Cybersecurity company Rapid7 is cutting around 12 per cent of its workforce. The company is restructuring its organisation and redirecting resources towards its core platform and AI-led security capabilities.
The restructuring began in the second quarter of 2026 and is aimed at simplifying the company’s organisational structure and aligning teams more closely with its Core Platform Solutions. Around 12 per cent of employees have been informed that their roles will be affected.
Rapid7 expects the restructuring to result in charges of approximately $10 million to $11 million. Most of the costs are expected to come from employee severance and related expenses.
The workforce reduction comes as Rapid7 accelerates a broader shift towards an AI-first approach to cybersecurity. The company has strengthened its leadership team with the new appointments.
Rapid7 is also expanding its use of generative and agentic AI in security operations. Through OpenAI’s Trusted Access for Cyber programme, the company is incorporating models including GPT-5.5 into its Agentic SOC workflows. Rapid7 said the technology has helped speed up the process of analysing security telemetry and cut false-positive queue times by 25 per cent.
The company has also joined Anthropic’s Project Glasswing, where it will work on defensive engineering, detailed code reviews and automated vulnerability remediation.
The restructuring comes as Rapid7 manages modest declines in key financial metrics. The company reported quarterly revenue of $210.9 million and annual recurring revenue (ARR) of $824 million, both slightly lower than the previous year.
Despite the workforce reduction and restructuring costs, Rapid7 generated $31.9 million in free cash flow during the quarter. It ended June with $702.6 million in cash, cash equivalents and government securities.
The changes indicate a broader shift in the company’s workforce strategy, with resources being moved away from existing organisational structures and towards AI-driven products and cybersecurity capabilities.
