Rein Security Raises $25M to Secure AI Agents

Key Takeaways

• <ul><li>Rein Security raised $25 million in Series A funding, bringing total raised to $35 million, to secure enterprise AI agents using runtime sidecar technology.</li><li>Rein Security's revenue grew eightfold and customer base fivefold since its January 2026 launch, with clients including Flex, Swimlane, Lemonade, and Dun & Bradstreet.</li><li>Jamie Dimon said cyber risks rose 10-fold after Anthropic's Claude Mythos model, which can reason, write code, and break advanced security systems.</li><li>Jamie Dimon warned AI could eliminate jobs faster than society can adjust and urged gradual phase-ins for major workforce changes.</li><li>DeepSeek is doubling its funding target to roughly $15 billion, with plans for a 2027 IPO on Shanghai's STAR Market.</li><li>Cloudflare open-sourced Clef, a set of AI decision models including a 27B multimodal model and a 9B Clef-Flash model with 38.8 ms median latency.</li><li>EnforceAuth launched Evidence IQ, a tool that records AI agent authorization decisions, available by end of Q4 2026.</li><li>Sanas introduced Supervised AI, where human supervisors oversee multiple AI agents handling customer calls in real time.</li><li>NVIDIA outlined a five-step workflow using OpenUSD and SimReady standards to prepare ABB YuMi robot assets for simulation.</li><li>Federal agencies are urged to update zero-trust controls for autonomous AI agents by assigning each agent its own identity and enforcing rules at machine speed.</li></ul>

Rein Security raises $25 million to protect enterprise AI agents

Rein Security raised $25 million in Series A funding to help companies secure AI agents. The round was co-led by Glilot Capital and Sienna Venture Capital. Rein uses runtime sidecar technology to monitor AI agents and block harmful actions. Since launching in January 2026, the company's revenue grew eightfold and its customer base grew fivefold.

Rein Security raises $25 million Series A for AI agent security

Israeli startup Rein Security raised $25 million in a Series A round, bringing its total funding to $35 million. The company was founded in 2024 by Matan Bar-Efrat and Netanel Rubin and has offices in Tel Aviv and New York. Rein's platform uses sidecar technology to watch AI agents in real time and stop harmful actions. Its customers include Flex, Swimlane, Lemonade and Dun & Bradstreet.

Rein Security raises $25M to secure insecure AI agents

Rein Security raised $25 million to address new security risks from autonomous AI agents. The company says AI agents create two types of risks that existing cybersecurity tools cannot handle. Rein's platform monitors agent behavior at runtime and blocks harmful actions without shutting down the agent. Its platform is already deployed by Flex, Swimlane, Lemonade and Dun & Bradstreet, securing thousands of agents executing millions of actions weekly.

Rein Security raises $25M to secure enterprise AI agents

Rein Security raised $25 million in funding led by Balderton Capital. The company plans to expand its AI-native runtime security platform for enterprise AI agents. Rein's technology provides real-time security monitoring and threat detection for AI systems. The company says its total funding now stands at $30 million.

Rein Security raises $25 million Series A for enterprise AI agents

Rein Security raised $25 million in Series A funding, bringing its total raised to $35 million. The round was co-led by Glilot Capital and Sienna Venture Capital with participation from Corner Ventures, Atlacle and RNP Capital Advisors. Rein's platform uses patented sidecar technology to give security teams visibility into AI agent actions and block harmful activity in real time. Since its January 2026 launch, revenue grew eightfold and customers grew fivefold, with clients including Flex, Swimlane, Dun & Bradstreet and Lemonade.

Rein Security secures $25 million to protect AI agents

Rein Security raised $25 million in a Series A funding round, bringing its total raised to $35 million. The company built an AI-native security platform that watches AI agents in real time. It already protects thousands of agents across many industries. The new money will fund product development and global hiring.

Jamie Dimon warns AI could eliminate jobs

JPMorgan Chase CEO Jamie Dimon said AI will change the job market whether people are ready or not. He warned that AI could move too fast for society to adjust. He said cutting truck driving pay from $150,000 to $25,000 too quickly could cause civil unrest. He advised phasing in changes gradually.

Jamie Dimon warns cyber risks from AI jumped sharply

JPMorgan Chase CEO Jamie Dimon said cyber risks rose 10-fold after Anthropic's Claude Mythos model. Mythos can reason in multiple steps, write code autonomously, and break advanced security systems. Anthropic shared the model with hundreds of groups to study its power. Dimon said AI created unknown vulnerabilities but urged calm and action.

