JFrog LtdFROG

JFrog released agentic security capabilities, AI Catalog and JFrog AppTrust features in 2025.

“In 2025, we released multiple JFrog Platform features across our core functionalities in DevOps, DevSecOps, DevGovOps and Artificial Intelligence and Machine Learning operations (“AI/MLOps”), as we continued to expand our position as the system of record for the unified software release lifecycle. We released significant enhancements to our platform, including the release of agentic security capabilities, AI Catalog and JFrog AppTrust for DevGovOps and compliance. We also released new subscription bundles to align more closely with the emerging market needs around application security and AI development. These enhancements and releases represent continuing expansion within and beyond our core DevOps business, delving more deeply into DevSecOps, DevGovOps and AI/MLOps. We may not be able to sustain the pace of improvements to our products successfully or implement systems, processes, and controls in an efficient or timely manner or in a manner that does not negatively affect our results of operations. Our failure to improve our systems, processes, and controls, or their failure to operate in the intended manner, may result in our inability to manage the growth of our business and to forecast our revenue, expenses, and earnings accurately, or to prevent losses.”

JFrog cites uncertainty and costs of integrating agentic AI and other emerging technologies.

“Uncertainty regarding new and emerging AI technologies, such as agentic AI, generative AI, and advances in ML, may require us to incur costly additional expenses to research and integrate AI technologies into our future product offerings and our internal systems. The emergence of token-based or consumption-based pricing models for AI-powered software tools across our industry may increase, and introduce unpredictability into, the costs we incur to operate our business, which could adversely affect our operating results and ability to forecast expenses. Seasonal or project-driven spikes in development activity, onboarding of new engineering teams, adoption of additional platform features, and changes in vendor pricing tiers or overage rates could each independently or collectively cause our actual expenditures under these arrangements to exceed budgeted amounts, which could adversely affect our financial condition. Additionally, AI may create content that appears correct, but is factually inaccurate, biased, insufficient, poor quality, flawed, or contain other errors or inadequacies, any of which may not be easily detectable. AI and ML technologies have been known to produce false or hallucinatory inferences or outputs. Our use of AI technologies may expose us to additional claims, demands, and proceedings by private parties, customers, and regulatory authorities and subject us to legal liability as well as brand and reputational harm, confidentiality or security risks, compet”

JFrog warns of security risks from customer use of AI-powered coding agents and agentic AI workflows.

“Furthermore, the increasing adoption of AI-powered coding agents by our customers and the broader developer community has resulted in an acceleration in the volume of software binaries and code artifacts that our platform must manage and secure. AI-generated code has been shown to reintroduce known vulnerability classes, including injection, cross-site scripting, and SQL injection vulnerabilities, at substantially higher rates than human-authored code. If our security products fail to detect or adequately flag these vulnerabilities, our customers could experience security breaches or other incidents, which could result in reputational harm, liability claims, and loss of customers. These non-human actors may also introduce novel security risks, including unauthorized package pulls, unvetted dependency chains, and actions that circumvent human review. The security and governance frameworks for agentic AI workflows are still evolving, and there is limited regulatory clarity, industry standardization, or case law addressing liability for actions taken by autonomous software agents. If our platform fails to adequately govern these workflows, or if the regulatory environment evolves in a manner that imposes obligations we cannot efficiently satisfy, our business and reputation could be harmed.”

JFrog notes employee use of AI technologies and agents may increase security breach risks.

“Many of our and our service providers’ employees and other personnel work remotely, and are increasingly using AI technologies, all of which may increase our and their susceptibility to security breaches and incidents. We maintain certain policies and controls designed to address identified risks relating to our employees’ use of generative AI, but personnel may unknowingly or otherwise inappropriately enter proprietary or sensitive information, data, or code into AI tools, may use tools making use of AI agents or other types of AI inappropriately, or may use unauthorized AI tools, all of which may lead to loss of intellectual property protection, inappropriate disclosure or other processing of proprietary or sensitive information, data, or code, noncompliance with applicable laws, regulations, or other obligations, and other liabilities. As we continue to expand our offerings, including through the acquisition of complementary businesses, such as our acquisition of Vdoo in 2021 and our acquisition of Qwak in July 2024, our products will likely have access to more sensitive and personal information of our customers, which could result in greater adverse effects from security breaches and other security incidents. Further, the more widespread our platform and products become, the more they may be viewed by malicious actors as attractive targets for attacks.”

