CalypsoAI vs Prediction Guard

CalypsoAI was an AI security company founded in the US in 2018 by Neil Serebryany, its first CEO; it describes itself as founded in Silicon Valley, and its early press releases carry Washington datelines. It began with software for testing, validating and monitoring AI and machine learning models for national security, finance and other regulated customers, and moved to securing generative AI with CalypsoAI Moderator in 2023. Donnchadh Casey became CEO in August 2024. In October 2024 CalypsoAI named Dublin its global co-headquarters, alongside a US headquarters in New York; IDA Ireland says the CEO, the CTO and most of its R&D were based in Dublin. Paladin Capital Group led its $13 million Series A (July 2020) and a $23 million Series A-1 (June 2023), and F5 says it raised more than $40 million in venture funding. Investors also included Lockheed Martin Ventures and Hakluyt Capital. F5 announced on September 11, 2025 that it would acquire CalypsoAI for $180 million in purchase consideration, mostly cash. F5's 10-Q says the deal closed on September 26, 2025 and records $145.2 million in cash consideration. CalypsoAI is no longer sold separately: calypsoai.com now redirects to F5's AI Guardrails page. F5 introduced F5 AI Guardrails and F5 AI Red Team when it announced the close, and made both generally available in January 2026. Buyers should evaluate those products. **What it sold** By 2025 CalypsoAI sold an inference security platform with two main products. Inference Defend checked prompts and responses in real time to block prompt injection, jailbreaks and sensitive data leakage, using custom scanners that could be versioned, tested and enforced per project. Inference Red-Team ran automated and agent-driven multi-turn attacks, drawing on more than 10,000 new attack prompts a month, and produced a risk score. The platform logged AI interactions for audit, sent logs to Splunk, and ran as SaaS, on premises or hybrid. In August 2025 CalypsoAI said an on-premises version of Red-Team, for organizations with strict data residency or air-gapped requirements, would soon be available in early access. **Frameworks** In August 2025 CalypsoAI added an EU AI Act scanner package to Inference Defend, covering four Article 5 prohibited-practice categories: manipulation, social scoring, biometric data harvesting and surveillance. F5's deal announcement said its governance features help customers meet GDPR and EU AI Act requirements. We found no CalypsoAI mapping to the NIST AI RMF or ISO 42001.

Prediction Guard sells a self-hosted "sovereign AI control plane" for regulated organizations. It sits between applications and the models, agents and MCP tools they call, and applies access, safety and policy controls to each request inside the customer's own infrastructure. The company was founded in 2023 in Lafayette, Indiana; CEO Daniel Whitenack is a co-founder. It raised a $3.7 million seed round led by Sovereign Capital in May 2025, and Noblis Ventures made a strategic investment of undisclosed size in July 2025. At the time of the seed round it had eight employees and said it had 18 customers in health care, defense, transportation and manufacturing. The platform registers each agent with an identity and role and limits it to the models and tools that role needs. Runtime controls include PII masking, prompt injection detection, toxicity filtering, input sanitization, MCP tool scoping and human-in-the-loop approval; requests that fail a policy are blocked and logged. It produces an exportable AI bill of materials (CycloneDX) of models, guardrails and MCP servers, runs model security scans and streams security events to SIEM tools such as Splunk, Datadog and Grafana. Applications call it through OpenAI- and Anthropic-compatible APIs. It deploys into a customer VPC on AWS, Azure or Google Cloud, on Kubernetes on premises, or air-gapped. Prediction Guard says its control overhead is under 200 ms per request. **AI compliance coverage** Prediction Guard says its policies align with the NIST AI RMF, NIST AI 600-1, the OWASP Top 10 for LLM and agentic applications, OMB M-26-04 and AIUC-1, with one-click baselines and an exportable "proof of alignment" for the NIST AI RMF. A company blog post (August 2026) adds the EU AI Act and ISO/IEC 42001. The same post labels its OWASP mapping "self-published" and not independently verified, and notes that the audit log records enforcement events rather than full replayable traces. We found no published control-by-control mapping. The homepage shows badges for a SOC 2 Type 2 report by Sensiba and HIPAA compliance; there is no public trust center. **Who it's for** Regulated enterprises and public sector bodies that want AI guardrails and agent controls to run inside their own network, including disconnected environments. It is sold to government through Carahsoft. Its homepage quotes executives at Noblis (also an investor), SimWerx and iBase-t.