Anthropic has expanded Project Glasswing to around 150 new organisations in more than 15 countries, including operators and vendors tied to power, water, healthcare, communications and hardware. The programme gives vetted partners access to Claude Mythos Preview for defensive vulnerability discovery. The strategic signal is that frontier AI is being deployed first as controlled offensive defence, while the hard bottleneck moves from finding vulnerabilities to validating, disclosing and patching them.
AI company developing Claude models and controlled cyber-defence access programmes
Anthropic is moving frontier AI capability into defensive cybersecurity workflows that may reshape critical infrastructure security practice.
Anthropic is moving frontier AI capability into defensive cybersecurity workflows that may reshape critical infrastructure security practice.
The event signals faster vulnerability discovery for communications, energy, water and healthcare operators while exposing a remediation bottleneck in human-led triage and patching.
The event signals faster vulnerability discovery for communications, energy, water and healthcare operators while exposing a remediation bottleneck in human-led triage and patching.
Anthropic扩展Glasswing项目，将AI漏洞狩猎引入关键基础设施防御。
The event signals faster vulnerability discovery for communications, energy, water and healthcare operators while exposing a remediation bottleneck in human-led triage and patching.
| 0.90–1.00 | A | High — direct sources |
| 0.75–0.89 | A/B | Strong |
| 0.55–0.74 | B/C | Medium |
| 0.35–0.54 | C/D | Weak–medium |
| 0.10–0.34 | D | Weak signal |
| 0.00–0.09 | D | Internal monitoring |
Direct public sources
• Glasswing项目覆盖15个国家150个经过审查的合作伙伴，推动关键基础设施防御 另见: Anthropic.
• AI驱动的漏洞发现速度现已超过人工修复能力 另见: 台积电权衡AI芯片涨价，面临短缺.
事实
Anthropic已将Glasswing项目扩展至超过15个国家的约150个组织，新增了电力、水务、医疗、通信和硬件等领域的关键基础设施运营商。经过审查的合作伙伴可获得Claude Mythos Preview的访问权限，用于防御性漏洞扫描。自4月初启动以来，合作伙伴已识别出超过10,000个高严重性或关键软件漏洞，其中1,700多个经独立审查的结果显示确认有效率超过90%。 另见: Anthropic 扩展 Glasswing 项目，强化 AI 漏洞防御.
评估
Glasswing项目将强大的AI模型用于受控防御用途，以免类似能力扩散到敌对行为者手中。瓶颈已从发现缺陷转向分类和补丁部署，从而催生了对AI辅助修复和供应链安全服务的需求。该计划还突显了一个更广泛的风险：核心互联网协议和广泛使用的开源库正受到大规模扫描，这将暴露基础设施堆栈中长期隐藏的漏洞。 另见: IBM 与 Google 扩展企业 AI 代理.
关注要点
关注AI驱动的漏洞狩猎能否以更低的成本改善关键基础设施安全，Glasswing是否会进入自动补丁领域，以及竞争对手AI公司是否会推出类似的防御计划。 另见: Pinterest 承诺向 AWS AI 基础设施投入 40 亿美元.
另请阅读：DNA增加网络级来电垃圾标签功能
Signal Brief
- Signal: Anthropic将AI漏洞狩猎引入关键基础设施防御
- Signal Type: AI cybersecurity programme expansion
- Region: Global
- Market Class: Cloud Service
Operating Surface
- Published sources should identify the affected parties, operating surface, and market exposure before this trend map is treated as complete.
Market Context
- The event signals faster vulnerability discovery for communications, energy, water and healthcare operators while exposing a remediation bottleneck in human-led triage and patching.
- Operational relevance: High
- Time horizon: Next quarter
What To Watch
- Watch for official statements, regulatory updates, customer or partner exposure, and follow-up disclosures.
Member Briefing
Deeper Trend Context
Login is required to unlock the full trend briefing and source notes.
Only for Strategy Circle
Strategic Circle Access
Open to all readers. Unlock trend briefings after joining and logging in.
Join Strategic CircleOnly for Leadership Alliance
Leadership Alliance Access
For operators, investors, and policy teams that need relationship evidence, failure paths, and source notes. Login required to unlock.
Join Leadership AlliancePublic Sources and Linked Organizations
6 linked-organization notes require member access.
