August 2025 Summaries

Tetragon Financial Group, a hedge fund managing approximately $30 billion, transitioned to Elastic Cloud Serverless to enhance AI applications integral to its investment strategy, specifically for optimizing the storage and retrieval of data embeddings. This migration aimed to address challenges such as scalability, performance, and management overhead that arose from previous solutions like ChromaDB and Pinecone. Elastic Cloud Serverless offered a managed solution that enabled the team to scale efficiently and improve query performance by 200%, while freeing developers to focus on core tasks without infrastructure concerns. The transition was swift and facilitated by Elastic's comprehensive documentation and support, resulting in improved data strategy and reduced management overhead. This change aligns with Tetragon's ongoing efforts to expand AI capabilities, including future plans to leverage Elastic's machine learning features, while ensuring familiarity with the Elastic ecosystem and AWS infrastructure.

In environments where internet access is restricted, deploying Elastic Agents can become a complex process due to the need for manual artifact downloads and updates. Pete Ward, a senior engineer, addresses this challenge with the creation of the artifacts-bundler tool, which streamlines the download and management of Elastic Stack components for offline use. This tool automates the fetching, packaging, and distribution of necessary binaries and security artifacts, allowing organizations to deploy Elastic Agents in air-gapped environments efficiently. It supports both traditional server setups and containerized deployments, providing flexibility and scalability. By simplifying artifact management, the tool reduces manual labor and potential errors, enabling organizations to focus more on configuration and optimization rather than logistical challenges. The artifacts-bundler is offered on GitHub, encouraging community adoption and further development to meet the needs of various sectors, including government, healthcare, and critical infrastructure.

Burgan Bank Türkiye, a subsidiary of the Kuwait Projects Company Group, has significantly enhanced its observability and security capabilities by implementing Elastic's solutions. Initially leveraging Elastic on OpenShift, the bank transitioned to a bare metal deployment to overcome storage limitations and improve performance, ultimately establishing a modern observability stack. Led by senior software manager Burak Düz, the bank developed a robust Elasticsearch architecture, integrating machine learning to proactively detect anomalies and maintain system health. The introduction of an on-premise AI assistant further streamlined transaction investigations, enhancing security by ensuring sensitive data remains within its local network. As a centralized monitoring hub, Elasticsearch reduced incident response times and facilitated faster root cause analysis, while sophisticated alerting, data masking, and integration with tools like Zabbix and Grafana ensured compliance and privacy. Burgan Bank's strategic partnership with Gantek and ongoing innovations, such as integrating APM data with SolarWinds, exemplify its commitment to scalable and secure IT systems to support its growth in the MENA region.

Version 9.0.6 of the Elastic Stack was released on August 28, 2025, and users are recommended to upgrade from the previous version, 9.0.5, due to the inclusion of fixes for potential security vulnerabilities. The release addresses various issues, and a complete list of changes for each product is available in the release notes. For additional details about the security fixes, users are advised to consult the security advisory.

Version 8.19.3 of the Elastic Stack has been released, addressing a specific bug in Kibana's Canvas application that caused progress charts to display an "Unable to load page" error, affecting both new and existing charts. Other chart types were not impacted by this issue. Users are encouraged to upgrade to this latest version over the previous 8.19.2 release. For a comprehensive list of fixes and changes included in this update, users are advised to consult the release notes.

Version 9.1.3 of the Elastic Stack was released on August 28, 2025, offering important updates and fixes over its predecessor, 9.1.2. This release addresses potential security vulnerabilities, underscoring the importance of upgrading to enhance system security. Additionally, it includes a fix for a bug in Kibana's Canvas application, which previously caused an "Unable to load page" error in progress charts, although other chart types remained unaffected. For a comprehensive list of changes and detailed issue resolutions, users are encouraged to consult the release notes.

Elastic has announced the availability of Elastic Cloud Serverless on Microsoft Azure in regions including North Europe (Ireland), Australia East (New South Wales), and West US (Washington), offering a streamlined way to scale security, observability, and search solutions without managing infrastructure. Built on the Search AI Lake architecture leveraging Azure Blob Storage, this service allows independent scaling of workloads through decoupled compute and storage while maintaining high performance with low-latency query capabilities. Elastic Cloud Serverless eliminates the need for operational management like cluster provisioning and performance tuning, offering a flexible usage-based pricing model for data ingested and retained in Elastic Security and Observability products or compute resources in Elasticsearch. Users can easily create a serverless project via the Elastic Cloud console by selecting the project type and region, with projects provisioned within minutes. Elastic is also working on expanding to more Azure regions and enhancing features, emphasizing a future where search, security, and observability can be achieved efficiently without compromising on speed, scale, or cost, and urging users to explore serverless possibilities or start a free trial on Azure. The post also notes potential involvement of third-party AI tools, reminding users to be cautious with sensitive data due to privacy and security concerns.

