Hello, stay ahead with CISO Brief 🚀

🚀 NVIDIA's Cosmos3-Edge, Cosmos3-Nano, and Cosmos3-Super are now available in Amazon SageMaker JumpStart, expanding AWS's foundation model offerings for physical AI. These omnimodal world models enable robots, autonomous vehicles, and vision AI to perceive, reason, plan, and act. Customers can deploy the models from the SageMaker JumpStart catalog or use the SageMaker Python SDK for integration.

🆕 Amazon SageMaker JumpStart now offers Meta's Muse-Glimmer-30B and Alibaba's Qwen 3.8-27B, expanding foundation model choices for enterprise deployments. Muse-Glimmer-30B targets autonomous agentic workflows with a ViT-G/14 perception encoder, 131K+ context, and selectable reasoning strength, while Qwen 3.8-27B provides strong multimodal reasoning and coding performance with a 262K context (scalable via YaRN). Customers can deploy either model from the SageMaker JumpStart catalog or using the SageMaker Python SDK.

🔔 Amazon Redshift now accepts Amazon Kinesis Data Streams (KDS) records up to 10 MiB, a tenfold increase from the prior 1 MiB limit. This update lets teams stream much larger payloads directly into Amazon Redshift without splitting records, simplifying ingestion pipelines and enabling new high-volume, large-record use cases. The feature is available now in all commercial AWS regions where Amazon Redshift is offered.

🔧 Amazon Redshift integrates with the Agent Toolkit for AWS to let AI agents like Claude Code, Kiro, and Cursor build, query, troubleshoot, and migrate Redshift data warehouses and data lakes. The integration pairs the AWS MCP (Model Context Protocol) server for authenticated AWS API execution with curated Redshift skills—packages of procedures and reference material—to reduce query errors and streamline common tasks. Skills cover SQL syntax, metadata discovery, data loading, materialized view best practices, function and type guidance, extensions, and end-to-end migration workflows for both provisioned clusters and Serverless workgroups. The capability requires no infrastructure changes, is available at no extra charge in supported Regions, and can be enabled via the aws-data-analytics plugin or discovered at runtime by agents with MCP access.

🚀 Amazon EC2 X8i instances are now available in Europe (Milan) and Europe (Spain). These instances use custom Intel Xeon 6 processors exclusive to AWS and are SAP-certified, offering up to 43% higher performance, 1.5x more memory capacity (up to 6TB), and 3.3x more memory bandwidth versus prior generation X2i. X8i is optimized for memory-intensive workloads like SAP HANA, large databases, data analytics, and EDA, and is available in 14 sizes including bare metal options.

🔒 The article argues that poorly designed external AI guardrails can erode defenders' advantages by blocking or delaying agentic SOC investigations, giving attackers time to succeed. It urges organizations to retain operational sovereignty by embedding customizable guardrails within their own systems and testing LLMs against real workflows. Cisco Talos evaluated many models and stresses balancing efficacy, cost, speed, and consistency when selecting AI for security.

🚀 Amazon Bedrock AgentCore is now available in US West (N. California) and Asia Pacific (Hyderabad), enabling developers to build and run agents closer to end users. The platform supports agent runtime, identity and access control, policy management, session persistence, tool connectivity, evaluations, and observability at launch. AgentCore emphasizes secure infrastructure-level enforcement so agents cannot bypass controls. Regional availability reduces latency and improves performance for local customers.

🔍 Amazon Connect Customer now offers generative AI summaries, real-time call analytics, and real-time rules in the Africa (Cape Town) Region. These features complement existing post-contact conversational analytics to provide AI-driven insights that improve customer experience, agent performance, and operational efficiency. Supervisors and agents receive concise AI-generated summaries after contacts, while live transcription and automated rules enable monitoring and immediate interventions during interactions.

