EU AI Lobbying 2026
TL;DR: Big Tech lobbying in Brussels reached €151 million annually in 2025, a 55% surge since 2021, with Meta leading at €10 million and Microsoft, Apple, and Amazon collectively spending €24 million. Corporate Europe Observatory documents reveal 890 full-time lobbyists now outnumber the 720 Members of European Parliament, holding 146 meetings with EU Commission officials in the first half of 2025 alone. This unprecedented influence campaign successfully watered down the EU AI Act’s Code of Practice, shifting “large-scale illegal discrimination” from mandatory systemic risks to optional considerations after concerted lobbying by Google and Microsoft. With OpenAI increasing U.S. lobbying six-fold to $1.9 million while simultaneously advocating for stricter European regulations, the divergence between public positioning and behind-the-scenes advocacy exposes how the world’s most powerful AI companies are reshaping regulatory frameworks that will govern artificial intelligence through 2030 and beyond.
The transformation of EU AI regulation represents more than a policy battle. It documents the most sophisticated corporate influence campaign in European legislative history, where technology companies deploy financial resources exceeding pharmaceutical, automotive, and financial industries combined to shape rules governing systems that will fundamentally alter democratic processes, labor markets, and individual rights across 27 member states and 450 million citizens.
The €151 Million Lobbying Machine Reshaping Brussels
In October 2025, Corporate Europe Observatory and LobbyControl published research documenting that the digital industry now spends €151 million annually on EU lobbying, representing a 33.6% increase from €113 million in 2023. This spending acceleration coincides precisely with the AI Act’s progression through implementation phases, creating what researchers describe as an “unprecedented lobby firepower” aimed at Europe’s digital rulebook.
The concentration of spending reveals the power dynamics shaping AI governance. Just ten companies account for €49 million of total expenditures, representing one-third of the digital sector’s lobbying budget. Meta leads with €10 million annually, making it the single largest corporate lobbyist in the entire European Union. Microsoft and Apple each spend €7 million, while Amazon invested €4.3 million more than its 2023 baseline, demonstrating commitment escalation as AI Act enforcement approaches.
These figures dwarf other industries known for lobbying influence. The top ten digital companies now spend three times as much as the top ten pharmaceutical corporations, twice the energy sector’s expenditure, and substantially more than automotive or financial industry leaders. This financial superiority provides structural advantages in policy access, technical expertise deployment, and sustained advocacy campaigns that traditional industries cannot match.
However, direct spending represents only the visible component of influence operations. Euronews reporting documents that Big Tech firms spend an additional €9 million annually on consultancies, PR firms, and think tanks to amplify messaging and provide seemingly independent validation for industry positions. Think tanks including Bruegel, Centre for European Reform, CEPS, and CERRE now receive funding from all five major digital corporations: Google, Meta, Apple, Amazon, and Microsoft.
The lobbying infrastructure extends beyond financial commitments. The number of “lobby actors” representing tech-friendly think tanks, associations, companies, and intermediaries in Brussels expanded from 565 in 2023 to 733 by mid-2025. This growth stems partially from AI-focused companies like Mistral AI and Aleph Alpha establishing Brussels presence, but also reflects strengthened EU transparency rules requiring companies meeting mid-level Commission officials to register publicly.
Most striking is the human capital deployment. The digital industry now employs 890 full-time lobbyists in Brussels, exceeding the 720 elected Members of European Parliament. Of these lobbyists, 437 hold accredited passes granting nearly unrestricted access to Parliament buildings, creating daily opportunities for influence that exceed democratic representatives’ capacity to counter-balance through constituent services or policy research.
Meta’s €10 Million Strategy: Leading the Deregulation Push
Meta’s position as the European Union’s largest corporate lobbyist reflects strategic calculation about AI Act implications for its business model. The company’s €10 million annual investment exceeds any single competitor and funds comprehensive advocacy campaigns spanning legislative, executive, and member state engagement.
Joel Kaplan, Meta’s Chief Global Affairs Officer, articulated the company’s position in July 2025 when announcing Meta would not sign the AI Act’s voluntary Code of Practice. In a LinkedIn statement, Kaplan declared the EU “is going down the wrong path on AI” and warned the Code introduces “legal uncertainties” and “measures which go far beyond the scope of the AI Act” that would “throttle the development and deployment of frontier AI models in Europe.”
This public opposition stands in contrast to OpenAI and Anthropic’s decision to sign the Code, creating a split among frontier AI developers that reflects divergent business models and regulatory risk assessments. Meta’s refusal carries material consequences because the company operates massive general-purpose AI systems including Llama models that serve European users, making non-compliance a strategic rather than operational choice.
The lobbying impact extends beyond Code rejection. Corporate Europe Observatory analysis documents that the second draft of the Code introduced a distinction between “systemic risks” and “additional risks for consideration” after concerted lobbying by Google and Microsoft. This categorization matters enormously because “systemic risks” trigger mandatory compliance obligations while “additional risks” remain essentially optional.
Documents obtained by CEO reveal that “large-scale illegal discrimination” initially categorized as a systemic risk requiring mandatory prevention was downgraded to an “additional risk for consideration” following industry pressure. This shift fundamentally alters the regulatory framework because it transforms discrimination prevention from a legal obligation to a voluntary best practice, despite documented cases of biased AI systems deployed in welfare programs, hiring processes, and law enforcement across EU member states.
