Tag: AI 보안

  • Armadin Raises $255.5M — Mandia’s ‘Agent Swarm’ Bet Returns After 4 Years

    Armadin Raises $255.5M — Mandia’s ‘Agent Swarm’ Bet Returns After 4 Years

    Key Takeaways

    • Armadin is a new startup founded by Kevin Mandia, the founder of Mandiant, the security firm Google acquired in 2022 for $5.4 billion.
    • On October 1, Armadin announced it had closed a $255.5M Series B at a $2.5B valuation.
    • The Series B was co-led by Andreessen Horowitz and Accel, with participation from Bain Capital Ventures, Redpoint, 8VC, Ballistic Ventures, Google Ventures, In-Q-Tel, Kleiner Perkins, and Menlo Ventures.

    As AI agents begin to be weaponized on both offensive and defensive sides, Armadin’s strategy of converting penetration testing into an ‘always-on agentic swarm’ to eliminate the latency gap in enterprise defense—and the resulting shift in cybersecurity capital flows—deserve close examination.

    Table of Contents

    Four years after selling Mandiant to Google for $5.4 billion in 2022, Kevin Mandia’s new security startup Armadin announced on October 1 that it had closed a Series B. The Armadin funding round totals $255.5M at a $2.5B valuation. Including the follow-on to the $190M Series A from six months ago, cumulative capital raised now exceeds $445M.

    The pace of this Armadin funding round is itself the message. In the cybersecurity market, a ‘promising’ company typically takes 18–24 months to reach Series B—but Armadin closed its second major round in just six months. A look at the investor lineup in this Armadin funding round reveals why.

    Armadin Funding — Round Details and Investor Lineup

    The Series B was co-led by Andreessen Horowitz (a16z) and Accel. The participant lineup is striking: Bain Capital Ventures, Redpoint, 8VC, Ballistic Ventures, Google Ventures (GV), In-Q-Tel, Kleiner Perkins, and Menlo Ventures—alongside Silicon Valley’s traditional VCs, the venture arm of the U.S. intelligence community (In-Q-Tel) and Google’s venture fund sat at the same table.

    Category Investors
    Lead Andreessen Horowitz, Accel
    Private VCs Bain Capital Ventures, Redpoint, 8VC, Ballistic Ventures, Kleiner Perkins, Menlo Ventures
    Strategic & Government Capital Google Ventures (GV), In-Q-Tel

    The most meaningful signal in this lineup, in my view, is the simultaneous entry of GV and In-Q-Tel. Google was the party that acquired Mandia and his Mandiant four years ago, and In-Q-Tel is the informal fund the U.S. intelligence community uses to procure technology. The fact that both pools of capital came in together means Armadin’s ‘always-on agentic swarm’ product concept has been recognized not as flashy marketing, but as security infrastructure that works for both the U.S. government and Big Tech.

    Agentic Swarm — From One-Off Pen Tests to Always-On Automation

    Traditional penetration testing typically happens once or twice a year, with outside firms probing for vulnerabilities over a defined window. Armadin’s approach is different. Instead of humans manually scripting scenarios, multiple AI agents run continuously, automatically chaining vulnerabilities together to attempt deeper system penetration. In the company’s own words, the goal is to operate under the premise that an attacker could reach ‘an AI lab with logging agents’—and to help companies block the same attack paths first, before a real hack occurs.

    Put simply, if traditional pen testing is an ‘annual physical exam,’ Armadin’s agent swarm is a ‘cardiothoracic surgeon on call 24/7.’

    That said, what stands out to practitioners is the volume of logs and the question of accountability created by the ‘always-on’ model introduced with the Armadin funding. When agents continuously probe a system, that probing itself becomes load—and a malfunction could cause normal traffic to be misidentified as an attack, bringing operations to a halt.

    The Timing of Armadin’s Raise — Why Now

    It’s no coincidence that the Armadin funding is happening at this moment. On the same day, reports surfaced that xAI’s Grok model is being used in actual U.S. Department of Defense and White House decision-making. AI agents are starting to exert influence not only on defense but in policy and military domains. That means agent swarms will inevitably connect to the broader question of agent governance, well beyond private-sector penetration testing. The stronger the autonomy, the more unavoidable the question: how far should agents be allowed to decide on their own?

    The Dilemma — When the Defender Becomes the Attacker

    The most uncomfortable question the Armadin funding poses to the industry is this. An ‘always-on agentic swarm’ is, by its very nature, a system that continuously runs attack scenarios inside a company. If logs persist after the contract ends, or if an agent unintentionally scans external systems, the swarm itself becomes an ‘insider attacker.’ As the Grok case demonstrates, the more powerful agents become, the more likely it is that controlling the misuse of defensive agents will become cybersecurity’s new frontier.

