SentinelOne acquires Prompt Security: Agentic AI protection takes centre stage

Hot on the heels of Palo Alto Networks’ announcement that it plans to buy CyberArk, SentinelOne (aka S1) has flexed its own M&A muscles to snap up Prompt Security, the Israeli AI security startup. S1 says that the acquisition will advance its GenAI security and agent security strategy, providing the foundation for secure AI adoption at scale.

Details of the acquisition

Before too many comparisons are made, it’s important to put the two purchases in perspective. While Palo Alto Network’s acquisition of CyberArk is a huge, $25 billion deal – in the same order of magnitude as Google’s $32 billion purchase of Wiz, for example – S1 is buying Prompt Security for relative pocket change. Although undisclosed, the value of the stock and cash transaction is likely to be between $250-$300 million, according to Israeli business news outlet Globes. This would represent a healthy multiplier vs. Prompt’s value. Since Prompt Security’s founding in 2023, the firm has raised around $23 million in capital. S1 is clearly expecting to see the return on such an investment.

Why Prompt Security?

The reason why S1 may consider this to be a good deal is what S1 is acquiring: the IP that Prompt Security has developed, and that is increasingly pertinent for firms looking to scale their cybersecurity defences as GenAI and AI agents are increasingly deployed in the enterprise. S1 is leaning heavily on the case for protecting against the risks associated with AI. This will be music to the ears of countless security practitioners who have been flagging the cyber threats associated with “shadow AI”. With AI-related data leakage, misuse, ungoverned access and LLM-specific risks such as prompt injection attacks on the increase, Prompt Security’s capabilities will be a significant enhancement to S1’s portfolio and Singularity Platform.

Describing itself as “The Platform for AI Security”, Prompt Security can boast strong credentials, especially via its founders, who have over 15 years’ experience in the AI security space. The company is a member of the OWASP research team and has an array of enterprise customers – including The New York Times, Royal Caribbean, Amdocs and others. It also provides specific solutions for the healthcare and financial services industries.

How will this acquisition benefit S1’s profile and market position?

Publicly owned S1 has a market cap of $5.5-$6 billion as of mid-2025. The firm has faced significant pressure on its stock: in mid-July it was trading down 21% on the year and is down around 60% over the past 4 years since its IPO in June 2021. The acquisition announcement might have been expected to change that, but between market open on 4th August and market close on 8th August S1’s stock dropped around 11%. During the same period, several of S1’s publicly-traded competitors – including CrowdStrike, Palo Alto Networks and Rapid7 – saw share price drops, but only between 4.2%-7.4%. S1’s announcement of the acquisition of PingSafe in 2024 resulted in a 5.1% share price drop on the day, while the news of its planned purchase of Attivo Networks in March 2022 saw a boost to its share value for around two months before dropping back below $30 dollars in May 2022.

It is still too early to tell, but does this purchase change any of the fundamentals underlying S1’s business? Pessimists would say probably not. Previous acquisitions have not obviously boosted its share price or increased the perception of S1’s value directly over the long term. There is also the question of why Prompt Security? When it comes to LLM and GenAI governance, there are firms such as Securiti, Prisma, Blueteam AI, Unbound Security, Lasso and others. Prompt Security appears to be staking a claim to being the leader in Agentic AI threat protection at the Model Context Protocol (MCP) level – its MCP Gateway is designed to monitor, control, and protect MCP interactions “in real time” i.e. as they are happening between AI agents and MCP hosts and servers, as opposed to identity and/or access management aspects when the AI agents interact with other applications, data sources or tools. Nevertheless, startups including Straiker, Splx.ai, Noma and others offer similar forms of governance capability and are listed by OWASP in their Agentic AI Security Landscape – Q2/3 2025 “Cheat Sheet” report.

At the same time, Prompt Security represents a potentially smaller investment than other options and may just be a better ‘fit’ for S1. Its small size means it will likely be quicker and easier to integrate, and its focus has been squarely on AI security from day one. Perhaps the best indication has come from Itamar Golan, Co-founder and CEO of Prompt Security, confirming the potential synergy with S1 in his blog. He describes the discussions around joining forces with S1, enthusing: “It was clear from our first conversation: this is home”

The Gist

The purchase of Prompt Security is important for S1’s platform-led approach to integrating functions across endpoint, networking, cloud and identity functions. GenAI tools and AI agents are already being used by many enterprises, and the expectation is that their adoption will rapidly increase. While the security risks are still emerging, they need to be addressed by specialist solutions such as those offered by Prompt Security and others that drill-down to the level of MCP interactions. The ability to discover, identify, and act against both ungoverned internal use as well as external threats is key to providing the reassurance enterprises need to adopt GenAI and agentic AI technologies.