Endor Labs 构建工具来扫描 AI 生成的代码以查找漏洞，4月23日 宣布完成9300万美元B轮融资（总融资达1.63亿美元），由DFJ Growth领投，Salesforce Ventures等跟投，估值较A轮"高出几个数量级"。 Endor 目前有 133 名员工，主要分布在帕洛阿尔托和班加罗尔的办公室。 公司的CEO Badhwar 表示，本轮融资对 Endor 的估值比其 A 轮估值 " 高出几个数量级 " 。他补充说，所得款项将用于扩展 Endor 的平台。 B 轮融资使这 家初创公司的总融资额达到 1.63 亿美元。 " 新一轮融资使我们能够继续交付产品，即使在比 5 到 10 年前同类公司面临的更严峻的宏观环境中， "Badhwar 告诉 TechCrunch 。 " 我们现在加息是因为 我们看到了强劲的势头 —— 自 2023 年 A 轮融资以来，年度经常性收入增长了 30 倍 —— 这使我们能够加倍努力为客户提供成果。 " AI 生成的代码无疑正在改变软件的构建方式，但它也带来了新的安全挑战。根据开发人员安全平台 Synk 在 2023 年底的一项调查，超过 50% 的组织有时 会或 ...

Core Insights - Endor Labs has developed tools to scan AI-generated code for vulnerabilities and recently completed a $93 million Series B funding round, bringing total funding to $163 million, with a valuation significantly higher than its Series A round [1][2]. Funding and Growth - The company has 133 employees primarily located in Palo Alto and Bangalore [2]. - CEO Badhwar indicated that the Series B funding allows Endor to continue delivering products despite a challenging macro environment, with annual recurring revenue increasing 30 times since the Series A funding in 2023 [2][5]. - The funding will be used to expand Endor's platform [2]. Market Opportunity - Endor initially focused on protecting open-source package dependencies but shifted its focus to address the growing demand for identifying vulnerabilities in AI-generated code [3]. - The platform not only reviews code and identifies risks but also recommends precise fixes and can automatically apply them [3]. Product Offerings - Endor has launched a tool to help organizations discover where AI models and services integrate with their codebases and assess potential security vulnerabilities [4]. - The company claims to protect over 5 million applications for clients such as OpenAI, Rubrik, Peloton, Snowflake, Egnyte, and Dropbox, conducting over 1 million scans weekly [4]. Industry Context - DFJ Growth's partner Ramin Sayar noted that Endor is positioned well as generative AI changes coding practices, creating a need for visibility and control in code generation [6].