F5 agreed to acquire CalypsoAI for $180 million in cash to expand its security platform with purpose-built guardrails for AI applications. CalypsoAI, founded in 2018 and headquartered in Dublin, develops inference-layer defenses that protect against prompt injection, jailbreak attacks, and sensitive data leakage while ensuring compliance with regulations such as the EU AI Act. The acquisition is expected to close in F5’s fiscal Q4 2025.
The deal reflects growing enterprise demand for AI-specific security as companies embed generative and agentic AI into operations. CalypsoAI’s technology offers adversarial red-teaming at scale, adaptive risk scoring, and centralized governance across clouds and models. These capabilities will be integrated into F5’s Application Delivery and Security Platform (ADSP), enabling unified observability and policy control for apps, APIs, and AI models.
CalypsoAI has raised more than $40 million in venture funding from Paladin Capital Group, Lockheed Martin Ventures, and Hakluyt Capital. Its customers include Palantir and SGK. F5 said the transaction will be immaterial to its revenue and earnings but strategically important as enterprises demand AI security that is independent from model providers.
• F5 to pay $180 million in cash for CalypsoAI, closing expected by September 30, 2025
• CalypsoAI’s defenses include real-time threat detection, large-scale red-teaming, and runtime guardrails for AI inference
• Unified observability and compliance tools will integrate into F5’s ADSP platform
• Enterprise customers include Palantir; investors include Paladin Capital and Lockheed Martin Ventures
• Gartner highlights the need for model-agnostic AI Trust, Risk, and Security Management (TRiSM) solutions
“AI is redefining enterprise architecture—and the attack surface companies must defend,” said François Locoh-Donou, President and CEO of F5. “The addition of CalypsoAI will give enterprises the confidence to innovate faster and better protect critical data.”
🌐 Analysis: AI inference security is becoming a critical segment as generative models move into production at scale. F5’s move aligns with a broader industry trend where security vendors race to add model-agnostic guardrails, while hyperscalers such as Microsoft and Google are embedding native safeguards into their AI stacks. Independent platforms like CalypsoAI are positioned as neutral arbiters across multi-model and multicloud environments, a capability that will resonate with enterprises wary of vendor lock-in.
