SentinelOne’s $250M Bet on Prompt Security: Teh Rise of AI Security and the Fight Against Shadow AI
Table of Contents
Meta Description: Discover how Prompt Security, acquired by SentinelOne for an estimated $250 million, is leading the charge in securing the rapidly expanding world of generative AI and combating the risks of shadow AI.
The generative AI (GenAI) revolution is here, but so are the security risks. In August 2025, SentinelOne finalized the acquisition of Prompt Security for an estimated $250 million, a move signaling the critical importance of securing the burgeoning AI landscape. Prompt Security’s rapid ascent – from founding in August 2023 to a successful exit in under two years, fueled by $23 million in funding and a 50-person team – underscores the urgency organizations face in addressing the unique threats posed by large language models (LLMs).
The genesis of Prompt Security: From Transformer Architectures to a New Attack Surface
Itamar Golan,co-founder and CEO of Prompt Security,didn’t stumble into AI security; he built his foundation in it. His academic work focused on transformer architectures – the very technology powering today’s LLMs – long before their widespread adoption. Golan’s early experience building security features leveraging GPT-2 and GPT-3 revealed a essential truth: these powerful models created an entirely new attack surface.
“that experience made it clear that applications powered by GPT-like models were opening an entirely new and vulnerable attack surface,” Golan explained. This realization spurred the creation of Prompt Security, with a mission to empower organizations to safely harness the potential of AI.
Shadow AI: A $4.63 Million Per Breach Problem
The timing couldn’t be more critical. Shadow AI – the unauthorized use of AI tools within organizations – is exploding. VentureBeat analysis reveals that shadow AI breaches now cost enterprises an average of $4.63 million, 16% above the average breach cost. Alarmingly, 97% of breached organizations lack basic AI access controls, according to IBM’s 2025 data. Estimates suggest shadow AI apps could double by mid-2026, growing at a rate of 5% per month. Cyberhaven data paints a stark picture: 73.8% of ChatGPT workplace accounts are unauthorized, and overall enterprise AI usage has increased 61x in just 24 months.
As Golan previously noted, “We see 50 new AI apps a day, and we’ve already cataloged over 12,000. Around 40% of these default to training on any data you feed them, meaning your intellectual property can become part of their models.”
Beyond Prompt Injection: A Thorough AI Security Platform
Prompt Security quickly became known for its expertise in prompt injection defense, but the platform’s scope extended far beyond that. It was designed to address a broad range of GenAI security challenges, including data leakage, model governance, compliance, and red teaming. The platform provided runtime protection across the association, safeguarding every interaction with AI.
A key differentiator was its ability to discover the extent of shadow AI within organizations. Many customers were surprised to find dozens of unmanaged AI services in use outside of IT’s visibility. Equally impo
Now integrated into SentinelOne’s Singularity Platform, Prompt Security’s focus is on extending AI security across the entire ecosystem, encompassing endpoints, identities, and cloud workloads. The ultimate goal, according to Golan, is to build a future “where AI itself becomes part of the defense fabric: not just something to secure, but something that secures you.”
The acquisition of Prompt Security, alongside similar moves by Palo Alto Networks, Tenable, and Cisco, highlights a growing trend: the companies that will thrive in the age of AI-enabled attacks are those that prioritize security from the outset. Post-acquisition, Prompt Security’s capabilities will extend across SentinelOne’s Singularity Platform, including MCP gateway security between AI applications and more than 13,000 known MCP servers, and model-agnostic coverage across major LLM providers like OpenAI, Anthropic, and Google.
