of 23,113 finished hours
$597 per hour of video
MBC-led consortium (with iMBC, DataMaker, an LG AI research arm and CrowdWorks) in Korea's 2025 broadcast-video AI training data programme: a planned dataset of about 5,000 hours built from MBC broadcast video
One of the four consortia of the programme the held KMCC row records in total (MBC, MBC Chungbuk, KT ENA, KBS, each given the same amount), so this restates part of the same money at the consortium's own planned hours and is not an independent sale; the medians count the programme once. The hours are a plan (구축할 예정이다), not delivered hours, and the content mixes broadcast genres, so it fits no single source segment.
Video and footageCost to produceAug 2025Press report
$3.5M
per consortium building about 5,000 hours of broadcast video data
$695 per hour of video
Video highlight and description AI training data (AI-Hub data 41), Korea's 2021 AI training data programme, KAIST consortium with KBS: 6,001 source hours of KBS broadcast video (1,001 hours) and YouTube video (5,000 hours), labelled into 18,123 processed hours
The science ministry and NIA fund a KAIST consortium (KBS, Metrix Research, Sorizava, DataMaker, ActionPower and others) to build, label and release the set free on AI-Hub, so this prices collection, labelling and model work together and is an upper bound for the footage alone; it is a planned figure that includes private money. News-led by name, but KBS supplied 1,001 of the 6,001 source hours and YouTube the other 5,000, so it is not filed as news footage. The hours are AI-Hub's record of the delivered set (6,001 source hours, 18,123 processed hours): https://www.aihub.or.kr/aihubdata/data/view.do?currMenu=115&topMenu=100&dataSetSn=616
Video and footageCost to produceJul 2021Press report
$3.7M
per project of 6,001 source hours of video
$609 per hour of video
Where this data comes from, and who lists it
The whole supply mapmatch video bought outright with permanent rights from Satba TV and MBC; 62,769 video sets
AI-Hub's page for the state-funded ssireum set (62,769 video sets and 220,000 image sets) says its match video was bought from Satba TV (샅바TV) and the broadcaster MBC with permanent rights, while its training video was filmed at the Yongin University ssireum hall. The price paid is not published. The derived set is released on AI-Hub.
Reaches buyers: State-funded dataset programme buying existing sports broadcast footage outright, then releasing the labelled derivative on AI-Hub
NIA AI-Hub dataset page: 한국 씨름 동작 데이터 (Korean ssireum motion data), 30 Oct 2026
3,005 hours of broadcast video, 485 hours of it news
AI-Hub's 2023 broadcast-video set holds 3,005 hours in 5,958 items from TBS and EBS programmes in eight genres, of which news is 485.4 hours in 1,224 items. The page gives no budget. Released on AI-Hub.
Reaches buyers: State-funded set built from broadcaster archives and released on AI-Hub, a free Korean substitute for broadcast news footage
1,000 hours or more, 4,000 videos (239 news and politics, 323 sports)
AI-Hub's 2020 video-summary set uses 4,000 source videos of 2 to 60 minutes, 1,000 hours or more, in 15 categories collected from YouTubers and broadcasters, including 239 news and politics videos and 323 sports videos in its category table. The page gives no budget; a 2020 press report on the consortium's task (an emotion-recognition public database that also covers video summarisation data) gives KRW 4.6 billion with no split by hours, so no per-hour figure can be taken from it.
Reaches buyers: State-funded set released on AI-Hub, a free Korean substitute for general video summarisation training data
7 news, sports and esports holders among 16 featured data owners
Troveo's page for content owners names, among its featured data owners, a US local TV news group (Sinclair Broadcast Group), a South Asian news agency (Asian News International), Australia's Nine Network, the financial news site Benzinga, an NCAA Division I conference (Mountain West Conference), Barstool Sports and the esports brand 100 Thieves. It lists 'Sports & Movement' and 'Aerial/Drone' among the categories it licenses. No rate or royalty share is published.
Reaches buyers: Holders license through Troveo, which licenses to AI labs under NDA and pays a royalty on every use
Troveo, License Your Content to Top AI Companies (content owners page), 27 Oct 2026
155K+ hours
Troveo's catalogue lists a Broadcast Sports dataset of more than 155,000 hours of live and recorded broadcasts across 25+ sports with 8+ camera angles, commentary, replays and graphics intact. It is offered by request with no price, under Troveo's private, NDA-bound direct-to-lab licensing.
Reaches buyers: Broker pooling rights holders' sports broadcasts into datasets licensed privately to AI labs
408K+ hours
Troveo's catalogue lists a Broadcast News dataset of more than 408,000 hours of daily-news-cycle output (anchor reads, packages, live field shots, interviews and raw B-roll), in finished and raw form. It is offered by request with no price.
Reaches buyers: Broker pooling news broadcasters' output into datasets licensed privately to AI labs
- Programmes the broadcaster produced and owns outright
- Transcripts, logs and descriptions with the footage
- Decades of archive in one format
- Speech in languages buyers lack
A broadcaster licenses what it produced itself. Agency footage, acquired programmes, music and archive clips it licensed in stay their owners', and contracts with presenters and contributors may limit new uses.
Who has bought, offered or listed broadcast archives
- Protege
- Shaip
- ELRA
- Linguistic Data Consortium
- Korea Media and Communications Commission
- Korea Radio Promotion Association
- KBS
- MBC
- Crowdworks
- KAIST
- NIA Korea
- Korea Ministry of Science and ICT
Protege
Shaip
ELRA
Korea Radio Promotion Association
KBS
MBCBuyers that take video, and what each publishes
13 buyers taking video and footage, largest published figure first
Hold broadcast archives? A valuation takes about two minutes: it says what buyers pay and whether we would run the sale. Prices last read 30 Oct 2026.
