Veritone Launches Data Marketplace — A Governed Exchange Connecting Rights Holders and Accredited AI Developers Through Ethically Sourced, Rights-Cleared Training Data
Built on Veritone's FedRAMP-authorised aiWARE platform and paired with the Veritone Data Refinery tokenisation engine, the new marketplace creates a unified pipeline from unstructured content transformation to trusted distribution — enabling media archives, sports libraries, and content owners to monetise previously dormant assets while giving AI developers access to rights-cleared, multi-modal datasets formatted for LLMs, world models, and foundational architectures.
5 min read
Veritone, Inc. (NASDAQ: VERI), a leader in enterprise AI and data solutions, has announced the launch of Veritone Data Marketplace (VDM) — a new platform designed to accelerate the industry's shift from scraped data to ethically sourced, multi-modal datasets, helping advance AI innovation while ensuring rights holders are compensated for their contributions. Previously available only to a limited group of customers, VDM is now being expanded to the broader market to accelerate data monetisation and responsible AI development. Rights holders and AI developers can access the platform at marketplace.veritone.com.
The launch arrives at a pivotal moment for the AI data economy. As the market for AI training data matures and the supply of publicly available, unencumbered content contracts under the pressure of copyright litigation, regulatory scrutiny, and platform access restrictions, enterprises building large language models, world models, and foundational AI architectures are increasingly prioritising governed, rights-cleared data over raw volume. VDM is designed precisely for this transition — creating a trusted exchange that connects data suppliers directly with accredited AI developers, enabling both sides of the AI data economy to scale with integrity rather than with legal and reputational exposure.
"AI innovation is entering a new phase where provenance, governance, and quality matter more than raw volume. With Veritone Data Marketplace, rights holders gain a transparent path to monetize their archives, while developers access the clean, rights-cleared datasets designed to support next-generation AI."
— Ryan Steelberg, Chairman, President, and CEO, Veritone
VDM and VDR — A Unified Pipeline From Data Transformation to Trusted Distribution
VDM does not operate in isolation. It is paired with Veritone Data Refinery (VDR), the company's tokenisation engine that transforms unstructured data — video archives, audio libraries, image catalogues, sports footage, news content — into structured, AI-ready assets. The two products form a complementary pipeline: VDR creates the high-quality datasets, and VDM delivers them through a governed, monetisable marketplace. Together they address both ends of the data supply chain problem: the transformation challenge (turning raw content into usable training data) and the distribution challenge (connecting that data to the developers who need it, with the rights chain intact and verified).
Both VDM and VDR are powered by Veritone's proprietary aiWARE platform, which is used by hundreds of enterprise customers and is FedRAMP-authorised and CJIS-compliant — credentials that reflect the platform's deployment by the US government, including the Department of War and state and local law enforcement, for public safety modernisation. That compliance foundation matters for the marketplace: it means the governance architecture underpinning VDM has been validated to federal security standards, not just commercial ones.
For Rights Holders — Unlocking the Value of Dormant Archives
For content owners and media organisations, VDM opens new monetisation opportunities for valuable archives that have historically generated no revenue outside traditional licensing models. Rights holders maintain full control of their assets while offering clean, AI-ready datasets to a growing network of accredited developers — entering the AI training data economy without surrendering ownership, surrendering rights, or exposing themselves to the ambiguous use terms that have made unsolicited web scraping such a contentious issue in copyright law.
"We knew we were sitting on a massive archive, but we didn't fully understand its value. Veritone Data Marketplace helped us unlock and structure our content in a way that allows us to participate in the growing AI economy while protecting our rights and turning previously dormant assets into a meaningful new revenue opportunity."
— Gunnar Dedio, Managing Director, Progress Film
"Our content library was always valuable, but it wasn't accessible as usable data. Veritone's technology transformed our unstructured media, opening the door to new data-driven revenue and automated workflow opportunities."
— Stef Hoffer, Owner, Hoffer Media
For AI Developers — Rights-Cleared, Multi-Modal Data Formatted for LLMs and Foundational Models
For AI developers, VDM provides access to a deep and expanding catalogue of proprietary, rights-cleared data spanning media, entertainment, sports, public safety, and human behaviour — verticals where high-quality labelled and contextualised data is both scarce and strategically valuable for training specialised models. All datasets are formatted for streamlined ingestion into LLMs, world models, and foundational architectures, and crucially, can be evaluated for metadata density and model fit before acquisition — allowing developers to assess dataset quality and relevance before committing to a purchase, reducing the risk of acquiring training data that proves unsuitable for the intended model architecture.
Governance Architecture — Developer Accreditation, Chain-of-Custody Rights Tracking, and Hardened Delivery
The governance architecture built into VDM is what distinguishes it from informal or unverified data exchange arrangements. Three layers of protection are embedded in the platform: a developer accreditation process with a governance framework designed to verify the identity and intended use of developers seeking access to sensitive data; chain-of-custody rights tracking from data owner to end-user ingestion — ensuring the provenance of every dataset is documented and auditable throughout the supply chain; and hardened, IP-protected delivery pipelines that prevent unauthorised redistribution or use outside the agreed licensing terms.
Underpinning the platform's governance is Veritone's team of content and data licensing experts with over a decade of experience executing thousands of agreements across media, sports, news, and entertainment. From negotiating usage rights and managing approvals to supporting compliance with evolving copyright and distribution standards, this expert layer provides the human oversight that automated rights-tracking technology alone cannot fully substitute — particularly as copyright law and AI data licensing frameworks continue to evolve rapidly across jurisdictions. As Steelberg noted, the marketplace represents the next evolution of the AI supply chain: high-quality data sourced ethically, governed transparently, and delivered at the scale modern AI demands.
Key Takeaways
- Veritone has launched Veritone Data Marketplace (VDM) — now broadly available after a limited customer phase — as a governed exchange connecting data rights holders directly with accredited AI developers for ethically sourced, rights-cleared, multi-modal training datasets spanning media, entertainment, sports, public safety, and human behaviour
- VDM pairs with Veritone Data Refinery (VDR), the company's tokenisation engine that transforms unstructured content into AI-ready assets — together forming a unified pipeline from raw content transformation to governed marketplace distribution, both powered by the FedRAMP-authorised, CJIS-compliant aiWARE platform
- For rights holders, VDM unlocks new monetisation for previously dormant archives — content owners maintain full asset control, participate in the AI training data economy on defined terms, and avoid the legal and reputational exposure of unsolicited web scraping that has made ambiguous data use such a contested issue in copyright law
- For AI developers, all datasets are formatted for direct ingestion into LLMs, world models, and foundational architectures, and can be evaluated for metadata density and model fit before acquisition — reducing the risk of purchasing training data that proves unsuitable for the intended use case
- Three governance layers protect all parties: developer accreditation with identity and intent verification, chain-of-custody rights tracking from data owner to end-user ingestion, and hardened IP-protected delivery pipelines — backed by Veritone's team of licensing experts with over a decade of experience across thousands of media, sports, news, and entertainment data agreements