Armadin, a new cybersecurity startup led by Kevin Mandia, has raised $255.5 million in a Series B funding round. The deal values the company at more than $2.5 billion. The size of the deal shows how much trust investors now place in AI-based security tools.
The company has a clear goal. It wants to use AI agents to find weak points in company systems before real attackers find them. Instead of relying only on regular security checks, Armadin wants its software to test a company’s systems on a constant basis.
The idea is simple. If a company has a security hole, it is better for the company to find it first. Armadin wants AI agents to act like attackers, test possible ways into a system, and show security teams where the real risks are.
The new capital also gives the young company a very strong financial base. Armadin has now raised more than $445 million in total. This is a remarkable amount for a company that became public only months ago.
Kevin Mandia Is Behind the Startup
Armadin was founded by Kevin Mandia, a well-known name in cybersecurity. Mandia is best known as the founder of Mandiant, a major cybersecurity company that worked on some of the world’s most serious cyber incidents.
Mandiant was later sold to Google for $5.4 billion in 2022. That past success gives Mandia a strong reputation among investors and large companies. His new company is now taking a different approach to cyber defense, with AI at the center of its work.
Mandia believes the threat from AI is changing the security world. AI can help attackers find weaknesses much faster than before. It can also help them connect several small problems into one larger attack.
That creates a new problem for companies. A security team may know about one weak password, one exposed service or one software flaw. Each issue may seem small by itself. But an attacker could connect those weaknesses and use them to reach an important system.
Armadin wants to find those attack paths before criminals can use them.
A $255.5 Million Series B Round
The latest deal is a $255.5 million Series B round. Andreessen Horowitz and Accel co-led the round.
Several other major investors also took part. They include Bain Capital Ventures, Redpoint, 8VC, Ballistic Ventures, Google Ventures, In-Q-Tel, Kleiner Perkins and Menlo Ventures.
The new deal came only about six months after Armadin raised $190 million in its Series A round in March. The company has therefore raised more than $445 million in a very short period.
That pace is unusual even for the current AI market.
Armadin had also received an earlier $24 million seed investment before its public launch. The company later came out of stealth with about $190 million in funding in March 2026.
The latest round gives Armadin the money it needs to expand its technology, research work and business operations. It also gives the company more room to hire experts and build its AI security platform for large customers.
What Armadin Actually Does
The main idea behind Armadin is called agentic security.
The word may sound complex, but the basic idea is easy to understand. Armadin uses AI agents that can carry out security tests on their own.
These agents can look at a company’s digital systems and search for weaknesses. They can then test whether those weaknesses can be used in a real attack.
The key difference is that Armadin does not want to show customers a long list of possible problems. It wants to show which problems can actually work together.
For example, a company may have three separate security weaknesses. One may affect a public-facing system. Another may allow access to an internal service. A third may give access to important cloud data.
A normal security report may show all three issues as separate problems.
Armadin wants its AI agents to see the larger picture. They can try to connect those weaknesses and find a real route from the first problem to the final target.
This can help security teams focus on the risks that matter most.
From Regular Tests to Constant Tests
Traditional penetration testing usually happens at set times.
A company may hire a security team to test its systems once a year or several times each year. Experts try to break into the system in a controlled way. After the test, they give the company a report.
That process remains useful. But modern companies change very quickly.
A company may add a new cloud service today, change its network tomorrow and release new software next week. Each change can create a new security problem.
A test from six months ago may not show the risks that exist today.
Armadin wants to change this model.
Its platform can run what the company calls always-on agentic swarms. These AI agents can test systems on a regular basis instead of waiting for the next scheduled security review.
This is one of the main reasons investors see a large market opportunity.
If the technology works as planned, companies could have a much more current view of their security risks.
AI Agents Act Like a Team
One of the most interesting parts of Armadin’s technology is its use of many AI agents.
Instead of one AI system doing every task, Armadin uses groups of specialized agents. Each agent can focus on a different part of the security test.
Together, these agents can act more like a large security team.
The agents can search for weaknesses, test possible attack routes and connect separate issues. The system can then give the security team a clearer view of what an attacker could do.
This is the idea behind the phrase agent swarm.
The word “swarm” refers to many agents that work together toward the same goal.
The approach is important because modern cyberattacks can involve many steps. An attacker may need to find one weakness, use it to gain access, move to another system and then find a way to reach valuable data.
Armadin wants its AI agents to test those steps in a controlled environment.
A Large Real-World Test
Armadin has already carried out a major test of its technology.
