Anthropic Scales Back Landmark Safety Policy to Maintain Global AI Competitiveness Amid Pentagon Contract Dispute
Anthropic updates its Responsible Scaling Policy, ending mandatory pauses for dangerous models to stay competitive as Pentagon contract threats loom.
By: AXL Media
Published: Feb 26, 2026, 4:29 AM EST
Source: The information in this article was sourced from City A.M.

The Transaction or Development
Anthropic, a company that originally distinguished itself through a rigorous commitment to AI safety, has officially loosened its internal "Responsible Scaling Policy." Under the new 3.0 framework, the firm will no longer actively stop the development or deployment of a model deemed to cross safety risk thresholds if a competitor has already released a comparable system. This marks a radical departure from the company's 2023 stance, which prioritized safety pauses regardless of market conditions, and reflects the intense commercial pressure facing the San Francisco-based developer.
Regulatory and Competitive Landscape
The policy shift is a direct response to a regulatory environment that Anthropic claims has stalled at the federal level. While the company previously championed safety-oriented discussions, it now argues that the policy landscape has shifted toward prioritizing national AI competitiveness and rapid economic growth. By abandoning its unilateral safety pauses, Anthropic aims to keep pace with primary rivals like OpenAI and Google, as well as agile Chinese developers who are currently flooding the market with low-cost, high-performance models.
Strategic Rationale and Market Impact
Valued at approximately $380 billion, Anthropic is currently navigating a high-stakes transition from a safety-first research lab to a major government and enterprise contractor. The company admitted that sticking rigidly to its earlier, more cautious approach would risk leaving it behind in the race for dominant market share. This pivot is intended to ensure that Anthropic remains a viable alternative for massive capital investments and infrastructure deals, even if it means moving away from the "industry-leading" safety standards that defined its brand during its founding years.
Categories
Topics
Related Coverage
- Pentagon’s Blacklisting of Anthropic Triggers Unprecedented Confrontation Between Washington and Silicon Valley Over AI Safety
- Anthropic Resists Venture Capital Overtures Valuing AI Firm At Over Eight Hundred Billion Dollars
- OpenAI Faces Executive Resignations and Mass User Exodus Following Controversial Artificial Intelligence Contract with U.S. Pentagon
- Pentagon Ousts Navy Secretary John Phelan Amid Rapid Military Leadership Purge and Iran Blockade