In this topic
What sells
- Start hereWhich data do AI labs want, and who holds it?Each kind of data the market prices, the businesses that hold it, what labs are asking for now and what they turn away.
- GuideWhat Are RL Environments? Who Builds, Buys and Prices ThemAn RL environment is a sandbox where an AI agent attempts tasks with tools and a verifier scores each try. Who builds them, who buys them and what they cost.
- GuideHow to Sell Your Videos for AI Training: Prices and BuyersAI companies pay to license your videos for training, by the hour or the minute. What each buyer pays, which footage sells, and the rights every frame needs.
- GuideSelling audio and speech recordings for AI trainingWhat AI companies pay for call recordings, podcasts, radio, meeting audio and voice data, why archive audio sells cheapest, and the recording-consent law state by state.
- GuideSelling photos and image archives for AI trainingWho buys photos for AI training, what stock agencies pay contributors (shares and bonuses, not prices), what image archives sell for, consent and releases agency by agency, and how to prepare a library for a sale.
- GuideSelling your company's Slack data for AI training: what the export allows and who buys itYes, a company can license its Slack history to AI buyers. What Slack's export and API terms allow on each plan, whose messages are inside, what buyers pay and the steps.
- GuideSelling company documents and shared drives for AI trainingYes, a company can license its documents, wikis and shared drives to AI buyers. Which documents they pay for, what Google Workspace, Microsoft 365, Confluence and Notion let you export, who owns the files, and the steps.
- GuideSelling customer support tickets and chats for AI trainingOften yes, after three checks: your customer contracts and privacy notice, your help desk's export terms, and removing personal data. What buyers want in a ticket, what records with tickets fetch, and the steps.
- GuideSelling a company's email archive for AI training: exports, consent and priceYes, a company can license its own email to AI buyers. How Google Workspace and Microsoft 365 let you export a whole organisation's mail, the deadlines when a subscription lapses, the law on staff and customer messages, and what buyers pay.
- GuideSelling CRM data for AI training: Salesforce and HubSpot records, consent and priceA company can license its CRM history to AI buyers, but CRM records are customer data first. What buyers value in Salesforce and HubSpot, how each lets you export, why the contact list usually stays out, and what CRM history fetches.
- GuideSelling meeting recordings and transcripts for AI trainingCan a company license its Zoom, Google Meet or Teams recordings and notetaker transcripts to AI buyers? Recording consent state by state, what Zoom's terms allow, what meeting audio and transcripts fetch, and the steps.
- GuideSelling accounting and finance records for AI training: QuickBooks, Xero and NetSuiteA company can license its accounting history to AI buyers, but finance records carry customers', suppliers' and lenders' confidential terms and staff payroll. What buyers value in QuickBooks, Xero and NetSuite books, how each lets you export, what has to come out, and what finance records fetch.
- GuideSelling Jira and Linear issue history for AI trainingYes, a company can license its Jira or Linear issue history to AI buyers. It is valued when each issue links to the code that closed it. What each export keeps, what buyers ask for, and what it fetches.
- GuideSelling design files and product history for AI training: Figma and beyondYes, a company can license its Figma files and design history to AI buyers. What buyers value, what Figma lets you export, client work, fonts and other assets that are not yours, who owns the designs, and the steps.
- GuideSelling IT service desk records for AI training: ServiceNow tickets, runbooks and incidentsOften yes. IT tickets, incidents and runbooks teach agents to resolve real problems, once credentials, hostnames and people are removed. What ServiceNow and Jira Service Management let you export, what buyers ask for, and what the records fetch.
- GuideSelling Books to AI Companies: Prices, Deals and Who Gets PaidAI companies license books by the title, mostly from publishers' backlists. What deals paid per book, how much reaches the author, and how to sell a catalogue.
- GuideSelling Game Code and Game Data to AI CompaniesAI companies license game studios' code with its history and gameplay recorded with the player's inputs. Who buys, what studios are offered, and what you may not sell.
- GuideSelling a company's paper archive for AI trainingYes, a paper archive can be licensed to AI buyers once it is scanned into searchable text. What scanning costs, what text fetches, who owns old records, and the steps.
- GuideSelling robot and first-person video data for AI trainingRobotics companies pay for first-person video and teleoperation data, mostly recorded to their spec. What each kind pays, why bulk footage collapsed, consent, and the steps.
- GuideSelling sales call recordings and sales emails for AI trainingYes, if each call was lawfully recorded and the prospects come out. Who buys sales calls and sales emails, what they fetch, and the consent and AI-agent rules.
- GuideSelling CAD files and engineering drawings for AI trainingYes, if the company owns the designs and export rules allow it. Who buys CAD files and drawings for AI training, what they pay, and what to check first.
- GuideCan You Sell AI-Generated or AI-Assisted Data for AI Training?Rarely if a model made it: buyers screen generated data out. Work people made with AI tools still sells when you can show which parts a model wrote.
- GuideAI Content Licensing: How Publishers Get Paid, and How MuchAI companies license content to train models or to ground answers. What publishers have been paid, which platforms pay per use, and how to choose a route.
- PageWhere data comes fromWho holds each kind of data, how much exists and how it reaches buyers.
- RegisterCompanies closingCompanies closing, sold or gone dark, and the records they leave.