Version 8.18.6 of the Elastic Stack was released on August 28, 2025, with recommendations to upgrade from version 8.18.5 to benefit from the latest updates. Users are encouraged to consult the release notes for a detailed list of fixed issues and changes across each product included in this version.

In 2025, The Hague is emerging as a center of digital sovereignty, where technology intertwines with public values to foster trust in government digital strategies. As European countries transition from policy visions to regulated realities, compliance frameworks like NIS2, DORA, and the EU AI Act are shaping the landscape of cyber resilience and responsible AI. The Dutch National Digital Strategy emphasizes the importance of open source and open standards for transparency and interoperability, while initiatives like sovereign clouds ensure data sovereignty and legal compliance. Elastic's offerings, such as its Search AI Platform, are designed to enhance digital infrastructure resilience and AI explainability, aligning with public values. The blog introduces a year-long series of discussions on these themes, emphasizing that trust in government systems is achieved through effective execution. The upcoming Elastic{ON} event in Amsterdam will continue this conversation, highlighting how Elastic's solutions are transforming data management and AI applications in the public sector.

Elasticsearch, a Java-based search and analytics engine built on Apache Lucene, relies on the Java Virtual Machine (JVM) for its operation across various platforms. The JVM abstracts the underlying operating system and hardware, enabling Java applications to run platform-independently, while handling crucial tasks such as memory management and garbage collection. The blog emphasizes understanding JVM memory management and troubleshooting issues like java.lang.OutOfMemoryError through garbage collection processes, including the transition of objects between the young and old generations in heap memory. Elasticsearch now uses the Garbage-First (G1) garbage collector for improved performance over its predecessor, the concurrent mark sweep (CMS) garbage collector. Administrators can monitor JVM states and memory metrics using Elasticsearch APIs or the jstat tool for real-time JVM statistics. The blog advises against altering default JVM settings without consulting Elastic Support, highlighting that these defaults are generally sufficient for most use cases. Further insights and potential tuning of JVM settings will be discussed in subsequent posts.

The Elasticsearch storage engine team has introduced significant enhancements in versions 8.19 and 9.1, focusing on storage efficiency and indexing performance through features like time series data streams (TSDS) and logsdb index mode. These advancements aim to address growing data volumes faced by DevOps and SRE teams, offering up to a 70% reduction in data storage footprint and improved indexing throughput. This translates to lower operational costs, improved service availability, and enhanced compliance capabilities, empowering teams to make informed decisions based on actual operational experience. The changes help eliminate data silos, enabling unified data access and analysis through machine learning, which facilitates faster incident resolution. Additionally, these improvements support seamless compliance and reporting by allowing data to be archived cost-effectively without going offline. Elasticsearch's advancements also bolster AI adoption, providing context-rich data for better insights and decision-making, ultimately offering customers an opportunity to optimize their infrastructure and prepare for AI integration.

Security operations centers (SOCs) are facing significant challenges due to the overwhelming volume of alerts, with a large number of false positives causing analyst burnout and inefficiencies. The integration of AI-driven tools and automation, such as Elastic Security's Attack Discovery and AI Assistant, is transforming how alerts are managed by reducing noise and enabling analysts to focus on genuine threats. These tools enhance detection by identifying actual attacks, prioritizing alerts based on risk factors, and providing context-aware guidance that accelerates response times and improves the precision of threat investigations. This shift from manual alert handling to strategic, AI-powered processes not only reduces the workload on analysts but also improves the overall security posture by enabling more proactive and insightful threat management. The approach leverages AI to turn high alert volumes into actionable intelligence, thereby enhancing the efficiency and effectiveness of defense security teams.