    Closing — It’s Time to Rewrite Pen Testing

    The Armadin funding is more than a fundraising headline. It signals a shift in the basic unit of cybersecurity work—from ‘one-off scenarios scripted by humans’ to ‘self-evolving collectives of agents.’ The era Mandia defined at Mandiant, of ‘post-incident analysis,’ is fading; an era of ‘continuous simulated intrusion before an incident occurs’ is opening. For security leaders, the question is no longer ‘should we buy an agent swarm?’ but ‘who will control how the agent swarm behaves inside our systems?’

    What to Do Right Now

    • Pull out your current penetration testing contract and mark whether it falls under ‘one to two annual one-off assessments’ or ‘continuous automation.’
    • From your internal asset inventory, separately classify any ‘log nodes with permanent agent access’ and review access-control policies on each.
    • Add an explicit line item to vendor RFPs asking about the retention period of agent-generated logs and whether they are transmitted externally.
    • Check whether your security team has a playbook for ‘agent malfunction response,’ and if not, draft an initial version within two weeks.
    • Inform the executive team in the next security report that ‘AI agents will reside in our systems on a permanent basis for defensive purposes.’

    Key Issues Summarized

    • Autonomy vs. Controllability: The more ‘always-on’ an agent swarm becomes, the more essential it is to have control mechanisms capable of tracing accountability when incidents occur.
    • Convergence of Private and Government Capital: The simultaneous entry of GV and In-Q-Tel implies expansion into the national-security domain. Ordinary enterprises need to review in advance where they will stand in this flow.
    • Potential for Misuse: There are still no industry standards governing the possibility that an agent designed for defense could scan outward or leak data.

    Frequently Asked Questions

    What is Armadin?

    It is a cybersecurity startup newly founded by Kevin Mandia, the founder of Mandiant, the security firm Google acquired in 2022 for $5.4 billion. Its core product is an ‘agentic swarm’ in which multiple AI agents run continuously to perform penetration testing automatically.

    Why are Armadin’s funding rounds coming in so quickly back-to-back?

    The gap between the $190M Series A (March) and the $255.5M Series B (October) is just six months. The fact that a16z, Accel, In-Q-Tel, and GV have gathered around the same capital table is read as a signal that the product is in demand from both the market and the government.

    How is Armadin’s agent swarm different from traditional penetration testing?

    Traditional pen testing relies on humans scripting scenarios and is performed one to two times a year over a defined window. Armadin’s agent swarm has multiple AI agents running continuously, automatically chaining vulnerabilities together to attempt attacks.

    Isn’t an agent swarm dangerous?

    Autonomous agents that run continuously could misidentify normal traffic as an attack when they malfunction, or leave residual logs after a contract ends and become an internal attacker. This is an area where control and governance standards are not yet established.

    Reference

    This article was prepared with reference to the following original source: TechCrunch — Kevin Mandia’s new ‘agent swarm’ security startup Armadin raises $255.5M at $2.5B valuation

  • Big Three Unveil AI Security Models in Coordinated Push: Inside the Google, Anthropic, and OpenAI Cyber Race

    Big Three Unveil AI Security Models in Coordinated Push: Inside the Google, Anthropic, and OpenAI Cyber Race

    Key Summary

    • Google unveiled its cybersecurity-focused AI security model, ‘Gemini 3.8 Flash Cyber,’ and self-evaluated it as the “most capable cybersecurity model to date.”
    • Google provides early access to the model to critical defense organizations in government, healthcare, and telecommunications through the ‘Fairwind Program.’
    • Based on the article headline, Anthropic and OpenAI also appear to have released cybersecurity AI models along with safeguards and access programs, but the main body of reporting only covers Google-related content.

    Analyzing how three big tech companies are applying AI in earnest to the cybersecurity field, and the policy and practical implications of restricted, trusted defender-led early access programs

    Table of Contents

    The race to develop AI security models has accelerated all at once. In September 2026, Google unveiled its cybersecurity-specialized AI security model, ‘Gemini 3.8 Flash Cyber,’ self-praising it as “the most capable cybersecurity model ever.” Reports surfaced that Anthropic and OpenAI also released their respective AI security models and early access programs in the same breath. The fact that all three big tech companies moved on the same topic simultaneously reads not as a simple product release, but as a signal of a turning point in the trend.

