Enterprise data stores were already sprawling before generative AI arrived. Sensitive records sit in cloud buckets, SaaS applications, data warehouses and on-premises file shares, often copied, rarely classified, and governed by permissions inherited years earlier from a project that no longer exists. Security teams built their controls around a perimeter that assumed a human being was the one reaching for the data, and that assumption held for as long as it was true.
It is no longer true. AI agents, software processes that act on their own to complete tasks, have become some of the heaviest users of corporate data. They query warehouses, read documents, call internal services and touch knowledge stores, and many of them do so with credentials copied from a person or a service account that was granted broad access for convenience. The result is a governance gap that opens quietly and widens quickly.
Sentra Inc., a data security company co-headquartered in New York and Tel Aviv, has spent five years building a platform aimed at that gap. On November 4, 2026, the company said it had raised a $150 million Series C led by Capital One Ventures, with a Capital One representative taking a seat on its board. The company said the round follows a year in which its business grew more than 300 percent year over year.
The announcement lands in a crowded but fast-moving corner of cybersecurity. Venture money has flowed toward tools that inventory data and constrain who can reach it, because regulators, insurers and boards now ask questions that legacy visibility products cannot answer. Sentra is betting that the specific problem of agentic access, not data discovery in general, is where the next round of enterprise budget will be spent.
Key Facts
The $150 million Series C was led by Capital One Ventures, the venture arm of Capital One Financial Corp., and brings total funding in Sentra to more than $260 million since the company launched, according to SecurityWeek, which reported the figure on November 4, 2026. Calcalist (Ctech) reported on November 4, 2026 that existing investors Bessemer Venture Partners, Zeev Ventures, Standard Investments, Munich Re Ventures and Key1 Capital all participated in the round. Key1 Capital had led a $50 million Series B in April 2025. Sentra did not disclose a valuation, but chief executive and co-founder Yoav Regev told Calcalist that the company's valuation has tripled since its previous funding round. Calcalist reported that while Sentra has not reached a $1 billion valuation, estimates suggest it is close.
Calcalist reported on November 4, 2026 that the round includes a small secondary component allowing employees to sell shares, with most of the capital going directly into the company. Sentra employs 130 people, 65 of them at its Tel Aviv offices, and counts around 100 major enterprise customers across the United States and Europe. The company says it generates tens of millions of dollars in revenue. Sentra was founded in 2021 by Regev, Asaf Kochan, Ron Reiter and Yair Cohen. Regev once headed the cyber department of the Israeli military's Unit 8200 and later commanded its Cyber Center; Kochan is a former commander of Unit 8200 and now serves as president; Reiter is chief technology officer and previously headed Oracle's development center in Tel Aviv; Cohen is vice president of product and held senior product roles at Microsoft and Datadog.
The platform protects data at rest across cloud environments, SaaS applications, data warehouses, on-premises file shares and AI knowledge stores. Before an incident, it classifies sensitive data using models tuned to each customer's business, maps which people, service accounts and AI agents can reach a given dataset including inherited permissions, enforces least privilege for non-human identities, and removes redundant or high-risk data. During an incident it flags unusual access and automatically revokes permissions or removes risky sharing. After an incident it reconstructs what an agent could have reached, with business context attached.
SiliconANGLE reported on November 4, 2026 that all analysis runs inside the customer's own environment, so sensitive data never leaves it. That architectural choice matters to regulated buyers, who can adopt the product without shipping copies of their most sensitive records to a third party. Lyft chief information security officer Chaim Sanders is quoted in the company's announcement saying his teams need to know what data agents can reach before those agents are turned on.
Sentra plans to use the proceeds to accelerate visibility into agents, copilots and models, to enforce least privilege continuously rather than as a periodic audit, and to speed containment when something goes wrong. It also intends to expand its go-to-market organization, grow its brand and marketing teams, double its workforce in Israel and the United States, open a London office and enter the United States federal market, with sales there expected to begin in the second half of 2027. Regev has said the company aims to reach $100 million in sales and is considering an acquisition in areas such as data loss prevention and identity management. Sentra names Cyera and Varonis as competitors.