In August, the company worked with TENEX.ai on a three-day security exercise. Armadin said it used 26,000 autonomous agents during the test.
Those agents carried out about 17 million offensive actions against more than 25,000 services. The exercise produced 238 security findings, including 98 significant findings.
The agents also found 38 validated attack paths.
These figures show the scale that AI can bring to security testing.
A human security team would not be able to perform millions of separate actions at the same speed. AI agents can carry out many tests in a short period.
Armadin says the agents did not have privileged credentials or access to source code during the exercise. The company also says its agents worked through a control layer with oversight from a safety model trained with feedback from human security experts.
These details matter because security testing must be controlled. A system that attacks its own customer without limits could create the very problem it was meant to prevent.
Why This Matters as AI Gets Stronger
The rise of AI has created a new problem for cybersecurity teams.
AI can help defenders. But it can also help attackers.
A person who wants to attack a company may use AI to search for weaknesses, write code, study systems or automate parts of an attack. This can reduce the time and skill needed for some cyber tasks.
That means companies need faster ways to defend themselves.
Armadin’s idea is to use AI on the other side of the fight.
If attackers can use AI to find weaknesses, defenders can also use AI to find those same weaknesses first.
This creates a race between offensive and defensive AI.
For Armadin, the goal is not to wait until a breach happens. The goal is to find a possible attack route before a criminal can use it.
Investors See a Large Market
The huge funding round suggests that major investors see strong demand for this type of technology.
Companies now depend on cloud services, software tools, remote access and large amounts of digital data. This makes their systems more complex.
More complexity can create more places for attackers to enter.
At the same time, security teams face a problem with the sheer number of alerts they receive. A company can have thousands of possible security issues, but not every issue is equally dangerous.
Armadin wants to help solve that problem by showing which weaknesses can form a real attack path.
That approach can save security teams time. Instead of checking every alert with the same level of urgency, they can focus first on risks that could lead to serious damage.
Armadin Has Moved Very Fast
The speed of Armadin’s growth is another notable part of the story.
The company has now raised more than $445 million. Its latest round alone added $255.5 million. Its valuation has crossed $2.5 billion.
All of this has happened only months after the company came out of stealth.
The company has also started work with Fortune 500 companies and government customers, according to Armadin. The names of those customers have not been made public.
That early customer interest is important. Funding can give a startup time to build, but real customers show whether the product can solve a real business problem.
Armadin now has to prove that its technology can work at a large scale and provide clear value to customers.
The Mandiant Legacy Helps
Mandia’s history with Mandiant is a major advantage for Armadin.
Mandiant became known for its work after major cyber incidents. Its experts helped companies understand how attackers entered their systems, what they did after gaining access and how to stop further damage.
Google’s $5.4 billion acquisition of Mandiant in 2022 was one of the biggest signs of the company’s value.
That history gives Mandia credibility with large businesses.
Cybersecurity buyers often deal with very sensitive systems. They need to trust the company that gets access to those systems. A founder with decades of security experience can make that decision easier.
But reputation alone will not be enough.
Armadin now has a very high valuation. The company must show that its technology can deliver results that justify that value.
What Comes Next for Armadin
The next stage will be about execution.
Armadin has a large amount of capital, major investors and an experienced founder. It now needs to turn those advantages into a strong business.
The company can use the new money to expand its AI systems, hire more security experts, work with more customers and improve its platform.
It will also need to make sure its AI agents remain safe and controlled.
That point is especially important because Armadin’s software is designed to behave like an attacker. The same technology that helps find a security hole must not create a new security problem.
Trust will therefore be just as important as speed.
A New Chapter for Cybersecurity
Armadin’s $255.5 million funding round shows how quickly AI is changing cybersecurity.
The company is not simply building another vulnerability scanner. Its larger idea is to create AI agents that can test systems, connect weaknesses and show security teams how a real attack could happen.
Its $2.5 billion-plus valuation shows that investors believe this approach could become a major part of enterprise security.
For companies, the appeal is clear. Cyber threats do not wait for the next security review. Digital systems change every day, and attackers can use new technology to find weak points faster.
Armadin wants defenders to have the same advantage.
With $255.5 million in new funding, more than $445 million raised in total, a valuation above $2.5 billion, and Kevin Mandia at the helm, Armadin has quickly become one of the most closely watched AI cybersecurity startups.
The big question now is not whether investors believe in the idea. They clearly do.
The real question is whether Armadin can turn that belief into a reliable product that large companies can trust every day.
Also Read – Anthropic IPO May Arrive Before Thanksgiving