    Google Gemini 3.8 Flash Cyber: Where the AI Security Model Focuses

    This AI security model, called the Gemini Cyber line, narrowed its functionality to defender-centered tasks such as threat detection, breach analysis, and code security review. Rather than grafting a general-purpose generative model onto security as-is, the decisive difference from existing product lines is that this is a specialized model tailored to the workflow. The basis for Google’s use of the phrase “capable cybersecurity model” lies in its internal benchmark scores, but unless the evaluation criteria and datasets are made public, it’s difficult for readers to accept it at face value. This was the most disappointing point for me as well. The fact that benchmarks do not equal real-world performance has been repeatedly confirmed across the industry.

    Fairwind Program: The Controlled Deployment Approach of AI Security Models

    The Fairwind Program, which Google unveiled alongside the model, is a channel that provides early access to the AI security model for critical defense organizations. Organizations directly connected to national critical infrastructure—government, healthcare, telecommunications, and the like—are the priority targets, and a select group of “trusted defenders” uses the model in a controlled environment. The key point is that the program is not immediately open to general companies or individual developers. Google has chosen controlled access under the framework of responsible deployment, but the dual nature of this choice becomes the next point of contention.

    Anthropic and OpenAI Response: Limits of Reading the Headline Alone

    Based on the article headline, Anthropic and OpenAI also appear to have simultaneously released cybersecurity AI models and safeguards. However, since the main body of reporting remained at the level of an RSS summary, the specific model names, access program structure, and target organization scope of the two companies have not been confirmed. The fact that all three companies promoted AI security models at the same time is itself evidence that industry standards are forming quickly.

    Three-Company Comparison: AI Security Model Release Status at a Glance

    Item Google Anthropic OpenAI
    Model Name Gemini 3.8 Flash Cyber Unconfirmed Unconfirmed
    Early Access Program Fairwind Program Unconfirmed Unconfirmed
    Target Organizations Government, Healthcare, Telecommunications Unconfirmed Unconfirmed
    Release Scope Controlled Early Access Unconfirmed Unconfirmed
    Emphasis Selection of Trusted Defenders Estimated Safeguard-Centric Estimated Safeguard-Centric

    Issue Analysis: The Dual Edge of Responsible AI Security Model Deployment

    Controlled early access is a double-edged sword. While the advantage of preemptively reducing the possibility of model misuse is clear, it simultaneously concentrates more technological superiority in the hands of the group with access. The argument that security tools themselves can amplify asymmetry has been raised consistently in academia and policy circles. Depending on who defines the criteria for a “trusted defender,” the market may shrink, or game rules favorable to a specific group may become entrenched.

    From a practitioner’s perspective, the notable question is whether this system advances the democratization of security. If the baseline set by big tech becomes the default for global security practices, it becomes a standard in its own right. More details on the trend can be found in the original The Hacker News article.

    Practical Application Points

    • Within one week, review the scope of work that AI security models can replace in your organization’s security operations.
    • Determine in advance whether your organization qualifies as a target institution for the Fairwind Program and check the application eligibility requirements.
    • Track when competing models from the three companies release their benchmarks and build a comparative evaluation plan.
    • Check whether your team has internal guidelines for generative AI use, and draft one if it doesn’t exist.

    Actions You Can Take Right Now

    • Add “Review of AI security model adoption” as a one-line agenda item to today’s security operations meeting.
    • Update the contact information between Google Cloud Console and your security team point of contact.
    • Bookmark the official Fairwind Program guidance page and turn on notification alerts.
    • Check Anthropic and OpenAI official channels once a week for follow-up cybersecurity model announcements.
    • Add a one-page section on “Cases of generative AI used in attacks” to your internal security training materials.

    Frequently Asked Questions

    Can general companies use Gemini 3.8 Flash Cyber right away?

    A general public release date has not yet been set. The model is structured to be provided first to critical defense organizations such as government, healthcare, and telecommunications through the Fairwind Program, and general companies must wait for follow-up announcements.

    What are the eligibility requirements for the Fairwind Program?

    According to Google, organizations with critical infrastructure defense missions are the priority target. Specific eligibility requirements and procedures should be confirmed through official channels, and a CISO-level point of contact is recommended.

    What AI security models did Anthropic and OpenAI release?

    Based on the article headline, both companies appear to have released cybersecurity models, but due to limitations in the main body of reporting, the model names and access program details have not been confirmed. There is a need to watch for follow-up announcements through official channels.

    Will the introduction of cybersecurity AI reduce the demand for security personnel?

    While simple repetitive tasks may decrease, the demand for personnel responsible for model output verification and governance is likely to increase. Reorganizing the operating system, rather than just introducing tools, is the key task.

    Reference Original

    This article was written with reference to the following original: The Hacker News — Google, Anthropic, and OpenAI Unveil Cyber AI Models, Safeguards, and Access Programs