字号·· | 护眼
美国消费者新闻与商业频道

CloudSEK推出网络安全初创企业孵化器,并发布AI安全网关CloudSEK launches cybersecurity startup incubator, introduces AI security gateway

点「原文对照」整页切到原文,或双击某段只看那段的原文。

总部位于班加罗尔的网络安全公司CloudSEK推出了一项为期12周的孵化计划,面向早期网络安全初创企业,并发布了一款AI网关,旨在保障应用程序、AI智能体与第三方AI模型之间的交互安全。

Bengaluru-based cybersecurity company CloudSEK has launched a 12-week incubation programme for early-stage cybersecurity startups and introduced an AI gateway designed to secure interactions between applications, AI agents and third-party AI models.

这项名为CloudSEK Launchpad的初创企业计划首期将遴选10家公司。据该公司介绍,入选的初创企业将有资格获得最高5000美元的AI模型额度,并获得CloudSEK的威胁情报资源、导师指导、企业网络以及潜在投资者的对接机会。申请截止日期为2026年11月20日。

The startup programme, CloudSEK Launchpad, will select 10 companies for its first cohort. Selected startups will be eligible for up to $5,000 in AI model credits and receive access to CloudSEK’s threat intelligence resources, mentorship, enterprise networks and potential investors, according to the company. Applications are open until November 20, 2026.

此举出台之际,企业正不断扩大对生成式AI和自主智能体的使用,这引发了人们对数据暴露、凭证管理、提示注入以及运行AI应用成本的担忧。对于早期网络安全公司而言,能否接触到企业客户和真实世界的安全数据,也可能成为产品开发和商业化的障碍。

The initiative comes as businesses expand their use of generative AI and autonomous agents, raising concerns around data exposure, credential management, prompt injection and the cost of running AI applications. For early-stage cybersecurity companies, access to enterprise customers and real-world security data can also be a barrier to developing and commercialising products.

CloudSEK的第二项发布产品KeyVigil旨在通过为访问多个AI模型的应用程序和AI智能体提供集中式网关,来应对上述部分风险。该公司表示,该平台可以管理API凭证、检测并遮蔽敏感信息、标记提示注入和越狱尝试、记录模型交互并设置使用限额。

CloudSEK’s second launch, KeyVigil, is designed to address some of these risks by providing a centralised gateway for applications and AI agents accessing multiple AI models. The company says the platform can manage API credentials, detect and mask sensitive information, flag prompt-injection and jailbreak attempts, log model interactions and set usage limits.

KeyVigil还允许开发者通过单一端点访问多个AI模型,这有可能减少在更换供应商时修改应用程序代码的需要。该平台面向企业客户提供,而入选Launchpad的初创企业将获得使用该服务的额度。

KeyVigil also allows developers to access multiple AI models through a single endpoint, potentially reducing the need to modify application code when switching providers. The platform is available to enterprise customers, while startups selected for Launchpad will receive credits to use the service.

CloudSEK表示,KeyVigil由其AIVigil平台背后的团队开发,该平台专注于识别暴露的AI资产和攻击面。

CloudSEK said KeyVigil was developed by the team behind its AIVigil platform, which focuses on identifying exposed AI assets and attack surfaces.

这款新产品旨在将这些能力扩展到应用程序、AI智能体与AI模型之间的交互。Launchpad将聚焦种子前轮和种子轮的网络安全公司,不过处于创意阶段的创始人如果拥有强大的团队和明确定义的安全问题,也可以申请。关注的领域包括AI安全、自动化安全运营、威胁情报、数字风险管理、身份安全、欺诈防范以及软件供应链安全。

The new product is intended to extend those capabilities to interactions between applications, AI agents and AI models. Launchpad will focus on pre-seed and seed-stage cybersecurity companies, although founders at the idea stage can also apply if they have a strong team and a clearly defined security problem. Areas of interest include AI security, automated security operations, threat intelligence, digital risk management, identity security, fraud prevention and software supply-chain security.

这个为期12周的项目将包括由CloudSEK高管提供的指导、接触其网络安全研究成果的机会,以及向印度、海湾合作委员会地区和美国的潜在客户及渠道合作伙伴进行引荐。参与的初创公司还将在结项演示日上获得向投资者和潜在客户展示产品的机会。CloudSEK创始人兼首席执行官拉胡尔·萨西表示,该项目旨在帮助网络安全领域的创始人应对客户获取、产品定位和获取行业专业知识等挑战。

The 12-week programme will include mentorship from CloudSEK executives, access to its cybersecurity research and introductions to potential customers and channel partners in India, the Gulf Cooperation Council region and the United States. Participating startups will also have an opportunity to present their products to investors and potential customers at a concluding demo day. Rahul Sasi, founder and chief executive officer of CloudSEK, said the programme was intended to help cybersecurity founders navigate challenges such as customer acquisition, product positioning and access to industry expertise.“AI has made it significantly easier to build sophisticated technology, but building a successful cybersecurity company still requires access to real threat intelligence, trusted customers and the right industry network,” Sasi said.

“AI让构建复杂技术变得容易得多,但打造一家成功的网络安全公司仍然需要获取真实的威胁情报、值得信赖的客户以及合适的行业网络,”萨西说。该公司表示,希望帮助参与的初创公司从产品开发阶段迈向获得首批付费客户。不过,该公司尚未披露分配给Launchpad的整体投资金额、除已公布关注领域之外的选择标准,以及规范其企业网络访问权限的商业条款。

The company said it wants to help participating startups move from product development to acquiring their first paying customers. However, it has not disclosed the overall investment allocated to Launchpad, the selection criteria beyond its stated areas of interest or the commercial terms governing access to its enterprise network.

阅读媒体原文 ↗