🔁 AWS Elastic Disaster Recovery (AWS DRS) introduces Recovery Plans to automate the sequenced launch of multi-server applications during recoveries and drills. You define server groups and ordered steps once, with configurable wait times, approvals, and non-disruptive drill mode to validate procedures. Recovery Plans provide real-time monitoring, reduce recovery time, and remove manual dependency tracking. The feature is available in all Regions where AWS DRS is offered at no extra cost beyond standard usage.

🔄 Amazon Connect Customer now auto-refreshes scheduling metrics on the scheduling page to give managers immediate visibility into the impact of schedule updates. When changes such as a team meeting for several agents are made, metrics like net available headcount and projected service level update automatically. This feature helps workforce managers assess staffing coverage and make faster, informed decisions throughout the day.

🔒 Amazon FSx for NetApp ONTAP now supports copying backups within and across AWS Regions and across trusted accounts in your AWS organization. This capability lets you store secondary backup copies in different Regions or accounts for improved business continuity, data protection, and compliance. Backups remain secure, incremental, and durably stored across Availability Zones, and you can copy both new and existing backups in all Regions where the service is available.

🔔 AWS Backup now supports copying Amazon FSx for NetApp ONTAP backups across AWS Regions and accounts. You can copy FSx for NetApp ONTAP backups to another AWS Region, another AWS account, or both using policy-based backup plans or on-demand copy jobs. Copies can be managed centrally and automated across your organization with AWS Organizations. This capability is available in all commercial Regions where both services are offered.

🔒 Check Point Research warns that AI is accelerating cyberattacks, requiring faster and broader defensive measures. The company publicly supports OpenAI’s open letter calling for collective action across industry, government, and AI labs to expand access to security tools, share intelligence, and help under-resourced organizations. Check Point emphasizes its 30-year commitment to prevention and pledges to support customers adopting AI securely while contributing to shared defenses and practical remediation efforts.

🧠 Cloudflare’s Big Pineapple platform stores over 250 billion DNS cache entries and reduced per-entry memory by over 50% through five successive storage optimizations. These changes freed roughly 100 TB of RAM across the fleet while improving insert throughput by 43% and lowering lookup latency by 19%. The post details techniques like replacing Vec/String with Box, packing section offsets, inferring record owners, boxing large enum variants, and storing record data as a contiguous byte buffer to improve memory locality and reduce allocations.

🔎 This article compares self-managed PostgreSQL and managed PostgreSQL services, explaining how each model balances control, responsibility, cost, and operational effort. It highlights the operational tax of self-hosting versus the reduced platform responsibilities provided by managed services like Azure Database for PostgreSQL. The piece outlines implications for resilience, security, identity integration, backups, maintenance, and scalability to help technical decision makers choose the right operating model.

🛠️ Amazon Elastic VMware Service (Amazon EVS) now supports the i7i.metal-48xl bare-metal Amazon EC2 instance, providing a higher core-count option with 5th generation Intel Xeon Scalable processors. This instance class delivers up to 23% better compute and >10% better price performance versus i4i instances, plus larger memory and storage to host more VMs per EVS host. The feature is available in Regions where both Amazon EVS and EC2 i7i are offered and enables VCF 9.x capabilities and onboarding automations.

🛡️ This post shows how to extend Amazon Bedrock Guardrails beyond the model boundary to tool calls and external data flows using the Strands Agents SDK lifecycle hooks. It explains three validation checkpoints—BeforeInvocationEvent, BeforeToolCallEvent, and AfterToolCallEvent—that run without changing agent or tool logic. The guide includes implementation details using boto3 and a reusable GuardrailHook that can be scoped per tool and deployed to Bedrock Agent Core Runtime. Examples and testing steps help validate guardrail behavior.

🧾 Found a bank card? Don’t post photos or hand it to strangers. Contact the issuing bank, report the card as found, and then destroy it to prevent fraudulent use and avoid becoming a person of interest. For multi-use cards tied to schools or transit, return them to the institution’s office or security. Avoid keeping the card or leaving it where anyone can pick it up.