Meta’s lobbying strategy also targets the broader digital regulatory framework. The company held meetings with Commission officials and MEPs focused on the Digital Services Act and Digital Markets Act, seeking “simplification” and burden relief that would reduce compliance costs across Meta’s platforms including Facebook, Instagram, WhatsApp, and Threads. These efforts align with statements from U.S. government officials including Secretary of State Marco Rubio, who in August 2025 called on American diplomats to undermine the Digital Services Act.
The company’s influence extends through trade associations. DIGITALEUROPE, whose members include Meta alongside Microsoft, Google, Amazon, and Apple, increased its lobbying budget by €1.25 million to amplify industry messaging through a seemingly neutral business association. This layered approach enables Meta to advocate directly while also supporting collective industry positions that provide political cover for individual company stances.
Microsoft and Google: The €7 Million Giants Shaping Technical Standards
Microsoft and Apple each spend €7 million annually on EU lobbying, but Microsoft’s influence extends beyond direct expenditure through its position as both AI developer and cloud infrastructure provider. The company’s Azure platform hosts OpenAI’s models, creating aligned interests in regulatory frameworks favoring general-purpose AI providers over application-layer companies deploying AI systems.
Kent Walker, Google’s president of global affairs, publicly committed to signing the AI Act Code of Practice while simultaneously warning it risks “slowing down Europe’s development and deployment of AI.” This rhetorical balance allows Google to maintain compliance posture while lobbying for implementation approaches that minimize competitive disadvantages relative to U.S. market advantages.
The technical standards battleground represents where Microsoft and Google exert disproportionate influence. Corporate Europe Observatory research exposed how the AI Office relied on external consultancies Wavestone and Intellera to draft Code of Practice provisions, despite these firms having direct commercial ties to Microsoft. Wavestone received a “Microsoft Partner of the Year Award” in 2024 while simultaneously supporting the AI Office in developing AI regulations, creating conflicts of interest that enabled privileged access to regulatory drafting processes.
This structural advantage matters because harmonized standards will govern high-risk AI systems across the EU. Companies participating in standards development can ensure technical requirements align with their existing systems and development practices, creating de facto compliance advantages that smaller competitors cannot match. When consultancies with Microsoft commercial relationships draft these standards, the resulting framework inherently favors incumbent architectures over alternative approaches.
Google’s lobbying particularly targets general-purpose AI model provisions. TIME Magazine reporting documented that Google, alongside Microsoft and OpenAI, successfully lobbied to ensure the final AI Act did not classify general-purpose AI systems as inherently high-risk. This distinction shifted regulatory burden from model providers like Google (which develops Gemini) to downstream application developers deploying AI for specific use cases.
Both companies deployed lobbyists to meet with European Parliament’s European People’s Party (EPP), which Computing.co.uk reporting notes held a disproportionate number of meetings with Big Tech representatives. This partisan targeting reflects sophisticated political strategy recognizing that center-right MEPs are more receptive to competitiveness arguments than Green or Social Democratic counterparts emphasizing rights protection and environmental safeguards.
The companies’ lobbying also addresses copyright frameworks under the AI Act. Both Google and Microsoft advocate for broad fair use interpretations enabling training on copyrighted material without explicit consent, arguing restrictive copyright rules stifle innovation. This position directly opposes rights holders including publishing organizations, music labels, and authors’ associations that argue AI training constitutes unauthorized reproduction requiring licensing arrangements.
Amazon’s €4.3 Million Surge: Cloud Infrastructure and AI Services
Amazon increased its EU lobbying expenditure by €4.3 million between 2023 and 2025, representing the largest single company budget escalation documented by Corporate Europe Observatory. This surge reflects Amazon Web Services’ (AWS) strategic position as the dominant cloud infrastructure provider for AI workloads, creating incentives to shape regulations affecting both AI development and deployment environments.
AWS hosts AI systems developed by Anthropic (in which Amazon invested $4 billion), Stability AI, Hugging Face, and thousands of smaller companies deploying models through Amazon Bedrock and SageMaker services. Regulations affecting model transparency requirements, data governance frameworks, or liability structures for AI-generated outputs impact AWS’s competitive position relative to Microsoft Azure and Google Cloud Platform.
Amazon’s lobbying particularly targets the AI Act’s provisions on high-risk AI systems in areas like employment, credit scoring, and law enforcement. AWS provides infrastructure for companies deploying AI in these domains, making compliance frameworks directly relevant to Amazon’s service offerings. More permissive rules reduce customer compliance costs and maintain AWS’s market leadership, while strict requirements could advantage competitors offering compliance-as-a-service capabilities.
The company also participates in Computer and Communications Industry Association (CCIA) lobbying campaigns. In October 2025, CCIA launched a campaign pushing for simplification not only of the AI Act but of the EU’s entire digital rulebook, representing Apple, Meta, Amazon, and Google positions. This collective advocacy enables individual companies to distance themselves from controversial positions while still benefiting from association-led pressure on EU institutions.
Amazon’s lobbying extends to member state governments. The company meets with national authorities in Germany, France, Netherlands, and Ireland to advocate for favorable AI Act implementation approaches, recognizing that member states retain flexibility in enforcement priorities and technical interpretation of AI Act provisions. This multi-level strategy addresses both EU-level regulation and national implementation variations that could fragment the single market.
