Originals · Startup · Startups & Fintech
CloudSEK opens a 12-week incubator for security founders and ships an AI gateway
The Bengaluru threat intelligence company will pick 10 early-stage cybersecurity startups for Launchpad, each eligible for up to $5,000 in AI model credits routed through its new KeyVigil gateway.

CloudSEK, the Bengaluru-based AI-powered cybersecurity and threat intelligence company, announced two things on October 29: CloudSEK Launchpad, a 12-week incubation programme for early-stage cybersecurity startups, and KeyVigil, a gateway that sits between applications or AI agents and the frontier AI models they call. Applications for Launchpad's first cohort are open until November 20, 2026.
The first cohort will take 10 companies. Each is eligible for up to $5,000 in AI model credits, delivered through KeyVigil. Founders get weekly sessions with CloudSEK's leadership on research, product, enterprise sales and fundraising, access to its threat intelligence resources and researchers, and introductions to potential customers and channel partners in India, the Gulf Cooperation Council region and the United States. The programme ends with a demo day in front of investors and prospective customers. Pre-seed and seed-stage startups can apply, as can idea-stage founders with a strong team and a clearly defined security problem.
CloudSEK lists its areas of interest as security for AI systems and agents, AI-native security operations, threat intelligence and digital risk, identity and fraud prevention, brand protection, and software supply-chain and code security. Founder and CEO Rahul Sasi said the aim is to help promising founders "get to their first ten paying customers." The company has not disclosed how much it is putting into Launchpad or the commercial terms behind access to its enterprise network, CNBC TV18 noted.
KeyVigil gives apps and agents one endpoint for several frontier models, so teams can switch providers without rewriting code. Master API keys are held in the gateway rather than in repositories, laptops or CI pipelines. Personal data, secrets and customer records are masked before a request leaves, prompt-injection and jailbreak attempts are flagged and blocked, and every call is logged with its cost, with spending limits per team, application or agent. CloudSEK says the product was built by the team behind AIVigil, its tool for finding exposed AI assets, and it is available now to enterprise customers.