Startup Platform Combines Endpoint Activity, Business Context, and Security Policies
By Michael Novinson
July 22, 2026
In a notable development in the realm of cybersecurity, a new endpoint security startup has emerged from stealth mode, securing an impressive $180 million in funding aimed at addressing the pressing security challenges posed by artificial intelligence (AI). This startup, named Glow, was co-founded by Roi Tiger, a former engineering leader at Meta, and is poised to make significant strides in the large and complex landscape of endpoint protection.
The funding round, led by prominent investors such as Sequoia, Cyberstarts, Greenoaks, and Redpoint Ventures, allows Glow to delve deeper into the intricacies of endpoint activities. The company aims to integrate an understanding of every application operating on an endpoint, along with the identity of the employee utilizing it, the organization’s security protocols, and the broader business context. Tiger emphasizes that this comprehensive understanding is pivotal in leveraging AI to stymie the infiltration of malicious software, risky AI agents, and unauthorized applications into business environments.
"This is really a huge challenge—one that has persisted for decades on the endpoint without a viable solution," Tiger remarked in an interview with ISMG. "Now, thanks to AI, we have the capability to scale and address issues that were previously inconceivable." With a valuation of $1.2 billion following its initial funding, Glow is positioned to play a critical role in the evolving narrative of endpoint security.
The urgency behind Glow’s mission is underscored by the escalating use of AI-driven applications within organizations. Employees, regardless of their technical expertise, are increasingly adopting AI-powered coding assistants and productivity tools that operate within enterprise systems. These applications often possess the ability to download software, install plug-ins, and access sensitive corporate resources, which in turn elevates the risk of potential cyber threats. As if this weren’t enough, the attackers themselves have access to sophisticated AI technologies that enable them to execute attacks with unprecedented speed and complexity.
"The endpoint has become the primary battleground for security," Tiger asserts. As AI technologies become more commonplace among employees across various functions, the risks associated with them multiply, creating a complicated new layer of vulnerability for organizations. Traditional security teams often find themselves relying on numerous disjointed point solutions and manual policy enforcement, which can be increasingly inadequate in the face of rapid AI-driven business activity and the inventive strategies employed by cybercriminals.
Tiger explains that companies are expected to harness AI to enhance productivity while simultaneously safeguarding against AI-enabled threats, all without significantly expanding their security teams. "AI truly unlocks the scalability needed to address these problems by providing visibility and control over endpoint activities, thereby preventing the attack surface from expanding indefinitely," he notes.
To achieve its ambitious goals, Glow utilizes advanced foundation models through AWS Bedrock, including notable models from Anthropic and Google’s Gemini. However, Tiger stresses that the models are merely a piece of the solution. The company’s engineering team concentrates on ensuring that AI-generated outputs are reliable enough for security operations. This requires an intricate blend of organizational data, security protocols, and business context, resulting in highly accurate decision-making capabilities.
"The true advantage of AI lies in its ability to scale our operations without hindering the pace of business," Tiger claims. By refining decision-making processes, Glow can support employees in their work while tackling security challenges more effectively.
Glow adopts a flexible approach to how organizations implement AI within their security frameworks. Companies can start by having humans review AI-generated recommendations before approving actions or can allow the AI to autonomously execute lower-risk security policies. This flexible deployment model enables organizations to gradually increase AI responsibilities as their confidence in the technology matures.
Tiger elaborates, "You can choose to maintain a fully human oversight approach as the AI learns the unique dynamics of your organization, or you can begin with specific policies that can operate autonomously, allowing the AI to make decisions."
The early applications of Glow’s technology include tasks such as continuously scanning software marketplaces, investigating suspicious applications, assessing potential threats, and intercepting risky software from entering enterprise environments. The proactive nature of these actions means they are often perceived as lower-risk opportunities for AI automation.
"The AI will generate an action plan based on the organization’s existing tools and capabilities, which will vary from one entity to another," Tiger states. Furthermore, organizations can dictate which specific actions necessitate additional human approval, particularly for changes that could impact production systems or other critical business operations.
Ultimately, Glow aims to establish a system where humans supervise AI decision-making while allowing for quick and consistent execution of approved tasks. Tiger emphasizes the importance of a dual-layered approach to cybersecurity: "We believe that a prevention-first strategy serves as the frontline defense for every organization, while detection and response mechanisms should function as a safety net for any risks that manage to infiltrate."
Through its innovative approach, Glow is not just responding to current cybersecurity challenges; it is engineering a paradigm shift in how organizations can harness AI for enhanced security. This shift could prove instrumental in building a more secure digital landscape in an era where AI continues to reshape the fabric of business operations.