Tipalti, a global leader in payables automation, leveraged AutoOps to transition from manual to automated Elasticsearch performance management, achieving a 10% annual cost saving and enhancing system reliability. Initially, Tipalti faced challenges with their manual monitoring approach, which involved using dedicated clusters for each environment and required significant time for issue diagnosis. The adoption of AutoOps provided immediate visibility and actionable intelligence, allowing the team to optimize their indexing strategy, reduce CPU and RAM usage, and decommission outdated infrastructure. Additionally, AutoOps' machine learning-powered anomaly detection minimized alert fatigue, empowering developers with real-time performance data and fostering a collaborative engineering culture. This strategic shift enabled Tipalti's small team to manage a large-scale data platform efficiently, focusing on proactive optimization and fostering an innovative, data-driven environment.

Kseniia Ignatovych and Isai Anthony's blog post offers a comprehensive guide to creating custom detection rules in Elastic Security, focusing on using Elasticsearch Query Language (ES|QL) and Elastic AI Assistant to improve threat detection and provide context for analysts. The authors emphasize the importance of developing detection logic that captures behavioral patterns rather than relying solely on static indicators, thereby ensuring that detections remain effective despite evolving threats. The blog outlines a four-step process for creating and deploying detection rules, including defining a focus area, refining detection logic, previewing and testing rules, and deploying them to production. Through examples involving AWS CloudTrail, the post illustrates techniques for identifying potential threats and evasion tactics by querying for specific API calls and analyzing log data. The article also highlights best practices for rule creation, such as mapping detections to MITRE ATT&CK, providing investigation guides, and setting up automated responses to reduce mean time to respond (MTTR). Elastic Security's community involvement is encouraged, inviting users to share custom rules and engage with Elastic's community resources for continual improvement of security operations.

The article by Woody Walton discusses the evolution and importance of Zero Trust (ZT) security paradigms, emphasizing the need for organizations to adopt this approach due to the complexities of modern interconnected IT environments. Unlike traditional perimeter-based security models, Zero Trust requires continuous verification, microsegmentation, and identity-centric security to protect against breaches. The article highlights the challenges in implementing Zero Trust, such as system complexity and speed, and introduces Elastic's Search AI Platform as a solution for integrating and operationalizing data across various pillars of Zero Trust. Elastic offers a distributed data platform that enables real-time search and analytics, which can assist in dynamic risk scoring and decision-making processes essential for Zero Trust operations. The platform's ability to collect and analyze data from diverse sources ensures a comprehensive approach to security, moving beyond point-to-point integrations and enhancing the organization's Zero Trust posture.

The blog post from the Elastic Platform Team delves into the evolving landscape of artificial intelligence (AI), contrasting traditional AI with generative AI to help IT leaders make strategic decisions. Traditional AI, which is rule-based and deterministic, excels in automating repetitive tasks and solving well-defined problems, whereas generative AI creates new content by learning patterns from vast datasets, enabling more creative and adaptable solutions. The post explores various applications and use cases for both types of AI, such as fraud detection and predictive analytics for traditional AI, and content generation and personalized recommendations for generative AI. It highlights the different implementation requirements, with generative AI needing extensive computational power and data, while traditional AI relies on structured data and predefined algorithms. The post also addresses the ethical and security challenges associated with AI adoption, noting the regulatory differences between the US and the EU, and it emphasizes the transformative potential of AI technologies if developed responsibly. Elastic positions itself as a key player in AI with its advanced search capabilities and machine learning integration, offering tools for enhancing customer support, employee efficiency, and security operations.

Elastic's response to claims of a zero-day vulnerability in its Elastic Defend software highlights its commitment to security and transparency. After receiving a proof of concept and crash dumps from a researcher, Elastic engaged directly with the researcher and involved a neutral third party to ensure an unbiased review. The crash dumps were traced to a known stability issue in an earlier version, which had already been fixed prior to the researcher's report. The proof of concept did not demonstrate new vulnerabilities but involved the misuse of administrator rights to load a custom unsigned driver, leading to a blue screen error unrelated to Elastic Defend. Elastic emphasizes the importance of staying current with software updates and adhering to security best practices, such as least-privilege principles and enabling Secure Boot. Despite the researcher's claims lacking evidence of reproducible exploits, Elastic continues to value its collaboration with the security community, maintaining a proactive bug bounty program and robust security development framework. The company conducts thorough vulnerability assessments and publishes security advisories when necessary, ensuring users are informed of any significant security issues and associated remediations.