OpenAI’s Strategic Duplicity: Public Support, Private Opposition
OpenAI’s lobbying strategy represents perhaps the most sophisticated example of divergence between public positioning and private advocacy. CEO Sam Altman has spent 2025 touring world capitals delivering speeches emphasizing the need for global AI regulation and expressing commitment to safety-oriented governance frameworks. These public statements receive extensive media coverage and position OpenAI as a responsible AI leader supporting regulatory oversight.
However, documents obtained by TIME Magazine through freedom of information requests reveal that OpenAI lobbied EU officials in September 2022 to water down AI Act provisions that would classify general-purpose AI systems as inherently high-risk. In a seven-page white paper titled “OpenAI White Paper on the European Union’s Artificial Intelligence Act,” the company argued that “by itself, GPT-3 is not a high-risk system” but rather “possesses capabilities that can potentially be employed in high risk use cases.”
This framing successfully shifted regulatory burden from OpenAI to downstream application developers, meaning companies deploying GPT-4 or ChatGPT for hiring decisions, credit scoring, or law enforcement applications bear primary compliance responsibility rather than OpenAI itself. The final AI Act incorporated this framework, requiring foundation model providers to comply with limited transparency and documentation requirements while imposing stricter obligations on high-risk AI system deployers.
OpenAI’s U.S. lobbying spending increased nearly seven-fold in 2024, reaching $1.9 million according to MIT Technology Review analysis. The company hired Matthew Rimkunas, a former lobbyist for Senator Lindsey Graham who worked on nuclear safety issues, reflecting OpenAI’s strategic pivot toward energy infrastructure advocacy essential for training increasingly large models.
This energy focus culminated in September 2025 meetings where Altman, alongside leaders from Nvidia, Anthropic, and Google, visited the White House pitching the vision that U.S. competitiveness in AI depends on subsidized energy infrastructure. Altman proposed constructing multiple five-gigawatt data centers consuming as much electricity as New York City, requiring government support for nuclear power development and transmission infrastructure expansion.
Despite signing the EU AI Act Code of Practice in July 2025, OpenAI’s public commitment came only after the company successfully lobbied to weaken Code provisions during drafting processes. Documents show OpenAI participated in dedicated workshops with working group chairs developing the Code, enjoying privileged access that civil society organizations lacked. Several stakeholders including Reporters Without Borders withdrew from consultation processes citing overwhelming Big Tech influence.
The company’s lobbying also targets copyright frameworks. OpenAI, alongside Google, lobbies the Trump administration to classify AI training on copyrighted data as fair use, framing such training as essential for national security and maintaining competitive advantage over China. This position directly contradicts author lawsuits alleging OpenAI illegally scraped copyrighted books, including leaked documents from early 2025 revealing systematic harvesting of literary works without permission or compensation.
Anthropic’s Contrarian Position: Signing the Code While Opposing Federal Preemption
Anthropic presents a distinct lobbying strategy compared to OpenAI, Meta, Microsoft, and Google. The company signed the EU AI Act Code of Practice and publicly supports the framework, stating the Code “advances the principles of transparency, safety and accountability” that Anthropic champions for frontier AI development. This public support aligns with the company’s positioning as a safety-focused AI developer distinguishing itself from competitors prioritizing rapid deployment.
However, Anthropic’s U.S. lobbying reveals tensions in this positioning. The company spent over $1 million on lobbying in Q3 2025 for the first time, according to Axios reporting, focusing on digital assets, AI policy, financial technology, and open-source AI frameworks. This spending increase coincides with Anthropic raising billions in venture funding and scaling Claude model development, creating regulatory interests distinct from established incumbents.
Anthropic diverges from industry consensus on federal preemption of state AI laws. While Meta, OpenAI, Google, and Microsoft support federal legislation preventing states from enacting independent AI regulations, LessWrong analysis documents that Anthropic opposes preemption absent adequate federal framework. This position reflects concern that preemption without robust federal requirements would create regulatory vacuum enabling harmful AI deployments without meaningful oversight.
The company’s position on California’s SB 1047 further illustrates this approach. Anthropic submitted a letter to Governor Newsom stating the revised bill’s “benefits likely outweigh its costs” but expressing uncertainty and noting “concerning or ambiguous” aspects. This measured support contrasts with industry groups including CCIA that opposed the legislation entirely, positioning Anthropic as willing to accept reasonable safety-oriented regulation.
However, critics note that Anthropic’s public safety positioning coexists with business practices raising questions about implementation sincerity. The company faces litigation over downloading digital books from online pirate libraries to train Claude models, mirroring allegations against OpenAI and other foundation model providers. A federal court in California ordered Anthropic to stand trial over these practices, challenging claims that the company operates differently from competitors on copyright respect.
Anthropic’s lobbying through trade associations also complicates its independent positioning. The company participates in AI Alliance, which represents Meta interests, and engages with industry groups opposing state-level AI regulation including New York’s Responsible AI Safety and Education Act. Anthropic policy lead Jack Clark publicly criticized this legislation, demonstrating alignment with industry opposition despite the company’s stated support for thoughtful AI governance.
European AI Companies: Mistral AI and Aleph Alpha Lobby Member States
European-headquartered AI companies pursue lobbying strategies targeting both EU institutions and national governments, leveraging domestic political relationships that U.S. companies cannot replicate. France’s Mistral AI and Germany’s Aleph Alpha led efforts resulting in 56 EU-based AI companies signing a July 2025 public letter urging the Commission to pause and simplify parts of the AI Act.
