The Essential Role of Open Source in Sovereign AI

This blog was first published on Oct 23, 2025 at https://blog.irvingwb.com/blog/2025/10/the-essential-role-of-open-source-in-sovereign-ai.html and repurposed here with consent from the author. 

A new report, “The State of Sovereign AI: Exploring the Role of Open Source Projects and Global Collaboration in Global AI Strategy,” was recently published by Linux Foundation (LF) Research, LF AI & Data, and Futurewei Technologies. “The term ‘sovereign AI’ has been used to describe efforts aimed at developing AI capabilities with minimal reliance on external actors, enabling nations and organizations to retain control over their systems, data, and decision-making processes,” said the report in its Introduction. “Sovereign AI initiatives seek to address concerns about data sovereignty, national security, economic competitiveness, and cultural alignment by developing domestically controlled AI capabilities that can operate independently of external technology providers and geopolitical constraints.”

How do we steer AI rather than be steered by it? “is one question dominating boardroom and governments alike,” wrote Mark Collier, General Manager of LF AI & Infrastructure in the report’s Foreword. “This report responds to that question with clarity and evidence: open source is the answer. Nearly four out of five organizations call AI sovereignty a strategic priority, and 90% cite open source as essential to achieving it.”

Let me summarize the key points in each of the report’s sections.

Executive summary

The report is based on a survey of 233 respondents as well as on insights from industry leaders. Almost 80% or respondents consider sovereign AI a strategic priority for nation and organizations. “This consensus spans major geographic regions, with 86% of U.S. respondents, 83% of European respondents, and 79% of Asia-Pacific respondents viewing it as essential. The strategic importance manifests at both national (66%) and organizational (47%) levels, with 82% of organizations already developing customized AI solutions to maintain control over their capabilities and intellectual property.”

Sovereign AI is driven by four key motivations:

• Data control (72%): Organizations recognize data as a strategic asset, seeking to prevent external appropriation of sensitive information or intellectual property.
• National security (69%): AI systems function as instruments of soft power, making widespread reliance on foreign AI platforms a structural vulnerability.
• Economic competitiveness (48%): Sovereign AI creates advantages through domestic capacity building and long-term innovation ecosystem development.
• Regulatory compliance and cultural alignment (44% and 31%, respectively): AI systems can align with local regulations, values, and cultural contexts.

Open source software, open standards, and open data provide the foundation for AI sovereignty. The key benefits of these open foundations for sovereign AI include:

• Transparency and auditability (69%);
• Security and trust (60%);
• Flexibility for customization and fine-tuning (69%); and
• Innovation acceleration through collaborative development (41%).

“94% of respondents view global collaboration as essential to achieving sovereign AI. This finding reveals that participation in shared, community-driven open source development can work as a bridge to achieve sovereign AI. … Nevertheless, the path to open source sovereign AI includes obstacles such as data quality and availability issues (44%) and technical expertise shortage (35%). Obstacles to participating in global AI development include resource constraints (35%), intellectual property concerns (34%), geopolitical tensions (28%), national security restrictions (26%), and regulatory compliance challenges (26%).”

The relevance of sovereign AI

Survey respondents were asked whether sovereign AI is a strategic priority.

The term “Sovereign AI” has been used to describe efforts to develop AI capabilities with reduced external dependencies.  Based on this definition, do you consider this approach valuable? Yes (79%); Partially (16%); No, but it may become important (3%); No (2%).

Sovereign AI is recognized as a strategic priority around the world: US: ( 79%), Europe (83%), AP (78%).

Sovereign AI is strategic relevant at both national and organizational levels: National level (66%), Organizational level (47%), Supranational, e.g., European Union (45%), Regional/State (23%), Community (18%), and Municipal level (16%).

“The recognition of organizational-level sovereign AI relevance (47%) reflects the growing awareness that sovereign AI directly influences operational autonomy, competitive positioning, and long-term strategic flexibility in organizations. … Organizations are no longer merely asking, ‘Which AI solution performs best’ but rather, ‘Which AI solutions preserve our decision-making autonomy and align with our institutional values?’”

Drivers of sovereign AI

Is your organization developing customized AI solutions? Yes (82%); No, we use off-the-shelf AI solutions without modification (16%); No, we are not currently implementing AI solutions (6%).

In what organization sizes is the emphasis on customized solutions strongest? Enterprises with 10K or more employees (92%), large organizations with 1 to 10K employees (92%), mediums businesses with 50 to 1K employees (71%), small businesses with 1 to 49 employees (65%).

“For nations pursuing sovereign AI strategies, this organizational need of customized solutions validates the importance of maintaining domestic AI capabilities rather than depending entirely on foreign technologies that may embed different values or create potential vulnerabilities.”

What are your organization’s motivations for building its own custom AI solutions? Control over AI capabilities and IP (57%), addressing unique requirements (49%), meeting unique security requirements (41%), competitive advantage (37%), specific mission objectives (28%), equitable access to AI benefits (25%), reducing dependency on external AI providers (24%).

“Despite the widespread marketing of universal AI platforms, many organizations are encountering significant gaps between generic capabilities and their specific operational contexts, regulatory constraints, and strategic goals.”