Elastic has incorporated generative AI fields into the Elastic Common Schema (ECS), ensuring compatibility with OpenTelemetry (OTel) to streamline and standardize monitoring across various generative AI vendors. This integration allows users to normalize data from different sources using a common set of fields for storing event data in Elasticsearch, such as logs and metrics. With the surge in generative AI applications, there is a rising need for fields capturing the context of AI model requests and responses, including timestamps, token counts, and moderation activities. These enhancements facilitate observability, security, and log normalization for generative AI applications. ECS's collaboration with OTel provides a turnkey solution for monitoring and securing AI infrastructures while maintaining compatibility with existing ECS workflows. The initial batch of ECS GenAI fields is available in beta, offering users the opportunity to start integrating them into their workflows, though their final release is subject to change at Elastic's discretion. Users are advised to exercise caution when using third-party AI tools with sensitive information, as Elastic disclaims responsibility for third-party tool content and security.

Elastic's IT Help Desk Agent, integrated with Salesforce Service Cloud, leverages the Elastic Search AI Platform to enhance IT support by offering proactive, real-time assistance that transcends traditional ticket queues. The platform incorporates generative AI (GenAI), enabling seamless integration of instant AI-driven support within Salesforce workflows, allowing for either immediate resolution or escalation to human agents for complex issues. This dual-path approach improves ticket quality and reduces manual ticket volume while using curated enterprise knowledge from sources like Jira and ServiceNow. The technical architecture includes Lightning Web Components and MuleSoft-powered REST APIs, facilitating real-time knowledge retrieval through ElasticGPT. Elastic's internal deployment serves as a real-world example of AI's operational value, demonstrating significant improvements in support efficiency and response times, while Elastic Observability ensures performance monitoring, maintaining trust in automation. The integration of Elastic's AI capabilities with native enterprise platforms like ServiceNow and Zendesk offers organizations the opportunity to streamline their support systems and enhance user experiences.

Dominik "Dom" Toepfer, a principal software engineer at Elastic, combines his passion for engineering with a flair for creativity both in his professional and personal life. Known for his work on generative AI and search technologies, Dom emphasizes the importance of collaboration and community, often drawing parallels between his work and his home projects, like crafting chili sauces. His career at Elastic began with a curiosity sparked by a book, leading him through various roles and granting him broad insights into the company. Dom's workspace, carefully curated for inspiration and productivity, reflects his belief in the significance of environment and tools, including noise-cancelling headphones and a specifically configured digital toolkit. He sees AI as a collaborative partner, aiding in complex problem-solving and routine tasks, while his projects at Elastic, particularly involving retrieval augmented generation (RAG), aim to create meaningful impacts for both internal teams and customers. Dom values the feedback loop with users, which helps refine these tools and experiences, ultimately enhancing their effectiveness. His journey at Elastic, marked by a focus on building impactful features rather than grappling with hardware, has been both challenging and rewarding.

Elastic has joined the AWS Zero Trust Accelerator for Government (ZTAG) program, marking a strategic collaboration aimed at enhancing security information and event management (SIEM) integration tailored for the US federal government's Zero Trust architecture. This partnership combines Elastic’s data-centric approach with AWS’s Zero Trust framework to support federal agencies in accelerating their Zero Trust adoption efforts, following the CISA’s maturity model. The collaboration offers comprehensive visibility across Zero Trust ecosystems, strengthens security by supporting mandates and enhancing SIEM capabilities, and facilitates a seamless transition from legacy systems with its open, cloud-native platform. Elastic’s solution is designed to unify data across different systems, reduce silos, and support interoperability while maintaining secure, long-term visibility and compliance with FedRAMP standards.

Elastic has been awarded the 2025 Google Cloud DORA Award for its advancements in applying DORA principles to enhance software delivery and operational performance, particularly through AI collaboration with Google Cloud. DORA, a Google Cloud program, focuses on identifying capabilities that drive software performance, revealing that AI significantly transforms the software development lifecycle. Elastic's integration of AI tools alongside DORA practices has improved key performance metrics like deployment frequency and lead time for changes, and reduced change failure rates and mean time to recovery. This partnership has resulted in measurable business impacts, including cost savings, operational efficiency, and sustainability improvements, positioning Elastic as a leader in AI-driven software solutions. Elastic’s vision centers on empowering enterprises with production-ready AI capabilities, enhancing both customer solutions and internal productivity, while also encouraging the careful use of AI tools to ensure data security and compliance.

Version 9.1.2 of the Elastic Stack has been released, with recommendations for users to upgrade to this latest version over previous ones, such as version 9.1.1. For a comprehensive understanding of the improvements and issue resolutions in this update, users are directed to consult the release notes, which detail the changes for each product within this version.