This coalition creates political cover for U.S. companies’ deregulation advocacy by demonstrating that European firms also view the AI Act as competitively disadvantageous. TechPolicy.Press reporting documents that Mistral AI and Aleph Alpha convinced their home governments to advocate for regulatory simplification, with France driving changes weakening earlier Parliament positions on biometric identification bans.
Mistral AI’s lobbying particularly emphasizes competitiveness arguments. The company argues that European AI development lags U.S. and Chinese investment levels, citing data showing the EU accounted for only 7% of global AI investment in 2021 compared to 40% for the United States and 32% for China. According to European Parliament research, Europe invested approximately €5 billion in AI in 2023 compared to €20 billion in the U.S., creating competitive disadvantage that strict regulation would exacerbate.
These arguments resonate with member state governments facing domestic pressure to support national AI champions. France’s government backed Mistral AI’s positions despite the company receiving substantial public funding, creating tensions between industrial policy supporting domestic companies and consumer protection frameworks limiting those companies’ practices. Germany similarly supported Aleph Alpha’s advocacy for regulatory simplification, reflecting Chancellor Olaf Scholz’s technology competitiveness agenda.
The European companies’ lobbying also exploits institutional divisions within the EU. While the European Parliament initially advocated strict AI Act provisions including broad biometric identification bans, member state governments led by France pushed for national security exemptions. Article 2 of the final AI Act placed national security uses outside the law’s scope, effectively permitting EU governments to deploy AI systems for mass surveillance at protests or borders, representing a rollback from Parliament’s stricter stance.
Aleph Alpha particularly lobbied for exemptions benefiting military and defense applications. The company develops AI systems for German government use, creating business incentives to ensure the AI Act does not constrain defense-related deployments. These exemptions extend to private companies and potentially third countries providing AI technology to police and law enforcement agencies, creating loopholes that civil society organizations argue undermine the Act’s fundamental rights protections.
The 146 Meetings: Daily Lobbying Access to EU Commission Officials
Corporate Europe Observatory documentation reveals Big Tech companies held 146 lobbying meetings with EU Commission officials in the first half of 2025, averaging more than one meeting per day. This access level exceeds any other industry sector and provides continuous opportunities to shape regulatory implementation approaches during critical AI Act enforcement preparation phases.
The meetings occurred across Directorates-General including CONNECT (communications networks), GROW (internal market), and JUST (justice and consumers), reflecting the AI Act’s cross-cutting implications. Companies deployed technical experts alongside policy professionals to these meetings, enabling detailed discussions of implementation approaches that Commission officials, often lacking equivalent AI systems expertise, struggle to independently evaluate.
The Good Lobby investigation through Freedom of Information requests to seven European governments revealed that many national authorities refuse to disclose details of meetings with tech companies to discuss AI regulation. Requests to Bosnia and Herzegovina, Denmark, Germany, Hungary, Ireland, Spain, and the United Kingdom resulted in “minimal disclosures, vague justifications, and in some cases, outright refusals,” preventing public accountability for decision-making processes.
This lack of transparency creates information asymmetry where companies know Commission and member state positions while citizens cannot access equivalent information. The Good Lobby researchers note this prevents assessing the extent of Big Tech influence on AI regulation because governments operate “under a veil of secrecy” regarding what is discussed, which proposals are made, and how these interactions shape national policies.
The meeting frequency also enables relationship building that shapes regulatory culture. Commission officials meeting daily with industry representatives develop familiarity with company positions, technical constraints, and business models that inform their regulatory interpretation even absent explicit lobbying asks. This socialization effect means that industry perspectives become embedded in regulatory thinking through repeated exposure rather than through documented position papers.
The European People’s Party’s disproportionate meetings with Big Tech lobbyists documented by Corporate Europe Observatory reveals partisan dynamics in lobbying access. Center-right MEPs more readily accept industry competitiveness arguments, creating strategic targeting opportunities for companies that focus advocacy resources on receptive political factions rather than distributing access requests evenly across the political spectrum.
Trump Administration Pressure: External Leverage Amplifying Corporate Lobbying
The Trump administration’s return to power in January 2025 created external pressure on EU institutions that amplifies Big Tech lobbying effectiveness. Secretary of State Marco Rubio called on American diplomats in August 2025 to undermine the Digital Services Act, explicitly framing EU regulation as harmful to U.S. commercial interests requiring government intervention.
President Trump warned EU leaders in early 2025 to “show respect to America and our amazing tech companies or consider the consequences,” threatening tariffs on countries whose technology regulations harm U.S. companies. This linkage between trade policy and technology regulation creates diplomatic pressure that individual companies cannot generate independently, providing leverage that corporate lobbying alone would not achieve.
Carnegie Endowment analysis notes that U.S. government pressure coincides with Big Tech’s internal policy shifts responding to Trump’s regulatory priorities. Google reversed its policy against military AI applications, leading to internal resignations over the company’s ethical direction. OpenAI similarly announced it would develop AI models with defense-tech companies after previously refusing military contracts, aligning with Trump administration priorities framing AI as national security essential.
This government-corporate alignment creates coordinated pressure that EU institutions struggle to resist. When both the U.S. government and major technology companies simultaneously advocate for EU regulatory simplification, member states face not only competitiveness concerns but also geopolitical calculations about transatlantic alliance maintenance and trade relationship preservation.
