February 2026 Summaries

Secretary of War Pete Hegseth announced that Anthropic would be designated a supply chain risk by the Department of War, following stalled negotiations over the use of their AI model, Claude, for mass domestic surveillance and fully autonomous weapons. Anthropic, which has been involved with the U.S. government's classified networks since June 2024, opposes these applications due to concerns about reliability and potential rights violations. The company plans to legally challenge any such designation, as it could set a dangerous precedent for American companies negotiating with the government. Hegseth's statement suggests this designation might restrict business with military contractors; however, Anthropic clarifies that it should only affect Department of War contracts and not other customers. Anthropic reassures its clients that their access to Claude remains unaffected outside of military contracts and expresses gratitude for the support received from various stakeholders.

Anthropic, led by Dario Amodei, has been collaborating with the U.S. Department of War to deploy its AI models, including Claude, for mission-critical applications such as intelligence analysis and cyber operations. Despite this close relationship, Anthropic has taken a firm stance against the use of its AI for mass domestic surveillance and fully autonomous weapons, citing concerns over democratic values and the reliability of current AI systems. These exceptions have not hindered the adoption of their models, but the Department of War has threatened to label Anthropic a "supply chain risk" and invoke the Defense Production Act to force compliance. Anthropic remains committed to providing its technology to support national security while maintaining its ethical safeguards, expressing hope for continued collaboration without compromising its principles.

Anthropic's third version of its Responsible Scaling Policy (RSP) aims to address the evolving risks of AI systems by reinforcing effective measures, improving transparency, and introducing new measures for accountability. Initially introduced in September 2023, the RSP was designed to manage risks that could emerge rapidly with advancing AI technology, using a principle of conditional commitments tied to capability levels. While the RSP has successfully incentivized Anthropic and other companies to adopt stronger safeguards and inform early AI policy, challenges remain, particularly in achieving consensus on AI risks and implementing higher-level safeguards unilaterally. The updated RSP includes a clearer separation of company plans from industry recommendations, introduces a Frontier Safety Roadmap to outline risk mitigation strategies, and mandates regular Risk Reports with external reviews to enhance safety and transparency. As the AI landscape continues to evolve, Anthropic commits to revising its policy to adapt to new capabilities while encouraging multilateral action to address industry-wide risks.

Anthropic has acquired Vercept to enhance the computer use capabilities of its AI system, Claude, enabling it to handle complex tasks within live applications similar to human interaction. This acquisition aims to leverage Vercept's expertise in AI perception and interaction to solve intricate problems, following the recent improvement in Claude's computer use proficiency with the release of Sonnet 4.6, which now shows near human-level performance in tasks like navigating spreadsheets and web forms. Vercept's integration into Anthropic comes after the acquisition of another team, Bun, as part of Anthropic's strategy to collaborate with teams that share its commitment to technical ambition, safety, and rigor. In addition to expanding AI capabilities, Anthropic has introduced Claude Code Security, a feature in research preview that scans codebases for vulnerabilities and recommends software patches, reinforcing its focus on cybersecurity.

A recent announcement highlights industrial-scale distillation attacks by three AI laboratories—DeepSeek, Moonshot, and MiniMax—aimed at extracting capabilities from the AI model Claude to enhance their own systems. Distillation, while a legitimate AI training method, is being used illicitly by these labs, circumventing terms of service and regional access restrictions, posing national security risks by potentially stripping models of critical safeguards. These attacks exploit vulnerabilities in export controls and involve fraudulent accounts and proxy services to access Claude, with the labs generating millions of exchanges targeting specific capabilities such as reasoning, coding, and tool use. The announcement details the methods used by each lab, the scale of their operations, and the broader implications for AI security and export controls, emphasizing the need for a coordinated industry and policy response to mitigate such threats. Anthropic, the company behind Claude, is investing in defenses and intelligence sharing to detect and prevent further distillation attacks, while advocating for global cooperation to address this growing issue.

