何が起こったのか
Snyk announced that Evo, its agent-native security layer, now accounts for 60% of the company's new deal volume and is driving a 30% increase in average contract value. The company reported an 81.5% month-over-month customer growth rate since the general availability of its first capability in March 2026. Snyk stated that 76% of customers who purchased Evo in the second quarter deployed it in production before the quarter closed, significantly faster than the typical two-to-three-quarter rollout for enterprise security software. The company processes 2.4 million agent supply-chain scans monthly and 4.2 million agent behavior checks daily across more than 417,000 onboarded machines. Snyk cited a deployment at one of the largest US banks, where Evo is used to centralize risk assessment for approximately 1,000 internal agent skills used by 50,000 developers.
Snyk, an AI security company, announced that Evo, the agent-native layer of its AI Security platform, now represents 60% of its new deal volume. This product is driving a 30% increase in the company's average contract value. Since the general availability of its first capability in March 2026, Evo has maintained an 81.5% month-over-month customer growth rate.
The company reported that 76% of customers who purchased Evo in the second quarter of 2026 deployed it in production before the quarter ended. This is notably faster than the typical two-to-three-quarter deployment timeline for enterprise security software. Snyk attributes this speed to Evo's ability to install into existing agent workflows without requiring extensive change-management programs.
Snyk stated that it currently processes 2.4 million agent supply-chain scans per month and 4.2 million agent behavior checks per day across more than 417,000 onboarded machines. The company noted that newly introduced security issues across its 4,800-plus customers rose 108% between Q4 2024 and Q1 2026, a trend exacerbated by the rise of agentic code generation.
A specific deployment was highlighted at one of the largest banks in the United States, where Snyk runs pre-registry risk assessment and continuous scanning for approximately 1,000 internal agent skills used by 50,000 developers building with Claude Code. This deployment has been live since July 2026 and was built jointly with the bank's security team.
ソースの詳細: cyprusshippingnews.com ↗
なぜそれが重要なのか
The rapid adoption of Evo indicates a significant shift in enterprise security strategy toward governing AI agents and agentic workflows. As coding agents generate more code and vulnerabilities than human teams can review, traditional security tools that scan static artifacts are insufficient. Snyk's data suggests that enterprises are moving from experimental AI security to production-grade controls that validate agent behavior and supply chains in real-time. This trend highlights the growing market demand for independent validation layers that can manage the security risks introduced by autonomous AI systems in software development and operations.
The adoption of Evo reflects a broader industry recognition that traditional security tools, which scan static artifacts, are inadequate for securing AI agents that assemble toolchains at runtime and execute production actions. Snyk argues that the 'generator cannot be the validator,' necessitating independent, deterministic security layers beneath AI models.
Snyk's VulnBench research, cited in the announcement, found that LLMs alone flagged issues inconsistently, with nearly half of flagged issues appearing in only one of five identical runs. In contrast, Snyk claims that open-source issues remediated through its independent validation layer merge at a 94% higher rate than fixes produced by a frontier model working alone.
The rapid deployment timeline suggests that enterprises are treating security as a critical, immediate operational risk rather than a long-term strategic initiative. This shift is driven by the compounding nature of vulnerabilities introduced by coding agents and the ability of attackers to exploit weaknesses at machine speed.
インタラクティブなメカニズム: 実際にどのように機能するか
この開発の背後にある基盤となるテクノロジーをインタラクティブに探索します。
crm_get_transaction(id='4092').What most distinguishes an AI agent from a basic chatbot?
次に見るべきもの
Monitor the expansion of Snyk's Evo capabilities, particularly the transition of the Remediation Agent from open preview to general availability. Watch for further disclosures on how other large enterprises are integrating agent-native security controls into their development pipelines. Observe whether the reported 94% higher merge rate for independently validated fixes becomes a standard for AI-generated code security in the industry.
The transition of Evo's Remediation Agent from open preview to general availability will be a key indicator of the maturity of autonomous security fixes. Currently, the Agentic AppSec Agent is in private preview, orchestrating triage and remediation loops.
Further details on the specific security controls implemented in the US bank deployment may provide insights into how regulated industries are governing behavior and supply chains.
The industry response to Snyk's claim that independent validation is superior to model self-grading will likely influence the development of new security standards and tools for agentic AI.