The March 2025 revelation that the U.S. is willing to leverage control over Starlink satellite communications to pressure Ukraine demonstrated Europe’s technological dependency on systems controlled by single U.S. companies. This vulnerability extends beyond communications to cloud infrastructure, AI models, and operating systems where European alternatives lack competitive parity, creating strategic weaknesses that U.S. government and corporate actors can exploit.
Some EU lawmakers warned against weakening legislation to appease tech firms and the U.S. administration, but Euronews reporting indicates many express dissatisfaction with Tech Commissioner Henna Virkkunen’s responses to Trump’s threats. The Commission’s November 2025 digital simplification package includes possible burden relief for AI Act compliance, suggesting U.S. pressure combined with industry lobbying is achieving regulatory rollback despite public commitment to enforcement.
The Code of Practice Drafting: Structural Advantages for Model Providers
The AI Act’s Code of Practice on general-purpose AI represents where lobbying influence translated most directly into regulatory outcomes. Corporate Europe Observatory and LobbyControl research documented that companies developing AI models enjoyed structural advantages throughout the drafting process that civil society organizations lacked.
Model providers including Meta, Microsoft, Google, OpenAI, and Anthropic received invitations to dedicated workshops with working group chairs developing Code provisions. These closed-door sessions enabled direct input into regulatory language that companies would subsequently be required to follow, creating fox-guarding-henhouse dynamics where regulated entities drafted their own oversight frameworks.
The AI Office’s reliance on external consultancies Wavestone and Intellera further advantaged incumbents. As noted, Wavestone received Microsoft Partner awards while simultaneously supporting Code development, creating conflicts of interest that the Office failed to adequately address. Corporate Europe Observatory filed an Ombudsman complaint in June 2025 over the Commission’s decision to hire consultancies with direct commercial interests in AI markets.
The structural advantages manifested in specific provisions. The second draft introduced a split between “systemic risks” triggering mandatory obligations and “additional risks for consideration” that companies could optionally address. This categorization reflected intensive lobbying by Google and Microsoft, according to documents obtained by CEO, enabling companies to avoid binding requirements around discrimination, bias, and fairness issues.
Civil society withdrawal from consultation processes further tilted the balance. Reporters Without Borders and other organizations departed citing overwhelming Big Tech influence, leaving model providers as the dominant voices in remaining technical discussions. This exodus meant that rights-focused perspectives disappeared from drafting processes at precisely the moment when detailed implementation approaches were being determined.
The voluntary nature of the Code creates compliance asymmetry. Companies signing the Code benefit from legal certainty about compliance pathways, while those refusing (like Meta) remain subject to AI Act obligations but can develop alternative compliance approaches. This optionality enables strategic non-participation by companies viewing Code requirements as more burdensome than developing independent compliance frameworks.
However, the August 2025 Code finalization with signatures from OpenAI, Anthropic, Google, Microsoft, Amazon, and others demonstrates that industry successfully shaped provisions to acceptable levels. If the Code imposed genuinely burdensome requirements, companies would not voluntarily adopt it. The willingness to sign indicates lobbying achieved regulatory frameworks companies view as manageable rather than restrictive.
Member State Divisions: France, Germany, and Denmark Diverge
EU member states demonstrate divergent positions on AI Act simplification reflecting domestic political calculations and economic interests. These divisions create opportunities for lobbying campaigns that exploit institutional fragmentation by targeting receptive governments while bypassing skeptical authorities.
France, led by President Emmanuel Macron, advocated strongly for national security exemptions that Article 2 incorporated into the final AI Act. This positioning reflects France’s strategic autonomy agenda emphasizing European technological independence while simultaneously protecting French security services’ operational flexibility to deploy AI surveillance systems. French support for Mistral AI’s lobbying positions creates alignment between industrial policy supporting domestic companies and regulatory simplification benefiting those companies.
Germany initially opposed fundamental alterations to the AI Act’s structure, arguing in a “non-paper” sent to Brussels that the law must retain its original scope and integrity. However, Germany simultaneously supports Aleph Alpha’s advocacy for regulatory burden reduction, creating tensions between defending the Act’s framework while seeking implementation flexibility benefiting German companies. This ambivalence enables selective support for simplification measures that advantage domestic firms while maintaining rhetorical commitment to strict regulation.
Denmark advocates aggressively for “genuine simplification” determining what “should be retained, revised, or repealed” across the digital framework. Danish officials argue the AI Act is overly complex and requires streamlining beyond symbolic reform, positioning Denmark as the most receptive major member state to industry deregulation arguments. This stance partly reflects Denmark’s limited domestic AI industry, reducing political costs of supporting U.S. company positions.
The Netherlands takes a more balanced position supporting efforts to reduce regulatory burdens while defending the digital rulebook’s overarching goals. Dutch pragmatism recognizes that excessive regulation could handicap European companies relative to U.S. and Chinese competitors, but maintains that fundamental rights protections and consumer safeguards justify compliance costs that industry lobbying characterizes as excessive.
These divergences enable lobbying strategies targeting member states individually rather than addressing EU institutions monolithically. Companies meeting with French authorities emphasize competitiveness and strategic autonomy arguments, while Danish engagement focuses on regulatory complexity and administrative burden. This tailored approach exploits member state divisions to build coalitions supporting industry positions within Council negotiations.
The European Commission’s November 2025 Digital Omnibus package proposing AI Act amendments reflects member state pressure as much as industry lobbying. TechPolicy.Press analysis notes that ongoing discussions between EU officials and the Trump administration around digital rules adjustments create diplomatic dynamics reinforcing domestic competitiveness concerns that member states bring to Council deliberations.