Claude Code Security, a new feature within Claude Code, offers a fresh approach to cybersecurity by scanning codebases for vulnerabilities and suggesting software patches for human review, addressing the challenge of too many vulnerabilities and insufficient personnel to manage them. Unlike traditional static analysis tools that rely on known patterns, Claude Code Security employs AI to understand code like a human researcher, identifying complex vulnerabilities that are often overlooked. This capability is released as a limited research preview to Enterprise and Team customers, with accelerated access for open-source maintainers, to refine its features and ensure responsible deployment. Each finding undergoes a multi-stage verification process with severity ratings to prioritize fixes, and nothing is implemented without human approval. Built on extensive research and stress-testing, Claude Code Security aims to enhance defensive capabilities in cybersecurity, allowing teams to work within familiar tools. The tool is part of a broader effort to increase security across the industry by leveraging AI to detect previously undetected bugs and reduce the risk of attacks.

Anthropic and the Government of Rwanda have entered into a three-year Memorandum of Understanding to integrate AI into Rwanda's education, health, and public sector systems, marking Anthropic's first multi-sector government partnership in Africa. This collaboration aims to support Rwanda's national health goals, such as eliminating cervical cancer and reducing malaria and maternal mortality, while also enhancing public sector capabilities through AI tools like Claude and Claude Code. The agreement formalizes a previous educational partnership, providing AI literacy training and deploying an AI learning companion across eight African countries. The initiative emphasizes capacity building and responsible AI deployment, reflecting a commitment to leveraging technology for public good and enabling local autonomy in the adoption of new technologies.

Anthropic and Infosys have announced a collaboration to develop enterprise AI solutions for regulated industries such as telecommunications, financial services, manufacturing, and software development by integrating Anthropic's Claude models with Infosys Topaz. This partnership aims to enhance software development and AI adoption with governance and transparency, leveraging Infosys's domain expertise. A significant emphasis is placed on creating agentic AI systems that can autonomously manage complex, multi-step tasks across different sectors, such as modernizing network operations in telecommunications, automating risk detection in financial services, accelerating product design in manufacturing, and streamlining software development processes. The collaboration also focuses on modernizing legacy systems, and Infosys is one of the first partners in Anthropic's expanded presence in India, where nearly half of Claude usage involves advanced AI work. This initiative aims to redefine industry operations and drive innovation, making AI a transformative force for global enterprises.

Claude Sonnet 4.6 is the latest upgrade in the Sonnet series, offering significant improvements in coding, computer usage, reasoning, planning, and design. It introduces a 1M token context window in beta and maintains the same pricing as its predecessor, Sonnet 4.5. Users have reported a preference for Sonnet 4.6 over earlier models due to its enhanced consistency, instruction-following, and reduced overengineering, making it favorable even compared to the previous high-end Opus 4.5 model. It demonstrates human-level capabilities in tasks that require complex reasoning and multitasking, showing marked improvements in areas such as coding, document comprehension, and financial analysis. The model also features robust safety measures against prompt injection attacks, and its performance in various benchmarks, such as OSWorld and Vending-Bench Arena, highlights its capacity for handling complex tasks at a competitive price point. Sonnet 4.6 is available on all Claude plans and major cloud platforms, supporting adaptive and extended thinking, context compaction, and improved integration with tools like Excel for enhanced functionality.

Anthropic has launched a new office in Bengaluru, India, demonstrating its commitment to expanding AI technology across the country by forming partnerships with various sectors such as enterprise, education, and agriculture. Recognizing India as a significant market with immense potential, Anthropic is addressing the language gap in AI models by improving performance in 10 widely spoken Indian languages. Collaborations with local nonprofits are enhancing AI's applicability in agriculture and legal services, including support for judicial access through Adalat AI's national WhatsApp helpline. In education, partnerships with organizations like Pratham aim to provide AI-enabled learning tools to underserved communities. The company is also contributing to the public sector by working with the EkStep Foundation to leverage AI for population-scale impacts. Anthropic's Model Context Protocol (MCP) is being integrated into government and private sector operations, exemplified by Swiggy's use for seamless grocery ordering and dining reservations. With the opening of its Bengaluru office, Anthropic is focused on hiring local talent and expanding its influence in India's dynamic market, supported by a recent $30 billion Series G funding round that positions the company as a leader in enterprise AI and coding.