Forward-looking statements on adoption of AI and use of AI agents.

“This press release and the earnings call referencing this press release contain “forward-looking” statements, as that term is defined under the U.S. federal securities laws, including but not limited to statements regarding JFrog’s future financial performance, including our outlook for the third quarter and for the full year of 2026, expectations regarding the market and revenue potential for the JFrog Platform, including JFrog Artifactory, JFrog Xray, JFrog Curation, JFrog Advanced Security, JFrog ML, JFrog AppTrust, JFrog AI Catalog and JFrog Runtime Security, and including the efficacy and benefit of integrating of any of the foregoing with other products and platform, our expectations regarding the mission-critical nature of the “JFrog Platform” to our customers’ infrastructure and its growth potential, expectations regarding the adoption of AI and the use of AI agents, the growth potential of our cloud business, including hybrid and multi-cloud, our expectations regarding potential for growth in and market opportunities within DevOps, DevSecOps, DevGovOps, Security, AI, and MLOps, our ability to provide effective tools and solutions to detect and remediate security vulnerabilities, our expectations regarding our strategic integrations and collaborations, the ability of our strategic sales team to grow the business across top-tier accounts, our ability to expand usage of our platform in the government and commercial sectors, our ability to contribute data to global secur”

JFrog released agentic security capabilities, AI Catalog and JFrog AppTrust for DevGovOps and AI/MLOps.

“In 2025, we released multiple JFrog Platform features across our core functionalities in DevOps, DevSecOps, DevGovOps and Artificial Intelligence and Machine Learning operations (“AI/MLOps”), as we continued to expand our position as the system of record for the unified software release lifecycle. We released significant enhancements to our platform, including the release of agentic security capabilities, AI Catalog and JFrog AppTrust for DevGovOps and compliance. We also released new subscription bundles to align more closely with the emerging market needs around application security and AI development. These enhancements and releases represent continuing expansion within and beyond our core DevOps business, delving more deeply into DevSecOps, DevGovOps and AI/MLOps. We may not be able to sustain the pace of improvements to our products successfully or implement systems, processes, and controls in an efficient or timely manner or in a manner that does not negatively affect our results of operations. Our failure to improve our systems, processes, and controls, or their failure to operate in the intended manner, may result in our inability to manage the growth of our business and to forecast our revenue, expenses, and earnings accurately, or to prevent losses.”

Uncertainty around agentic AI and generative AI may require costly research, integration and expose JFrog to legal and reputational risks.

“Uncertainty regarding new and emerging AI technologies, such as agentic AI, generative AI, and advances in ML, may require us to incur costly additional expenses to research and integrate AI technologies into our future product offerings and our internal systems. Additionally, AI may create content that appears correct, but is factually inaccurate, biased, insufficient, poor quality, flawed, or contain other errors or inadequacies, any of which may not be easily detectable. AI and ML technologies have been known to produce false or hallucinatory inferences or outputs. Our use of AI technologies may expose us to additional claims, demands, and proceedings by private parties, customers, and regulatory authorities and subject us to legal liability as well as brand and reputational harm, confidentiality or security risks, competitive harm, ethical and social concerns, or other complications that could adversely affect our business, reputation, or financial results. If we do not have sufficient rights to use the output of such AI and ML tools, or other data or content on which the AI and ML tools we use rely, we also may incur liability by violation of applicable laws and regulations, third-party intellectual property or other rights, or contracts to which we are a party.”

Employee use of AI agents and unauthorized AI tools may lead to loss of IP, data disclosure and other liabilities for JFrog.