The Draghi Report: Policy Cover for Deregulation Advocacy
Mario Draghi’s landmark 2024 EU competitiveness report provided industry lobbying with authoritative policy cover for deregulation arguments. The former European Central Bank president and Italian Prime Minister concluded that “onerous” regulatory barriers to innovation in the tech sector, including the AI Act, hamper EU competitiveness and require simplification.
This high-level endorsement of regulatory burden concerns enables industry advocates to cite an influential European voice rather than relying solely on U.S. company position papers. When Draghi argues that excessive regulation disadvantages European innovation, tech lobbyists can frame their advocacy as aligned with EU strategic interests rather than corporate profit maximization.
However, critics note that the regulatory overreach narrative is “largely a strategic construct promoted by U.S. actors rather than an objective reality,” according to Carnegie Endowment research. European experts caution that competitiveness arguments often lack empirical grounding, instead reflecting industry messaging that regulatory simplification inherently drives innovation despite evidence that consumer protection frameworks can enhance market trust and adoption.
The Draghi Report’s influence extends beyond AI regulation to the broader digital rulebook including the Digital Markets Act and Digital Services Act. Industry advocates invoke the report to argue for comprehensive simplification rather than AI-specific adjustments, creating political momentum for deregulation across multiple frameworks simultaneously. This bundling strategy amplifies lobbying effectiveness by framing individual regulatory changes as components of broader competitiveness enhancement rather than isolated concessions to industry pressure.
Commission President Ursula von der Leyen’s embrace of competitiveness rhetoric in her second term further legitimizes industry positions. The Commission’s 2025 work program signaled a striking policy shift toward deregulation, exemplified by the abrupt cancellation of the proposed AI liability directive that would have established provisions for civil liability for AI-generated damages. This cancellation removed protections for individuals harmed by AI systems that industry lobbying characterized as innovation barriers.
Enforcement Reality: Fines Proceeding Despite Lobbying Pressure
Despite intensive lobbying and political pressure for simplification, EU enforcement actions against Google, Meta, TikTok, and other companies proceed as scheduled, demonstrating limits to industry influence over Commission regulatory implementation. Investigations and fines under the Digital Services Act and Digital Markets Act continue, suggesting that enforcement officials maintain independence from political pressure affecting legislative processes.
The Commission’s resistance to “stop the clock” proposals reflects institutional commitment to AI Act implementation timelines despite industry advocacy for delays. When companies and some member states proposed pausing enforcement to allow additional consultation and adjustment, the Commission publicly rejected these calls, stating it is not considering implementation freezes even as it develops burden relief measures.
However, the November 2025 digital simplification package introduces concerning precedents. While formal enforcement timelines remain intact, burden relief provisions could create practical exemptions enabling companies to claim compliance while avoiding substantive obligations. The distinction between simplification reducing genuine administrative overhead versus creating loopholes undermining regulatory effectiveness will determine whether enforcement proceeds meaningfully.
Computing.co.uk reporting notes that MEPs warned against weakening key legislation to appease tech firms and the U.S. administration, with some expressing dissatisfaction at the lack of forceful responses to Trump’s threats. This parliamentary resistance creates political counterweight to Commission simplification initiatives, suggesting enforcement approaches will face scrutiny from elected representatives skeptical of industry-friendly implementation.
The success of lobbying campaigns to weaken the AI Act Code of Practice demonstrates that regulatory outcomes reflect power dynamics rather than objective policy analysis. When companies enjoy privileged access to drafting processes, hire consultancies with commercial ties to regulated entities, and deploy resources exceeding other stakeholders combined, resulting regulations inevitably favor industry positions regardless of formal enforcement commitment.
The Copyright Battlefield: Training Data and Rights Holder Resistance
Copyright frameworks under the AI Act represent perhaps the most contentious lobbying battleground because they directly implicate foundation model training practices that companies view as commercially essential. The EU’s opt-out system for copyrighted content creates legal uncertainty that both AI companies and rights holders seek to resolve through regulatory influence.
AI developers including OpenAI, Google, and Anthropic lobby for broad fair use interpretations enabling training on copyrighted material without explicit consent or licensing payments. These companies argue that restrictive copyright rules stifle innovation and disadvantage European companies relative to U.S. and Chinese competitors operating under more permissive frameworks. OpenAI and Google specifically lobby the Trump administration to classify AI training on copyrighted data as fair use, framing such training as essential for national security and competitive advantage over China.
However, this position faces organized resistance from rights holders including publishers, music labels, authors’ associations, and creators’ organizations. Sony Music Group and Warner Music Group sent letters to AI companies explicitly stating they do not consent to having their music or lyrics used for training, activating the opt-out mechanism the AI Act established. Similar notices from book publishers, newspapers, and visual artists create legal compliance obligations that companies lobbying for permissive frameworks seek to minimize.
Litigation provides the enforcement mechanism for copyright claims that lobbying cannot fully neutralize. Anthropic faces trial in California federal court over downloading digital books from online pirate libraries, while Meta faced lawsuits from authors after leaked documents revealed covert scraping of copyrighted books. These legal proceedings create judicial precedents that could override lobbying success in establishing favorable regulatory frameworks.
