National security, export controls and military AI deals reshaping the AI competitive landscape

National Security and the Shifting Landscape of Military AI Infrastructure

The global race to develop and deploy advanced AI hardware and related infrastructure has taken on a new dimension, heavily intertwined with national security concerns, export controls, and geopolitical tensions. Recent developments highlight how government policies, corporate actions, and security considerations are reshaping the AI competitive landscape.

Anthropic’s Clashes with the Pentagon and Supply Chain Risks

Anthropic, a prominent AI startup, has recently become embroiled in security debates surrounding military and government use of AI technology. The Pentagon’s directive to cease using Anthropic’s AI models—citing espionage risks linked to Chinese AI laboratories—has sparked significant controversy. Anthropic has publicly called this supply chain risk designation “unprecedented” and “legally unsound,” and plans to challenge it in court.

This legal dispute underscores the broader security dilemma: the U.S. government’s efforts to control and vet AI vendors for military applications while balancing innovation. Notably, Anthropic’s AI, Claude, has been at the center of accusations by Chinese companies, which are alleged to have conducted distillation and mining activities of Claude to improve their own models—a practice that raises concerns about AI model theft and intellectual property security.

Further complicating the picture, the Defense Secretary recently summoned Anthropic’s CEO, Sam Altman, over military use of Claude, reflecting heightened scrutiny of AI vendors in the defense sector. Meanwhile, Anthropic has criticized the US debate over AI chip exports, which aims to curb China’s access to advanced semiconductor technology, as part of a broader effort to safeguard US technological dominance.

US Export-Control Policies and Global AI Infrastructure

The United States has intensified export controls on advanced semiconductor equipment, such as lithography machines, to limit China’s ability to develop cutting-edge chips domestically. However, recent approvals—like the export of H200 lithography systems to China—signal a nuanced approach that balances security with strategic interests. China has responded by accelerating its domestic chip manufacturing efforts, aiming for self-reliance and reduced dependency on foreign technology.

In this context, Chinese companies are distilling and mining AI models like Claude to circumvent export restrictions, raising concerns over the proliferation of AI capabilities outside US-controlled supply chains. These activities highlight the geopolitical stakes involved in AI hardware and software development, with countries vying for technological sovereignty.

The Pentagon’s Engagement with Industry and AI Vendors

Despite restrictions, some AI companies are securing government contracts for military applications. OpenAI, for example, has signed a Pentagon deal to deploy its models within classified military systems, following security safeguards and compliance measures. This contrasts with Anthropic’s ongoing legal conflicts and banishment from US agencies, illustrating a divided landscape where certain vendors are favored based on security assessments.

Additionally, the US government is actively pushing agencies to onshore critical AI infrastructure. The recent directives to stop using Anthropic’s technology reflect an effort to tighten security around AI supply chains and prevent espionage. Meanwhile, corporate investments continue to fuel regional AI ecosystems—India, for instance, announced a $110 billion plan to develop a comprehensive AI hardware and data center infrastructure. Indian companies like Tata and Reliance are investing heavily to position the country as a significant player in the global AI infrastructure race.

Regional and Geopolitical Dimensions

Regions worldwide are pursuing strategies to bolster technological sovereignty:

• India is rapidly expanding its AI hardware ecosystem, aiming to become a regional hub. The recent $200 billion boost in investments, along with partnerships involving Nvidia, OpenAI, and Tata, exemplifies this push.
• Japan and South Korea are investing in fabs and AI robotics, with Japan’s $17 billion fab project serving to reduce dependence on Taiwan and China.
• Europe is incentivizing domestic semiconductor manufacturing and data centers, seeking strategic autonomy amid supply chain vulnerabilities.

The Future of Military AI Infrastructure

The current landscape indicates a paradigm shift: AI infrastructure is increasingly viewed as a strategic battleground. The interplay of massive capital investments, regional alliances, and security policies is shaping a new era where technological sovereignty and supply chain resilience are paramount.

Legal disputes like Anthropic’s challenge to the Pentagon’s supply chain risk designation exemplify the tension between innovation and security. Meanwhile, the deployment of AI models within military systems—whether through open agreements like OpenAI’s or restrictions on others—reflects a nuanced security landscape where trust, compliance, and geopolitical considerations determine vendor selection.

In conclusion, the AI infrastructure race in 2026 is no longer solely about technological breakthroughs but about building resilient, secure, and sovereign ecosystems that can sustain national security interests. The ongoing efforts to control hardware supply chains, vet AI vendors, and forge regional alliances will define leadership in the AI era, emphasizing the importance of aligning technological ambition with geopolitical strategy.