“Many of our and our service providers’ employees and other personnel work remotely, and are increasingly using AI technologies, all of which may increase our and their susceptibility to security breaches and incidents. We maintain certain policies and controls designed to address identified risks relating to our employees’ use of generative AI, but personnel may unknowingly or otherwise inappropriately enter proprietary or sensitive information, data, or code into AI tools, may use tools making use of AI agents or other types of AI inappropriately, or may use unauthorized AI tools, all of which may lead to loss of intellectual property protection, inappropriate disclosure or other processing of proprietary or sensitive information, data, or code, noncompliance with applicable laws, regulations, or other obligations, and other liabilities. As we continue to expand our offerings, including through the acquisition of complementary businesses, such as our acquisition of Vdoo in 2021 and our acquisition of Qwak in July 2024, our products will likely have access to more sensitive and personal information of our customers, which could result in greater adverse effects from security breaches and other security incidents. Further, the more widespread our platform and products become, the more they may be viewed by malicious actors as attractive targets for attacks.”

JFrog launched Skills Registry with NVIDIA for trust layer on AI agent skills.

“Launched JFrog Skills Registry alongside NVIDIA to bring a trust layer to AI agent skills.”

Forward-looking statements on AI agents, platform efficacy, and market opportunities.

“This press release and the earnings call referencing this press release contain “forward-looking” statements, as that term is defined under the U.S. federal securities laws, including but not limited to statements regarding JFrog’s future financial performance, including our outlook for the second quarter and for the full year of 2026, expectations regarding the market and revenue potential for the JFrog Platform, including JFrog Artifactory, JFrog Xray, JFrog Curation, JFrog Advanced Security, JFrog ML, JFrog AppTrust, JFrog AI Catalog and JFrog Runtime Security, and including the efficacy and benefit of integrating of any of the foregoing with other products and platform, our expectations regarding the mission-critical nature of the “JFrog Platform” to our customers’ infrastructure and its growth potential, expectations regarding the adoption of AI and the use of AI agents, the growth potential of our cloud business, including hybrid and multi-cloud, our expectations regarding potential for growth in and market opportunities within DevOps, DevSecOps, DevGovOps, Security, AI, and MLOps, our ability to provide effective tools and solutions to detect and remediate security vulnerabilities, our expectations regarding our strategic integrations and collaborations, the ability of our strategic sales team to grow the business across top-tier accounts, our ability to expand usage of our platform in the government and commercial sectors, our ability to contribute data to global secu”

JFrog platform supports governance for hybrid teams of developers and AI agents.

“JFrog’s goal is to provide a system of record for the software supply chain in the Artificial Intelligence (“AI”) era. As a foundational platform that provides governance, security and trust to enterprise software organizations, JFrog’s unified platform (“JFrog Platform”) allows companies to successfully unify development speed, volume, security, governance and delivery of software applications across hybrid teams of developers, security professionals, Artificial Intelligence/Machine Learning (“AI/ML”) engineers and AI agents. We envision a world where innovative, secure, compliant and trusted software is consumed, built and continuously delivered from any user to any destination; a reality we refer to as “Liquid Software.””

Proliferation of open source, AI, ML models and Agentic AI increases software supply chain complexity and strains traditional workflows.

“The proliferation of open source software, AI and ML models and availability of newer and more efficient software development technologies such as Agentic AI and Generative AI, enable organizations to produce software at an increasing rate. However, with speed can come software supply chain complexity through disparate services, inadvertent adoption of malicious packages, introduced security vulnerabilities, unregulated containers, and unchecked deployment across hybrid and multi-cloud environments. Organizations’ existing approaches can create silos and bottlenecks around critical steps, such as planning, curating, building, testing, securing, and delivering intelligent software. The merging of new technologies with legacy approaches has placed significant strain on traditional software workflows, which are even further complicated by the introduction of these AI components.”

DevOps is increasingly influenced by AI such as generative coding assistants and AI agents.

“DevOps is a maturing discipline that integrates software development and operations, shortening, automating, and improving the software build and release workflow. DevOps is a combination of technologies, methodologies, and culture that powers a continuous, fast, and secure software release cycle. DevOps is increasingly influenced by AI, such as generative coding assistants and AI agents.”

Technologies include AI agent-driven remediation of security vulnerabilities and contextualized security analysis.

“technologies, such as those that provide AI agent-driven remediation of security vulnerabilities or agents that provide contextualized security analysis.”