DeepSeek doubles funding target to $15 billion

Chinese AI company DeepSeek is raising up to 100 billion yuan, about $14.9 billion, doubling its original goal. Investors include CATL, Geely Auto, Tencent, and Monolith Management. The money will fund AI computing infrastructure, including 160,000 Huawei AI accelerators in Inner Mongolia. DeepSeek plans a possible 2027 IPO on Shanghai's STAR Market.

New method cleans up AI-generated customer intent data

Researchers built a post-processing system for Large Language Models that cleans up messy customer intent extractions. The method removes duplicates, groups similar intents, and keeps provenance metadata. It uses clustering and constrained language models to produce one clear intent per group. The system also supports predicting intents for customers with limited data.

New method improves AI legal reasoning across long cases

Researchers proposed a training-free method called Test-Time Memory Evolution and Rubric-Aligned Collaboration to improve AI legal reasoning. The approach lets agents reuse experience from past cases and coordinate actions across roles and procedural stages. Tests on J1-EVAL and LegalWorld with five models showed consistent gains over baseline methods. The two components offer complementary benefits for experience adaptation and cross-role coordination.

Cloudflare releases Clef decision models for AI agents

Cloudflare open-sourced Clef, a set of AI models that choose between predefined options instead of generating text. The release includes a 27B multimodal model and a faster 9B Clef-Flash model with median latency of 38.8 ms. Clef can process text, JSON, images, and video and return probabilities for each option in one pass. Cloudflare also announced a fine-tuning service and plans to adapt Clef for tasks like support triage and bot classification.

EnforceAuth launches Evidence IQ for AI agent authorization

EnforceAuth announced Evidence IQ, a tool that records why an AI agent was allowed to take a specific action. It rebuilds past decisions using identity, data classification, risk signals, and policy rules. Evidence IQ can replay proposed policies on historical decisions and react after access is granted if context changes. The product will be available by the end of the fourth quarter of 2026 and works with existing identity providers, SIEM, and AI agent frameworks.

Five-step workflow creates SimReady robotics assets with AI

NVIDIA developers outlined a five-step workflow to prepare ABB YuMi robot assets for simulation using OpenUSD and SimReady standards. The process imports STEP files, matches visual appearance, configures physics properties like joints and collision geometry, validates against SimReady requirements, and runs pick-and-place tasks. Frontier AI models such as GPT-6 Astra help by writing Python code that calls Omniverse tools. Validation confirmed the asset met acceptance criteria for tested trajectories.

Sanas launches Supervised AI for customer conversations

Sanas introduced Supervised AI, a model where AI agents handle customer calls while a human supervisor oversees multiple conversations at once. The supervisor can steer the AI, take over a call, or hand it back without interrupting the customer. Every call begins with disclosure that the customer is speaking with AI, and all actions are recorded for audit. The platform can run inside a customer's own environment and supports containerized deployment on Sanas Fabric.

Federal agencies need zero-trust controls for autonomous AI agents

Federal agencies are shifting from AI tools that answer questions to AI agents that can take action across systems. These autonomous agents move faster than humans, so existing zero-trust security rules need to be updated. Experts recommend giving each AI agent its own identity, limiting its authority to specific tasks, and enforcing security rules at machine speed. Federal contractors should build these controls into their systems from the start to keep AI powerful but safe.

Sources

• Rein Security Raises $25 Million to Secure the AI Agents Enterprises Build and Stop the Ones That Attack Them
• Rein Security raises $25 million Series A as companies struggle to control their AI agents
• Rein Security raises $25M to rein in insecure AI agents
• Dealroom.co | Rein Security raises $25M to secure enterprise AI agents
• Rein Security Raises $25 Million Series A To Secure Enterprise AI Agents
• Rein Security Raises $25 Million to Guard AI Agents at Runtime
• Jamie Dimon Says AI Will Eliminate Jobs, 'It Is What It Is' — But Warns Cutting $150,000 Truck Driving Jobs To $25K A Year Would Bring 'Civil Unrest'
• JPMorgan Chase CEO Jamie Dimon Just Issued a Blunt Warning About Artificial Intelligence
• DeepSeek doubles funding target to $15 billion ahead of 2027 IPO: Who has committed and where will the money go?
• From High Recall to High Utility: Dataset-Adaptive Post-Processing of LLM-Generated Customer Intents
• Test-Time Agent Evolution for Long-Horizon Legal Reasoning
• Cloudflare Open Sources Decision Models for AI Agents
• AI Safety ≠ AI Security: EnforceAuth Announces Evidence IQ
• 5 Steps to Create SimReady Assets for Robotics with Frontier AI Models
• Sanas Introduces Supervised AI, Putting a Human at the Helm of Every AI Customer Conversation
• The zero-trust controls federal agencies need for autonomous AI