Attorney Maxwell Pritt, litigating multiple copyright cases against AI companies, testified to Congress that AI developers engaged in “what is likely the largest domestic piracy of intellectual property in [U.S.] history.” This characterization frames the issue as criminal conduct rather than regulatory ambiguity, raising reputational risks that companies cannot address through Brussels lobbying alone.
The Code of Practice includes copyright transparency provisions requiring companies to publish summaries of content used to train models. However, the level of detail required remains contested, with industry lobbying for high-level disclosures while rights holders demand granular information enabling verification that opt-out requests were respected. Implementation guidance will determine whether transparency requirements provide meaningful accountability or symbolic compliance.
2026 Outlook: Digital Omnibus and Continued Lobbying Escalation
The Commission’s November 2025 Digital Omnibus package proposing AI Act amendments represents the immediate policy battleground for 2026. This comprehensive reform addresses regulatory burden concerns while attempting to preserve the Act’s fundamental rights protections, creating tension between industry demands for simplification and civil society advocacy for robust enforcement.
Member states remain divided on how far reforms should extend. France and Denmark push for substantial changes potentially altering the Act’s structure, while Germany warns against fundamental modifications undermining legal certainty. The Netherlands seeks balance between burden reduction and goal preservation, positioning itself as a swing vote in Council negotiations. These divisions create opportunities for continued lobbying campaigns targeting individual governments rather than addressing EU institutions collectively.
The Trump administration’s ongoing engagement with EU officials around digital rules adjustments ensures external pressure will persist throughout 2026. Trade policy linkages between technology regulation and tariff threats create diplomatic leverage that individual companies cannot generate, amplifying lobbying effectiveness through government-corporate coordination. EU officials report discussions around digital rules modifications, suggesting Washington’s pressure influences Brussels policymaking beyond direct industry advocacy.
Lobbying spending will likely escalate further as AI Act enforcement begins and companies face concrete compliance obligations. The €151 million annual expenditure documented through September 2025 represents only the visible lobbying disclosed in transparency registers, excluding additional millions flowing through think tanks, academic sponsorships, and industry associations that amplify corporate messaging without direct attribution.
The expansion of AI industry presence in Brussels from 565 lobby actors in 2023 to 733 by mid-2025 suggests continued growth as more companies recognize regulatory outcomes will shape competitive dynamics for decades. Companies that earlier delegated Brussels advocacy to trade associations now establish direct representation recognizing that industry-wide positions may not align with firm-specific interests.
However, resistance to deregulation pressure also intensifies. Civil society organizations, rights holders, and consumer advocates recognize that lobbying campaigns threaten to undermine the AI Act before it meaningfully constrains harmful AI deployments. Parliamentary skepticism of Commission simplification initiatives creates political counterweight, though MEPs lack the technical expertise and resources to match industry’s regulatory influence capacity.
The fundamental question for 2026 is whether the EU will maintain regulatory frameworks protecting fundamental rights and consumer interests despite unprecedented industry pressure, or whether lobbying campaigns successfully transform the AI Act into symbolic legislation creating compliance burdens without substantive constraints on harmful AI applications.
FAQ: EU AI Lobbying and Regulatory Influence
How much money do tech companies spend lobbying the EU on AI regulation?
The digital industry now spends €151 million annually on EU lobbying as of 2025, representing a 55% increase from €113 million in 2021. Just ten companies account for €49 million of this spending, with Meta leading at €10 million, Microsoft and Apple at €7 million each, and Amazon at over €4 million. These figures dwarf other industries: the top ten digital companies spend three times as much as pharmaceutical firms, twice the energy sector, and substantially more than automotive or financial industries. Additionally, companies spend over €9 million annually on consultancies, PR firms, and think tanks to amplify their messaging.
Which companies have the most lobbying influence on EU AI regulation?
Meta is the single largest corporate lobbyist in the European Union with €10 million in annual spending, followed by Microsoft and Apple at €7 million each. Google, Amazon, and Qualcomm also rank among top spenders. However, influence extends beyond direct expenditure. OpenAI, despite smaller budgets, secured privileged access to AI Act drafting processes and successfully lobbied to prevent general-purpose AI systems from being classified as inherently high-risk. The 890 full-time lobbyists now working in Brussels exceed the 720 elected Members of European Parliament, with 437 holding accredited passes granting nearly unrestricted Parliament access.
Did tech companies successfully weaken the EU AI Act?
Yes, in multiple substantive ways. Corporate Europe Observatory documents reveal that the second draft of the Code of Practice split risks into “systemic risks” with mandatory obligations and “additional risks for consideration” that are essentially optional, following concerted lobbying by Google and Microsoft. Large-scale illegal discrimination was downgraded from systemic to additional risk. TIME Magazine reporting documents that OpenAI successfully lobbied to ensure the final AI Act did not classify general-purpose AI systems as inherently high-risk, shifting regulatory burden to downstream application developers. The Commission also cancelled the proposed AI liability directive that would have established civil liability for AI-generated damages.
How many meetings do tech companies have with EU officials?
Big Tech companies held 146 lobbying meetings with EU Commission officials in just the first half of 2025, averaging more than one meeting per day. These meetings occurred across multiple Directorates-General and provided continuous opportunities to shape regulatory implementation. However, national government meetings remain largely secret. Freedom of Information requests to seven European governments revealed minimal disclosures, vague justifications, and outright refusals to disclose details of meetings with tech companies discussing AI regulation, preventing public accountability for decision-making processes.
Which companies signed the EU AI Code of Practice and which refused?