Chris Liddell has been appointed to the Board of Directors at Anthropic, bringing over 30 years of senior leadership experience from companies such as Microsoft and General Motors, as well as from his role as Deputy White House Chief of Staff. His appointment is seen as a valuable addition given his expertise at the intersection of technology, public service, and governance, which aligns with Anthropic's mission of building responsible AI technologies. Liddell joins other notable figures on the board, including Anthropic co-founders and industry leaders, and emphasizes the importance of governance in transformative technologies. In addition to his corporate roles, Liddell contributes to philanthropy and environmental causes, notably serving as Chairman of New Zealand's largest environmental foundation. Meanwhile, Anthropic has successfully raised $30 billion in Series G funding, achieving a post-money valuation of $380 billion, which will support its continued leadership in enterprise AI and coding initiatives.

Anthropic has raised $30 billion in Series G funding, led by GIC and Coatue, valuing the company at $380 billion post-money, with significant contributions from investors including D. E. Shaw Ventures, Dragoneer, and others. The funding will support Anthropic's frontier research, product development, and infrastructure expansions, reinforcing its leadership in enterprise AI and coding. Anthropic's revenue has grown dramatically, with a current run-rate of $14 billion, driven by the widespread adoption of its AI platform, Claude, which is favored by eight of the Fortune 10 companies and has seen a sevenfold increase in customers spending over $100,000 annually. Claude Code, launched in 2025, has become a major force in agentic coding, contributing to 4% of global GitHub public commits and seeing business subscriptions quadruple in 2026. Anthropic's latest model, Opus 4.6, excels in generating professional documents and is the top performer on GDPval-AA, reflecting its strength in economically valuable tasks. The company plans to use the Series G funds to expand its infrastructure, ensuring Claude's availability across major cloud platforms and enhancing performance through diversified AI hardware.

Anthropic has partnered with CodePath, the largest collegiate computer science education provider in the U.S., to integrate its AI tool, Claude, into CodePath's curriculum, aiming to reshape software development education for over 20,000 students, including many from underrepresented backgrounds. This collaboration will incorporate AI courses such as Foundations of AI Engineering and AI Open-Source Capstone, allowing students to work with tools like Claude Code and participate in real-world open-source projects. The initiative seeks to democratize access to advanced AI education, which was traditionally available only at wealthier institutions, and aligns with Anthropic’s broader commitment to expanding AI education globally, including partnerships with the American Federation of Teachers, the Icelandic Ministry of Education, and the Rwandan government. Additionally, this partnership is part of Anthropic's efforts to explore the evolving landscape of AI in coding education and economic opportunities, while also supporting their expansion through significant new funding, positioning Anthropic as a leader in enterprise AI and coding innovation.

Anthropic is donating $20 million to Public First Action, a bipartisan organization aimed at promoting effective AI governance and public education on AI risks and benefits. As AI rapidly advances, bringing both substantial benefits and significant risks, Anthropic emphasizes the need for robust policies to regulate its use, advocating for flexible regulations that promote transparency, safeguard against misuse, and maintain America’s competitive edge. Public First Action, working with both Republican and Democratic strategists, seeks to fill the current policy gap by supporting nonpartisan initiatives for AI governance, including transparency in AI development, federal oversight, smart export controls, and targeted regulations against high-risk AI applications like biological weapons and cyberattacks. Anthropic’s contribution is part of its broader commitment to ensuring AI serves the public good, highlighting the necessity for companies to responsibly manage AI’s transformative potential and associated risks.