Version 8.19.2 of the Elastic Stack has been released, with recommendations for users to upgrade from the previous version 8.19.1. For comprehensive details regarding the issues addressed and the changes implemented in this latest version, users are encouraged to consult the release notes.

Elastic Stack version 8.17.10 has been released, with recommendations for users to upgrade from previous versions, specifically 8.17.9, to benefit from the latest fixes and improvements. For comprehensive details on the issues addressed and a complete list of changes for each product included in this version, users are encouraged to consult the release notes.

Version 9.0.5 of the Elastic Stack was released on August 12, 2025, by Panagiotis Sfendourakis, featuring updates and improvements over the previous version, 9.0.4. Users are encouraged to upgrade to this latest version to benefit from the fixes and changes implemented. Detailed release notes are available for users seeking more information about the specific issues addressed and the enhancements made in each product included in this update.

Version 8.18.5 of the Elastic Stack has been released, and users are encouraged to upgrade from the previous version, 8.18.4. The release includes fixes and a comprehensive list of changes for each product, which can be reviewed in the release notes for further details.

The collaboration between Dell, NVIDIA, and Elastic aims to enhance creative industries like media, gaming, design, and animation by integrating Elasticsearch's vector database with the Dell AI Data Platform, offering a powerful infrastructure for instant access to vast unstructured data. This integration supports modern creative workflows by enabling fast semantic search and retrieval of data, crucial for production pipelines that involve machine learning, AI model training, and advanced analytics. Elasticsearch is designed to manage petabytes of data and provides lightning-fast search capabilities, supporting everything from animation files to video segments. Recent innovations like Better Binary Quantization and ACORN improve memory-efficient vector search and high-speed retrieval, respectively. The combined efforts promise to deliver an integrated, GPU-accelerated experience, boosting creativity and enabling smarter editorial tools, compliance checks, and personalized experiences. However, the release and functionality of these features remain at the sole discretion of Elastic, and users are cautioned about the privacy implications of using third-party AI tools.

In the era of AI-driven financial crime, financial institutions are urged to unify fraud, security, and compliance functions to enhance resilience, trust, and agility. IDC's report, sponsored by Elastic and AWS, underscores the necessity of moving beyond siloed risk functions to embrace cohesive, AI-powered platforms that integrate fraud prevention, cybersecurity, and regulatory compliance. Traditional threats have evolved, with generative AI enabling new forms of document forgery and identity spoofing, necessitating real-time, cross-channel detection and response capabilities. Customer expectations for seamless, secure digital experiences are rising, and fintechs are setting new benchmarks for agility and innovation. Legacy systems often hinder resilience, creating fragmented infrastructure and disjointed tools across various functions. Elastic's Search AI Platform aims to transform these fragmented defenses into a unified system, providing real-time intelligence through observability and security capabilities. This approach supports financial institutions in reducing operational complexity, enhancing customer trust, and meeting regulatory demands, ultimately fostering a more resilient and adaptive banking environment.

Elastic Security has introduced several enhancements to its platform, including the Agentic Query Validation workflow, Attack Discovery persistence, and automated scheduling and actions, all aimed at improving security operations through AI. These updates, available in Elastic Security versions 8.19 and 9.1, are designed to enhance reliability, automation, and efficiency in security investigations. The Agentic Query Validation ensures Elasticsearch Query Language queries are error-free before execution, reducing troubleshooting time and enhancing accuracy. The platform's AI Assistant now features time awareness, allowing users to access team schedules for better incident management. The Elastic AI Assistant is also accessible throughout the Elastic ecosystem, facilitating seamless collaboration and faster decision-making. Attack Discovery now retains historical findings for extended periods, aiding long-term threat analysis and collaboration. Additionally, the new scheduling and automated actions allow for continuous monitoring and response to threats, ensuring security operations are proactive and efficient. Elastic emphasizes the importance of these updates in providing deeper insights and faster actions for security teams, while also cautioning users about the responsible use of AI tools, particularly regarding data privacy and third-party tool interactions.

Version 9.1.1 of the Elastic Stack was released on August 7, 2025, with recommendations for users to upgrade to this latest version over the previous 9.1.0 release. The update addresses various issues, and users are encouraged to refer to the release notes for comprehensive details on the fixes and changes made to each product within the stack.

Version 8.19.1 of the Elastic Stack has been released, with recommendations to upgrade from the previous version 8.19.0. For comprehensive details on the fixes and changes included in this release, users are advised to consult the release notes.

