Anthropic's Claude Mythos Preview Shows Major Gains in Cybersecurity and Software Engineering

• Anthropic 的估值在約 14 個月內從 610 億美元飆升至 9,000 億美元，反映出投資者對這家 AI 公司的技術實力及成長軌跡展現了前所未有的興趣。 • 此次估值激增受強勁的營收表現以及投資者對 Anthropic AI 模型與研發的信心所推動，使其在生成式 AI 市場中成為 OpenAI 強大的競爭對手。 • 此次迅速崛起展現了尖端 AI 開發領域日益激烈的資本競爭，主要投資者同時支持多家 AI 初創公司，以對沖對該技術轉型潛力的賭注。

• Deep Tech Week 亮點在於 The AI Collective，該組織被譽為全球最大的 AI 社群，透過活動、論壇與 networking 連結了超過 100,000 名 AI builders、researchers 和專業人士。 • 此盛會匯集了 startups、corporate labs 與 academic groups，展示在 core machine learning、robotics 及 applied AI 方面的進展，並專注於 interoperability 與 responsible deployment。 • 美國的參與非常踴躍，許多領先的 AI research labs 以及 venture-backed startups 利用此場合宣布新的 collaborations、model releases 與 open-source tooling。

• 在 2026 年第一季，AI 驅動的保險科技公司吸引了巨額資金，其中包括 Harper Group 的 4,700 萬美元種子輪加 A 輪融資，以及 Indigo Technologies 的 5,000 萬美元 B 輪融資，兩者均專注於 AI 自動化核保與理賠工作流程。 • Mea Platforms、Pasito、General Magic 和 Qumis 等平台籌集了 430 萬至 5,000 萬美元不等的資金，旨在自動化文件密集型流程、降低賠付率，並為中小企業縮短報價與承保時間。 • 投資者正大舉支持總部位於美國的供應商，這些供應商利用 AI 在不按比例增加員工人數的情況下擴大分銷規模、減少福利選擇過程中的摩擦，並改善商業險種的風險篩選。

• Oracle 於 2026 年 4 月 30 日發布了新的客戶安全公告，搶先防範針對 AI 模型的威脅。 • 該公告揭示了網路安全的前兆，敦促更新 AI 部署的安全態勢。 • 這標誌著針對瞄準生成式 AI 系統之新型漏洞所採取的積極主動措施。

• Meta 報告 2026 年第一季雲端服務的合約承諾達 1,070 億美元，以推動 AI 模型開發。 • 積極投資包括數據中心擴張，以及 AI 硬體零組件的供應鏈協議。 • 公司預計 2026 年晚些時候將取得 AI 重大突破，並致力於將先進研究成果變現。

• Amazon Web Services achieved its fastest growth in nearly four years, fueled by AI solutions and enterprise adoption. • Key drivers include innovations like Bedrock and Kiro amid surging cloud spending. • AI is sparking a new wave of infrastructure investments in the US market.

• SoftBank is preparing Roze AI, combining robotics, data centers, and AI infrastructure for a US IPO targeting $100B valuation. • Masayoshi Son backs the venture, with potential public listing as early as late 2026 or 2027. • Executives plan an analyst day in July, though no products or revenue plans announced yet.

• Aviatrix Inc. announced the launch of a new platform designed to contain artificial intelligence agents and enforce security controls across AI workloads without requiring changes to AI agents or code. • The platform extends Aviatrix's Cloud Native Security Fabric with two new products: Zero Trust for AI Workloads (now generally available) and Aviatrix AgentGuard (in early access), enabling IT teams to set policies that allow or deny access to external AI services and block shadow AI with allowlists. • The solution provides network-layer enforcement across workloads and regions, addressing the growing need to secure AI agents, large language model proxies, and agentic frameworks in enterprise environments.

• The Pentagon launched GenAI.mil in December 2025 as part of Project Maven, providing secure generative AI access to approximately 3 million Department of Defense staff members. • The platform embeds multiple leading AI models including Anthropic's Claude, Google's Gemini, xAI's Grok, and OpenAI's ChatGPT, enabling widespread AI adoption across military operations. • The initiative reflects the U.S. military's strategic priority to integrate advanced AI capabilities into defense operations and demonstrates significant government investment in dual-use AI technology.

• Skye AI, a home screen app for iPhone powered by artificial intelligence, raised $3.58 million in pre-seed funding while still in private testing phase. • The funding supports development of the app, which aims to enhance iPhone user experience through AI-driven features accessible directly from the home screen. • This investment highlights growing investor interest in mobile AI applications amid competition in the smartphone ecosystem.

• Sevii announced the launch of its Autonomous Cyber Swarm Defense (CSD) capability on April 28, 2026, to counter high-volume AI-driven cyber attacks at machine speed and scale. • The ADR platform deploys Agentic AI Cyber Warrior agents that detect, contain, and remediate threats in under 15 minutes, reducing analyst workloads by 95% with a fixed per-asset pricing model eliminating AI token costs. • CSD Mode features Myrmidon Defense Technology (MDT) for efficient micro-orchestration, enabling predictable budgeting without additional hires or outsourcing, as noted by Synergem Technologies CEO Frank Holt.