Anthropic, an AI infrastructure company, is committed to mitigating the effects of electricity price increases caused by its data centers in the U.S., as it anticipates a significant demand for power due to the growth of AI technologies. The company plans to fully cover the costs of grid infrastructure upgrades required for connecting its data centers, ensuring these expenses are not passed on to consumers. Additionally, Anthropic aims to bring new power generation online to match its data centers' energy needs and will invest in systems to reduce grid strain during peak demand. The company is also focused on community investments, creating jobs, and minimizing environmental impacts. While taking these measures, Anthropic acknowledges the necessity for broader systemic changes in energy policy and supports federal efforts to expedite energy infrastructure development. These initiatives are part of its ongoing commitment to addressing the impact of data centers on energy costs, with more updates to follow as their work progresses.

Claude Opus 4.6, the latest upgrade from Anthropic, significantly enhances previous model capabilities with improved coding and reasoning skills, a 1 million token context window in beta, and a performance edge in complex, multidisciplinary evaluations. This model excels in tasks such as financial analysis, research, and document creation, while also showcasing state-of-the-art performance in agentic coding, long-context reasoning, and cybersecurity applications. It demonstrates superior abilities in retrieving and reasoning with vast amounts of information, outperforming both its predecessor and industry rivals in various benchmarks. With new features like adaptive thinking and context compaction, Opus 4.6 is designed to efficiently handle long-running tasks, and its safety profile remains robust with low rates of misalignment. Additionally, it integrates seamlessly with office tools like Excel and PowerPoint, enhancing productivity and collaboration. The Claude Opus 4.6 is available on multiple platforms, maintaining the same pricing as previous models, and offers developers enhanced control over model effort and task execution.

Apple's Xcode 26.3 now supports the Claude Agent SDK, enhancing developers' capabilities by allowing them to integrate Claude, an AI assistant, directly within the IDE for long-running and autonomous coding tasks. This integration provides developers with tools like subagents, background tasks, and plugins, and improves functionality such as visual verification with Previews, which is especially beneficial when working with SwiftUI views. Claude can reason across entire projects, autonomously execute tasks, and directly access Apple's documentation to aid in development. Additionally, it offers an interface through the Model Context Protocol, enabling users to capture visual Previews without leaving the command-line interface. Xcode 26.3 is available as a release candidate to Apple Developer Program members, with a full release imminent on the App Store.

Anthropic has decided to keep its AI assistant, Claude, free of advertisements to maintain its role as a genuinely helpful and unambiguous tool for users. The company believes that introducing ads could compromise the integrity of user interactions by introducing commercial incentives that may conflict with users' interests, particularly in sensitive or complex discussions. Unlike products like search engines or social media, where users expect a mix of organic and sponsored content, AI interactions are more open-ended and personal, making them susceptible to influence. Anthropic's business model focuses on generating revenue through enterprise contracts and subscriptions without resorting to selling user attention or data to advertisers. They aim to expand access to Claude globally while preserving its utility as a trusted tool for thought. Additionally, the company is exploring agentic commerce, where Claude could assist users with purchases or bookings, and plans to enhance productivity by integrating with third-party work tools.

Anthropic has announced partnerships with the Allen Institute and Howard Hughes Medical Institute (HHMI) to enhance scientific discovery using AI, specifically with their AI model, Claude. These collaborations aim to address the challenges of transforming massive amounts of biological data into actionable insights by integrating AI tools into research workflows. At HHMI, the partnership focuses on AI-enabled scientific discovery, leveraging technologies from Janelia Research Campus to develop AI models that respond to experimental needs. The Allen Institute collaboration aims to create multi-agent AI systems for data analysis and exploration, enhancing researchers' capabilities to discover and interpret complex patterns. Both partnerships emphasize the importance of transparency, scientific rigor, and researcher autonomy, positioning Claude as a tool that supports scientific judgment rather than replacing it, and offering valuable insights for the development of AI systems that can assist in diverse scientific contexts.