In the realm of cybersecurity and observability, enterprises often face a challenge of balancing extensive data logging for security purposes with the high costs associated with storing and managing such data. This issue is particularly prominent with PowerShell logging, where comprehensive script block logging is vital for threat detection but can lead to overwhelming data volumes and expenses. A modern solution involves using the Elastic Stack and the Elasticsearch Query Language (ES|QL) LOOKUP JOIN command to implement a data deduplication strategy. This approach involves hashing script data, storing it once, and using lightweight references for each execution, drastically reducing storage needs while maintaining full analytic capabilities. The new strategy shifts from traditional data ingestion models to a more cost-efficient "enrich at query time" paradigm, allowing analysts to access complete context on demand without compromising on security visibility. This innovative method not only addresses data storage challenges but also enhances the efficiency of forensic investigations by seamlessly integrating with existing security frameworks and detection rules.

Elastic has introduced the Elastic AI SOC Engine (EASE), which aims to enhance security operations by integrating advanced AI capabilities into existing SIEM and EDR systems without requiring a full migration to next-generation solutions. EASE helps security analysts detect hidden, complex threats more efficiently by reducing investigation time and alert fatigue through intelligent AI-driven alert correlation. It offers agentless integration with third-party platforms like Splunk, Google Security, Microsoft Sentinel, and others, allowing immediate application of AI analysis to alerts. The system provides a context-aware AI assistant that enriches investigations with organization-specific knowledge, supporting natural language queries and enhancing response workflows. Additionally, EASE offers transparent AI with model flexibility, operational metrics for tracking efficiency gains, and automated response actions to accelerate threat containment. This solution allows security teams to improve threat detection, reduce response times, and enhance security posture while maintaining current investments, with the option to transition fully to Elastic's open, AI-driven SIEM in the future.

Modern IT infrastructure has evolved dramatically, necessitating advanced observability tools to manage its complexity and scale effectively. The use of AI and machine learning (ML) is crucial for enhancing software development and orchestration in this context, addressing challenges like data volume and signal correlation to improve root cause analysis. OpenTelemetry (OTel) emerges as a pivotal development, offering a standardized method for collecting and integrating diverse observability data, minimizing vendor lock-in, and enhancing data value through better metadata application. Effective observability now hinges on four foundational capabilities: cost-effective storage, standardized data collection, signal correlation through unified metadata, and ML and AI-driven tools for knowledge democratization and actionable insights. These elements collectively aim to maintain a consistent mean time to resolution, despite the growing complexity of infrastructure, with future advancements likely focusing on storage innovations and AI-based workflow automation.

In the blog post "Harnessing Elastic APIs for custom AI-driven SOAR" by Kevin Umsted, the focus is on how Elastic Security's open, API-first approach enables the creation of customized workflows to enhance security operations. The article describes how Elastic's APIs, in conjunction with AI and collaboration tools like Slack, can be leveraged to automate and refine response playbooks, addressing challenges such as missing service level objectives for critical alerts in a Security Operations Center (SOC). The post explains a scenario where critical alerts are detected and managed using Elastic's APIs, with the AI Assistant generating actionable summaries and recommended commands, which are then communicated to analysts via Slack for approval and execution. The integration of Python and Windows Remote Management for command execution, along with the creation of auditable cases in Elastic Security, are highlighted as key components of improving operational workflows across the SOC. The piece underscores the flexibility and power of Elastic's platform to build tailored workflows and improve security operations by embedding AI and automation into existing infrastructures.

The White House's "Winning the Race: America’s AI Action Plan" outlines a strategic vision for the United States to secure its leadership in artificial intelligence by focusing on innovation, security, and a robust data foundation. It emphasizes the importance of a transparent and searchable data ecosystem as essential for managing and understanding the data that drives AI, addressing challenges such as bias detection, cybersecurity, and supply chain resilience. The plan advocates for the creation of a national "AI Evaluations Ecosystem" to build public trust by ensuring transparency and objectivity in AI outputs. It also highlights the necessity of developing a secure national AI infrastructure to protect against novel threats, advocating for a unified data platform to enhance defense capabilities. The plan suggests a preference for an open, interoperable approach to technology policy, which would prevent vendor lock-in and foster a competitive environment, thereby supporting national resilience. Elastic, a proponent of open-source principles, promotes its role in helping the government achieve this vision by providing a versatile data platform to maintain control over data and enhance both innovation and security.