AI and MLOps processes combine developers, AI agents, data scientists and ML teams to build and deploy models.

“Compounding the DevOps and DevSecOps workflows, increasing AI and MLOps processes combine the work of developers, AI agents, data scientists, and ML engineering teams who build, train, and deploy AI technologies - such as ML models - with operational motions that bring these models and AI-empowering technologies into production as part of applications.”

Software increasingly uses code from developers or generative AI with most binaries from open source.

“Modern society heavily relies on software, and most individuals understand that software comes from code generated by software developers and, increasingly, AI agents and coding assistants. But software, in code form, doesn’t equate to an application. Code is built (i.e., transformed) into binary files (or “software packages” and “artifacts”) that allow it to run alongside other components as a complete application for consumers as it runs on a server or device. The vast majority of technologies today are built using only a small percentage of original source code written by developers or generative AI, with an estimated >80% of an application’s binaries (including containers, traditional software packages and AI/ML models) coming from open source, existing software. JFrog’s unified platform is designed to universally secure, manage and deploy all types of software artifacts within an organization, making it the system of record for an organization’s software, whether bespoke by hybrid development teams or consumed from a third party.”

JFrog CEO discusses AI coding agents and platform for software supply chain management.

““Developers and AI coding agents are now building and releasing software together at unprecedented speed. Automation, security, and governance must be embedded from the get-go,” said Shlomi Ben Haim, CEO and Co-founder of JFrog. “Our 2025 performance reflects strong execution of our strategy to serve as the System of Record for all software artifacts, and it demonstrates how deeply companies are embracing the JFrog Platform as they adopt AI to manage and secure their software supply chains. We enter 2026 with strong momentum, positioned to power the next era of software delivery, jointly driven by AI agents and human developers.””

JFrog cites risks from uncertainty and use of agentic AI and generative AI in products and systems.

“Uncertainty regarding new and emerging AI technologies, such as agentic AI, generative AI, and advances in machine learning, may require us to incur costly additional expenses to research and integrate AI technologies into our future product offerings and our internal systems. Additionally, AI may create content that appears correct, but is factually inaccurate, biased, insufficient, poor quality, flawed, or contain other errors or inadequacies, any of which may not be easily detectable. AI and machine learning technologies have been known to produce false or hallucinatory inferences or outputs. Our use of AI technologies may expose us to additional claims, demands, and proceedings by private parties, customers, and regulatory authorities and subject us to legal liability as well as brand and reputational harm, confidentiality or security risks, competitive harm, ethical and social concerns, or other complications that could adversely affect our business, reputation, or financial results. If we do not have sufficient rights to use the output of such AI and machine learning tools, or other data or content on which the AI and machine learning tools we use rely, we also may incur liability by violation of applicable laws and regulations, third-party intellectual property or other rights, or contracts to which we are a party.”

JFrog announced AI agent-driven security remediation and the first agentic repository.

“Announced AI agent-driven security remediation & JFrog Fly, the first agentic repository.”

Uncertainty around AI technologies may require costly research, integration, and expose the company to legal, reputational, and security risks.

“Uncertainty regarding new and emerging AI technologies, such as agentic AI, generative AI, and advances in machine learning, may require us to incur costly additional expenses to research and integrate AI technologies into our future product offerings and our internal systems. Additionally, AI may create content that appears correct, but is factually inaccurate, biased, insufficient, poor quality, flawed, or contain other errors or inadequacies, any of which may not be easily detectable. AI and machine learning technologies have been known to produce false or hallucinatory inferences or outputs. Our use of AI technologies may expose us to additional claims, demands, and proceedings by private parties, customers, and regulatory authorities and subject us to legal liability as well as brand and reputational harm, confidentiality or security risks, competitive harm, ethical and social concerns, or other complications that could adversely affect our business, reputation, or financial results. If we do not have sufficient rights to use the output of such AI and machine learning tools, or other data or content on which the AI and machine learning tools we”

JFrog released remote MCP server to enable agentic AI interaction with its platform.

“Announced release of remote MCP server to enable agentic AI interaction with the JFrog Platform.”