Analysis
The interesting part of this round is not the size of the check. A $150 million Series C is a large but not extraordinary number in cybersecurity, and Sentra's total of more than $260 million puts it in the same bracket as several peers chasing the data security posture management market. What stands out is who led it. Capital One Ventures is the investment arm of a bank, not a generalist growth fund, and a bank taking a board seat signals that the buyer is no longer only the security organization but the enterprise that has to prove to a regulator that it knows where its sensitive data went.
That distinction matters because financial institutions have been among the most exposed to cloud data sprawl. A bank cannot simply prohibit AI adoption while competitors deploy it, and it cannot allow agents to inherit the accumulated permissions of a decade of service accounts. The pitch Sentra is making, that you secure the data rather than the perimeter because the perimeter no longer contains the workload, is precisely the argument a regulated buyer needs to hear. Regev framed the shift bluntly, saying traditional perimeter controls were not built for agents moving at machine speed, and that organizations can instead make sure the data those agents reach is already secured.
The bigger picture here is that identity, not network location, has become the control plane for enterprise data, and the companies that can map non-human identities at scale will hold leverage over the next generation of security budgets. Sentra's product is organized around exactly that map: which identities, human and machine, can reach which dataset, and what happens when one of them reaches for something it does not normally touch. That is a harder product to build than a scanner that reports on storage misconfigurations, and it is harder to displace once it is wired into an enterprise's access decisions.
The numbers also suggest a company in the middle of a growth spurt that has not yet settled into a stable shape. Growth above 300 percent year over year on a base of around 100 enterprise customers implies a business that is still adding large accounts rather than expanding within a fixed set, and a workforce of 130 people supporting that customer count is lean. Tripling the valuation while staying below $1 billion is a credible trajectory but not a guarantee. The plan to double headcount, open a London office and enter the federal market in the second half of 2027 will consume much of this round, and execution across three simultaneous expansions is where promising security companies most often stumble.
Why It Matters
The agentic data security category exists because of a specific failure mode. An agent is given a credential so it can complete a task, and that credential carries permissions no one audited. The agent then reads a dataset it was never meant to see, or writes to a system it should only read, and because the action looks like legitimate authenticated traffic, existing monitoring does not flag it. Sentra says only a small minority of organizations report having governance in place for this, and the anecdotal evidence from security leaders supports the claim.
Reuters reported on September 8 that enterprises were still struggling to inventory the data their AI systems could reach, a reminder that the visibility problem predates the agent boom and has worsened as agents multiplied. The practical consequence is that AI adoption timelines now depend on data governance work that many organizations deferred. A company that cannot say which datasets contain personal data cannot responsibly connect an agent to them, and that constraint is slowing deployments that executives have already announced publicly.
There is also a regulatory dimension. Data protection regimes in Europe and sector rules in the United States increasingly require organizations to demonstrate that access to sensitive records is limited to those who need it, and to show what happened after the fact. Sentra's post-incident reporting, which reconstructs what an agent could have reached with business context, addresses a question that regulators ask and that most security tools cannot answer. SecurityWeek reported on November 4, 2026 that the funding will go toward exactly this combination of visibility, continuous least privilege and rapid containment, which is a reasonable reading of where enterprise demand is heading.
Next Up
Sentra's immediate task is hiring. Doubling a workforce of 130 across two countries while opening a third office is an operational challenge that will test management as much as the technology. The federal market entry, slated for the second half of 2027, adds certification and procurement timelines that move on their own schedule. Regev has also signaled appetite for an acquisition in data loss prevention or identity management, which would widen the platform but also introduce integration risk at the moment the company can least afford distraction.
The competitive field will not stand still. Cyera and Varonis are named by Sentra as rivals, and every major cloud provider is shipping its own data security controls. The question investors are effectively answering with this round is whether a specialist that governs what agents can reach can hold its position as those larger platforms absorb adjacent features. Capital One Ventures has placed a bet that it can, and a board seat gives it a close view of whether the bet pays off.
Comments (0)
Log in or sign up to leave a comment.
No comments yet. Be the first to share your thoughts.