Artmarket.com said its controlling shareholder Groupe Serveur and the Ehrmann family intend to buy additional shares in the art-market data company, a move that coincides with a board-backed decision to rebuild its core databases around artificial intelligence.
The planned purchases will be disclosed to France’s financial markets regulator, the Autorité des marchés financiers, within statutory deadlines and carried out during permitted trading periods, Artmarket.com said on Aug. 13. The company did not disclose the number of shares sought, the expected value of the purchases, or a timetable.
The proposed increase in ownership arrives as Artmarket.com commits more resources to an “AI-First” overhaul of Artprice, its art-market information subsidiary. The company said the new plan replaces a gradual strategy of placing AI modules into existing systems with a deeper redesign of its data architecture and operating workflows.
Artmarket.com said the work will cause a minor revision to the public rollout schedule for the revamped service, while maintaining its existing financial trajectory. It also said revenue continues to show steady growth, without providing figures in the announcement.
A full rebuild of Artprice’s data infrastructure
The restructuring covers nearly 180 interconnected proprietary databases and a collection of auction catalogues and manuscripts spanning more than three centuries, according to Artmarket.com. The data includes auction results, artist records, images and historical sales documentation that form the basis of Artprice’s pricing and market-analysis products.
Rather than adding generative AI features over time, the company said it is refitting data collection, normalization and enrichment pipelines around deep-learning standards and proprietary algorithms. It has also begun equipping staff, departments and production units with dedicated AI hardware and edge-computing devices, which process data closer to where it is generated instead of relying entirely on centralized cloud infrastructure.
That approach places the company’s historical catalogue at the center of the project. Artmarket.com said its systems contain tens of millions of unique records protected by patented algorithmic architectures. Its database includes more than 30 million indices and auction results relating to more than 915,300 artists, while Artprice Images contains 181 million digital images of works dating from 1700 to the present.
For an art-data provider, the difficult part is less about producing a chatbot-style interface than converting uneven source material into reliable, comparable records. Auction catalogues may describe the same artist, work, provenance or condition in different ways across countries and decades. Artmarket.com’s AI-led redesign appears intended to automate more of that standardization while preserving the identifiers needed to trace individual works and sales histories.
The company said it will release the subscriber version of the new platform after recalibrating its internal processes. It described the planned product as a unified AI-First environment built on standardized big data, deep learning and algorithmic security.
Price-gap model expands beyond auction estimates
Artmarket.com also highlighted its “Blind Spot©” system, which it said began as a model designed to identify price gaps between publicly recorded auction sales. The tool was initially developed to address a familiar challenge in art-market valuation: a work may have a limited recent sales history, leaving substantial uncertainty between one public auction appearance and the next.
The company offered a Jackson Pollock example, stating that a work sold for $4 million in 1998 and later reached $58 million in 2026. Such examples illustrate the scale of long-term price changes that can complicate estimates based on sparse comparable sales, although they do not establish that an automated model can predict individual future auction outcomes.
Artmarket.com said Blind Spot has since grown into a broader analytical engine used across several categories of analysis. The success of that expansion will depend heavily on whether the company can make its underlying records consistent enough to distinguish meaningful market signals from differences in cataloguing, attribution, condition or sale terms.
Auction data can be particularly vulnerable to those distinctions. A headline hammer price alone may omit buyer’s premiums, currency effects, whether a lot was guaranteed, or whether a work had distinctive provenance. An AI system can speed up comparisons, but the output remains dependent on the quality and completeness of the inputs.
Auction-house network supplies the raw material
Artmarket.com said it works with 7,200 auction-house partners through contracts and partnerships. The company uses deterministic software agents to collect and organize market information from that network, according to the release.
Deterministic agents follow predefined rules, making them useful for structured tasks such as identifying standard lot fields, extracting sale dates and categorizing auction results. Artmarket.com said it is also deploying probabilistic agents to support commercial expansion in the United States. Such systems use statistical inference to assess uncertain information, a potentially useful capability when records are incomplete or expressed in inconsistent language.
The company’s art-market content operation, Artprice News, publishes dispatches nearly every hour with a focus on North America, it said. Artmarket.com also said it distributes art-market trend material across 121 countries in 11 languages and operates a fixed-price marketplace with 9.3 million members.
Its annual subscription service is priced between $1,600 and $2,500, or €1,600 to €2,500, the company said. The planned platform is intended to operate in more than 40 languages, while retaining artwork titles in their original form rather than translating them under Artmarket.com’s internal rules.
That language policy reflects a practical issue for art-market search tools: translated titles can make a work easier to discover, but can also create duplicate records or obscure the wording used in original catalogues. Maintaining original titles while standardizing associated metadata could help the company preserve historical accuracy across an international database.
Control remains concentrated as the rebuild advances
Artprice was founded in 1997 and is controlled by Groupe Serveur, which Artmarket.com said was established in 1987. The company reported that its latest TPI shareholder analysis counted more than 18,000 individual shareholders, excluding foreign shareholders, companies, banks, French investment funds and UCITS.
The intended share purchases by Groupe Serveur and the Ehrmann family would increase the stake held by the company’s existing controlling side rather than introduce a new strategic owner. With no purchase size disclosed, the immediate market effect cannot be measured from the announcement alone.
Thierry Ehrmann has also completed an 1,800-page essay on artificial intelligence covering the period from 1987 to the present, Artmarket.com said. The work is scheduled for global release in digital and print editions, with an English-language version planned before the French edition.
The company’s near-term test lies in executing the AI rebuild without disrupting access to the historical records that subscribers pay to use. Its decision to delay the public launch slightly while revising internal workflows suggests Artmarket.com sees data quality and system integration as more urgent than meeting its original deployment sequence.
Explore how traders harness AI for smarter portfolio decisions in crypto with our in-depth guide: AI copy trading.
Disclaimer: The content on this page is provided for general informational purposes only and does not represent the views or financial advice of Toobit. We make no guarantees regarding the accuracy or completeness of this information and shall not be held liable for any errors, omissions, or outcomes resulting from its use. Investing in digital assets involves risk; users should independently evaluate their financial situation and the risks involved. For further details, please consult our Terms of Service and Risk Disclosure.