OpenAI, Anthropic, Google, Microsoft, Amazon, IBM, and Mistral AI signed the Code of Practice as of August 2025. More than 25 major providers voluntarily adopted the framework. However, Meta refused to sign, with Chief Global Affairs Officer Joel Kaplan stating the Code introduces legal uncertainties and measures going “far beyond the scope of the AI Act” that would “throttle frontier AI model development in Europe.” xAI signed only the safety and security chapter rather than the complete Code. Companies declining to sign remain subject to AI Act obligations but can develop alternative compliance approaches.
How does U.S. government pressure affect EU AI regulation?
The Trump administration creates external leverage amplifying corporate lobbying effectiveness. Secretary of State Marco Rubio called on American diplomats to undermine the Digital Services Act, while President Trump threatened tariffs on countries whose technology regulations harm U.S. companies. This linkage between trade policy and technology regulation creates diplomatic pressure that individual companies cannot generate independently. EU officials report ongoing discussions around digital rules adjustments, suggesting Washington’s threats influence Brussels policymaking. The March 2025 revelation that the U.S. is willing to leverage Starlink control to pressure Ukraine demonstrated Europe’s technological dependency vulnerability.
What role do think tanks play in EU AI lobbying?
Tech companies spend over €9 million annually on think tanks to provide seemingly independent validation for industry positions. Bruegel, Centre for European Reform, CEPS, and CERRE now receive funding from all five major digital corporations: Google, Meta, Apple, Amazon, and Microsoft. This layered approach enables companies to advocate directly while also supporting “neutral” research validating their policy arguments. Think tanks provide academic credibility and expert voices that industry position papers cannot achieve, creating the appearance of broad consensus supporting regulatory simplification when underlying funding sources reveal commercial interests.
How did consultancies with Microsoft ties draft EU AI regulations?
The AI Office relied on external consultancies Wavestone and Intellera to help draft the Code of Practice, despite both firms having direct commercial ties to Big Tech. Wavestone received a “Microsoft Partner of the Year Award” in 2024 while simultaneously supporting the AI Office in developing regulations, creating conflicts of interest that enabled privileged industry access to regulatory drafting processes. Corporate Europe Observatory filed an Ombudsman complaint in June 2025 over the Commission’s decision to hire consultancies with commercial interests in AI markets, arguing this compromised the Code’s independence and fairness.
What is the European People’s Party’s role in tech lobbying?
The European People’s Party (EPP) held a disproportionate number of meetings with Big Tech lobbyists according to Corporate Europe Observatory reporting. This partisan targeting reflects sophisticated lobbying strategy recognizing that center-right MEPs are more receptive to competitiveness arguments than Green or Social Democratic counterparts emphasizing rights protection. Companies focus advocacy resources on receptive political factions rather than distributing access evenly across the political spectrum, exploiting ideological divisions to build coalitions supporting industry positions within Parliament.
How do European AI companies like Mistral AI lobby differently than U.S. firms?
European companies leverage domestic political relationships that U.S. firms cannot replicate. Mistral AI and Aleph Alpha convinced their home governments (France and Germany respectively) to advocate for regulatory simplification, creating political cover for U.S. companies’ deregulation advocacy by demonstrating European firms also view the AI Act as competitively disadvantageous. They led 56 EU-based AI companies in signing a July 2025 public letter urging the Commission to pause and simplify parts of the AI Act. France particularly drove changes weakening earlier Parliament positions on biometric identification bans, reflecting Mistral AI’s influence on national policy.
What happens to companies that don’t sign the Code of Practice?
Companies refusing to sign (like Meta) remain subject to AI Act obligations but can develop alternative compliance approaches rather than following the Code’s voluntary framework. However, they lose the legal certainty that Code signatories enjoy about compliance pathways. The EU AI Act establishes that for general-purpose AI model providers, the Commission may impose fines of up to €15 million or 3% of worldwide annual turnover for non-compliance. For general AI systems, fines reach €35 million or 7% of global annual turnover. The voluntary Code provides a structured pathway demonstrating compliance, while non-signatories must independently prove they meet legal obligations.
How does lobbying affect copyright protections under the AI Act?
AI developers lobby for broad fair use interpretations enabling training on copyrighted material without explicit consent or licensing payments. OpenAI and Google specifically lobby the Trump administration to classify AI training on copyrighted data as fair use, framing it as essential for national security. However, rights holders including Sony Music, Warner Music, publishers, and authors’ associations actively resist, sending explicit notices that they do not consent to content use for training. The Code of Practice includes copyright transparency provisions requiring disclosure of training content, but the level of detail remains contested. Litigation creates enforcement mechanisms that lobbying cannot fully neutralize, with Anthropic facing trial over downloading copyrighted books.
Will EU AI regulation be weakened further in 2026?
The Commission’s November 2025 Digital Omnibus package proposes AI Act amendments addressing burden concerns while attempting to preserve fundamental rights protections. Member states remain divided, with France and Denmark pushing substantial changes potentially altering the Act’s structure, while Germany warns against modifications undermining legal certainty. Lobbying spending will likely escalate as AI Act enforcement begins, suggesting continued deregulation pressure. However, parliamentary resistance and civil society advocacy create counterweight. The fundamental question is whether the EU maintains frameworks protecting rights despite unprecedented industry pressure, or whether lobbying transforms the AI Act into symbolic legislation without substantive constraints on harmful AI applications.
