Wednesday, November 18, 2026
HomeStartupOpinion: Sustaining the open net is central to Europe’s AI sovereignty ambitions

Opinion: Sustaining the open net is central to Europe’s AI sovereignty ambitions


Europe’s ambitions for AI sovereignty rely on giving its startups the assets to succeed, innovate and compete with established corporations. Because it stands, the EU stays closely reliant on digital applied sciences developed and managed from elsewhere, elevating questions on its skill to regulate the programs underpinning its financial system.  

In June this 12 months, the European Fee unveiled its European Technological Sovereignty Bundle. This contains the Cloud and AI Growth Act, which particularly goals to assist the analysis, improvement and innovation of next-generation AI applied sciences, together with frontier AI fashions.  

The Tech Sovereignty Bundle additionally goals to fast-track the implementation of the European Fee’s present AI Continent Motion Plan. This encompasses many areas which can be central to the EU’s AI sovereignty ambitions, together with bodily infrastructure, expertise improvement, and regulatory simplification. One of the vital notable of those is entry to giant volumes of high-quality knowledge, which the plan particularly highlights as a requirement for bolstering AI innovation. 

For European AI startups, publicly out there net knowledge is a helpful useful resource for coaching fashions and constructing companies that depend on present data. But rising licensing fashions that deal with public knowledge as unique, regulatory uncertainties and restrictions on automated entry danger placing that useful resource past the attain of smaller builders. Sustaining accountable entry to the open net is due to this fact central to enabling companies of all sizes to contribute to the EU’s AI sovereignty ambitions.  

AI sovereignty depends upon an open net 

The web itself was constructed on the precept that data will be revealed and accessed extensively, with none one firm controlling entry to it. Inside a couple of years of launching the World Extensive Net, CERN positioned the software program within the public area for anybody to make use of, modify and distribute. This openness allowed companies, researchers and builders to construct on shared foundations with out in search of permission from a single know-how supplier.  

This precept underpinned, and continues to assist, each the event and functioning of frontier AI programs. Various public net knowledge will help cut back gaps and biases in coaching knowledge, serving to AI fashions to ship correct, unbiased outputs. The rise of agentic AI has additional elevated the significance of entry to public net knowledge. It’s being deployed throughout a rising vary of companies, from private purchasing and journey assistants to enterprise analysis and IT assist. 

Microsoft’s Work Pattern annual index report discovered a 15-fold enhance in energetic brokers throughout its Microsoft 365 ecosystem between March 2025 and March 2026. For brokers finishing up duties on the internet, nevertheless, restrictions on entry to present data can restrict their effectiveness. For instance, an agent evaluating suppliers or reserving journey must test altering costs, availability and phrases earlier than performing. 

In October, analysis discovered that roughly a 3rd of the a number of hundred main US commerce, journey and reservation web sites it examined blocked Meta’s Muse agent, whereas one other third challenged or restricted its entry. The most important downside right here will not be for Meta, who’ve been negotiating leverage and huge monetary assets for authorized disputes, however many smaller agent builders don’t.

This limits their skill to innovate, as their brokers have much less entry to up-to-date data. Shoppers additionally lose out, with fewer competing companies to select from and brokers that could be unable to match the total vary of merchandise, costs and gives out there. 

Such arbitrary restrictions threaten the EU’s AI ambitions as its AI startups are excluded from the market by the selections of competing personal corporations. Additionally they elevate questions on shopper sovereignty within the free market and the way it could also be restricted. 

Whereas regulators focus on opt-outs, blanket blocks are carried out 

In September 2026, the European Fee launched a session on whether or not the EU’s copyright framework wants extra measures, together with these addressing the usage of copyright-protected content material in AI. The ensuing coverage selections may have important implications far past AI.  

A key danger is that overly broad measures have unintended penalties as knowledge assets beforehand used for innovation and worth creation change into nearly ineffective for society. They change into unique to the largest market gamers. For European startups, this interprets straight right into a aggressive drawback in opposition to corporations working below extra proportionate regulatory regimes. 

Whereas the EU remains to be contemplating the scope of copyright holders’ opt-outs and the way they need to work, personal corporations are already implementing their very own mechanisms, ready neither for regulation nor for the web site proprietor to truly make the selection. ​​If net hosts and platforms get to make such selections as an alternative of the customers that created the content material or the web sites that revealed it, copyright will slowly lose its that means.  

This won’t be due to AI corporations accessing content material however due to the dearth of management publishers could have over who can entry their content material. Each AI corporations and publishers are in search of readability over who can entry content material, for what functions and on whose phrases. These questions ought to be resolved by clear guidelines that give rights holders significant management, protect lawful entry to public net knowledge and maintain those that misuse it accountable. 

The online should stay open. This implies lawful data revealed overtly on the internet ought to stay accessible. Private, personal data ought to stay protected, whereas the gathering and use of private knowledge should adjust to data-protection legal guidelines. And that is the place the problem lies. We have to clearly distinguish what data is public and what data is deemed personal.  

Past these regulatory questions, the chance for EU startups is the creation of a pay-to-play financial system that favours giant know-how corporations and well-financed incumbents with negotiated entry to public knowledge. This raises the price of innovation and restricts European startups’ skill to catch up and compete, undermining the EU’s AI sovereignty ambitions. 

Enabling the EU’s AI sovereignty ambitions 

Making a industrial and regulatory surroundings that helps innovation and international competitiveness is central to growing the EU’s sovereign AI ecosystem. Current laws and ongoing coverage initiatives such because the AI Act present an necessary basis, however additional work is required to make clear compliance necessities and tackle obstacles that favour well-financed incumbents.  

Entry to the identical publicly out there data shouldn’t rely solely on an organization’s monetary assets or its skill to barter preferential offers. The EU must steadiness reputable entry to public net knowledge with the safety of copyright and platforms’ industrial pursuits. These goals can coexist, however builders and publishers want clearer guidelines about what data will be collected, for which functions and below what circumstances. 

Finally, the EU’s AI sovereignty ambitions rely not solely on funding in infrastructure and expertise but in addition on European corporations’ skill to entry and use numerous, up-to-date knowledge.



RELATED ARTICLES

Most Popular

Recent Comments