What are the key drivers of interest in sovereign AI? Data sovereignty (72%), security concerns (69%), economic competitiveness (48%), regulatory compliance (44%), cultural alignment (31%).

The sovereign AI blueprint

How important is open source to the development of Sovereign AI systems? Essential (45%), very important (44%), moderately important (8%), slightly important (2%), not important (0%).

Which open approaches are most critical to advancing Sovereign AI? Open source software (81%), open standards (65%), open data (65%), open governance (49%), open infrastructure (42%), open hardware (22%).

“Together, these preferences reflect a recognition that true sovereignty extends beyond control over AI models — it requires autonomy over the entire technological stack and data pipeline. This holistic approach to openness indicates that organizations understand sovereignty not as vendor substitution but as systemic independence.”

How important are the following aspects of open source for achieving sovereign AI?

• Access to model weights and architecture – very important (84%), somewhat important (15%);
• Ability to inspect and modify code – very important (79%), somewhat important (19%);
• Transparency of training methods – very important (76%), somewhat important (21%);
• Freedom from vendor lock-in – very important (69%), somewhat important (27%);
• Ability to fine-tune for specific use cases – very important (69%), somewhat important (28%);
• Community support for implementation – very important (56%), somewhat important (41%); and
• Reduced deployment cost – very important (44%), somewhat important (51%).

What types of customizations does your organization make to open source AI systems? Integrating with proprietary data systems (53%), creating domain-specific knowledge bases (48%), implementing customer security or privacy issues (48%), developing customer user interfaces (35%), adopting models to specific languages or dialects (33%), optimizing for specific hardware infrastructures (32%), coupling with local regulations (25%), no customizations (3%).

What benefits does open source offer to Sovereign AI efforts? Transparency and audibility (69%), security and trust (60%), faster innovation (41%), vendor independence (41%), ecosystem building (38%), cost savings (21%).

What challenges does your organization face when customizing or building AI solutions with open source components? Data quality and availability (44%), technical expertise and skill gaps (35%), security vulnerabilities (34%), integration with existing systems (29%), keeping up with rapid evolution of tools (29%), compliance and regulatory requirements (27%), maintenance and support concerns (20%), scaling challenges (15%), none (3%).

Global AI collaboration

“Despite involving essentially local initiatives, sovereign AI presents a global challenge that requires international coordination.” As a result, global collaboration on open source AI is essential.

Is global collaboration on open source AI technology important to your organization? Yes (94%), No (6%).

Do you agree with the following statement?: “Open collaboration is essential to building secure Sovereign AI systems.” Strongly agree (48%), agree (45%), neutral (6%), strongly disagree (1%).

At which levels of the AI stack is global collaboration most valuable? Base and foundation models (59%), data resources (59%), development tools/platforms (39%), hardware/compute infrastructure (38%), evaluation frameworks (36%), domain specific applications (24%), deployment and operations (17%).

Next steps for sovereign AI

Which of these forms of global AI collaboration would you or your organization be most likely to participate in? Contributing to open source projects and tools (59%), creating shared technical standards (45%), collaborating on responsible AI principles and practices (45%), establishing common evaluation metrics (40%), developing protocols for AI interoperability (30%), participating in research on emerging AI capabilities (29%), sharing non-sensitive training data (26%).

What barriers prevent your organization from participating more actively in global AI collaboration? Resource constraints (35%), IP concerns (34%), geopolitical tensions (28%), national security restrictions (26%), regulatory compliance (26%), data governance challenges (24%), lack of collaboration frameworks (23%), interoperability issues (16%), competitive concerns (15%), misalignment of priorities with global partners (12%).

What kind of global cooperation model do you believe would best support Sovereign AI development? Open-source community led governance (43%), public-private partnerships (32%), multilateral agreements and standards (20%), regional alliances (5%).

Which stakeholders should be most involved in shaping the future of Sovereign AI? National governments (66%), open source foundations (60%), supranational governments (53%), standards organizations (41%), academia (41%), industry/trade associations (31%), private sector (25%), civil society organizations (21%), regional governments (20%).

“Our findings evidence the strategic importance of sovereign AI development through open source collaboration,” said The State of Sovereign AI report in conclusion. “Open source emerges as the dominant pathway to achieving sovereign AI, primarily because it enables transparency and auditability, security and trust, and the flexibility needed for customization without vendor lock-in.”

The report concludes with six key recommendations:

• Invest in open source AI infrastructure: Prioritize contributions and adoption of open source AI as the foundation for sovereign AI capabilities.
• Develop sovereign AI talent through education: Address the critical skills shortage by investing in comprehensive AI education programs, and creating specialized training focused on open source AI technologies and governance.
• Establish community-led governance frameworks: Support open source foundations and community-driven governance models that enable collaborative development.
• Create shared standards and protocols: Collaborate with industry partners to develop open technical standards that enable sovereign AI systems.
• Address data quality and availability challenges: Help develop collaborative approaches to creating high-quality, diverse datasets through open data initiatives.
• Foster strategic international collaboration: Governments must establish diplomatic and policy frameworks that enable global AI collaboration while addressing legitimate security concerns.

Irving Wladawsky-Berger is an Advisor on the Linux Foundation Research Advisory Board and is a research affiliate at MIT.