Huskeys Raises $27M—AI Traffic Turns the Network Edge Into a Market

Huskeys’ September 2, 2026 financing release disclosed a $27 million Series A led by Blackstone Innovations Investments, bringing total funding to $35 million. The release lists Merlin Ventures, Skinos Ventures, Zscaler Ventures, Okta Ventures, Bright Pixel Capital and SV Angel among the participating investors.

The financing valued the roughly year-old Israeli cybersecurity company at about $100 million, according to The Times of Israel’s funding account. The investment backs Huskeys’ attempt to establish Network Edge Security Management, or NESM, as a distinct layer above existing internet-facing defenses.

The investment backs coordination, not a new firewall

Huskeys is not presenting NESM as a replacement for content delivery networks, web application firewalls or cloud-native security controls. The proposed platform connects those systems with load balancers, virtual private clouds and security groups, creating a common intelligence and control layer across infrastructure that remains responsible for enforcement.

The operational distinction matters. A WAF evaluates application requests, a CDN distributes content and may filter traffic, while cloud controls govern resources inside a provider’s environment. Huskeys is trying to translate their separate configurations into a common model so security teams can assess exposure and coordinate policy changes across several enforcement points.

That proposition occupies a potential market between individual security products and the teams managing them. NESM becomes a meaningful category only if customers treat cross-vendor policy translation, testing and lifecycle control as a separate function worth purchasing. If the platform mainly aggregates alerts and sends operators back to existing consoles, it adds an interface without resolving the underlying fragmentation.

AI traffic turns classification into a policy problem

Autonomous software complicates the old division between human users and unwanted bots. A non-human request may come from a malicious scraper, an internal automation process or an AI agent acting for a legitimate customer; identifying automation is therefore not enough to determine whether the request should proceed.

The decision may depend on identity, behavior, requested resources, rate, application sensitivity and prior activity. Relevant evidence can be divided among an identity provider, CDN, WAF and several cloud environments, giving a unifying policy layer a plausible role. It also raises the stakes because an incorrect centralized decision could be propagated across every connected control.

In Channel Insider’s account of Huskeys’ company-provided metrics, the startup says it analyzes more than one trillion web requests and thousands of network configurations daily; it also says revenue roughly quadrupled quarter over quarter after commercial activity began in early 2026 and its customer base roughly quintupled. These figures have not been presented as audited results or independent measurements of detection quality.

Traffic volume by itself cannot show whether the system makes better decisions. More consequential measures include the share of legitimate automated requests blocked, the share of malicious requests missed, the time needed to reverse a bad rule and whether identical policy intent produces consistent outcomes across vendors.

A control plane must expose vendor differences

The strongest version of the Huskeys proposition is consistent, testable enforcement, not centralized visibility alone. A credible control plane should expose contradictory rules, identify which product will act on a request and preserve policy meaning when translating an instruction into vendor-specific configurations.

That is harder than normalizing product names or rule labels. WAFs, CDNs and cloud platforms differ in rule languages, evaluation order, telemetry and rollback mechanisms. A capability available in one product may have no equivalent elsewhere, while overlapping controls may interact in ways that become visible only under production traffic.

A common abstraction must reveal those gaps rather than conceal them behind a universal setting. Buyers evaluating whether the layer reduces fragmentation can ask concrete questions:

• Which integrations can deploy and reverse policies, and which provide read-only telemetry?
• How is a control represented when one connected vendor cannot enforce it?
• Can proposed rules be simulated against representative traffic and introduced through staged deployment?
• Can false-positive and missed-detection rates be compared before and after a policy change?
• Does every automated action retain its evidence, approval, affected assets, vendor-specific translation and rollback state?
• When enforcement points disagree, which policy takes precedence and how is the conflict surfaced?

The answers define the platform’s risk boundary as well as its utility. Centralized orchestration may reduce repetitive work and policy drift, but it can also create a concentration of failure if an incorrect recommendation, compromised integration or faulty translation reaches several defenses at once.

NESM remains a vendor-defined category

The funding demonstrates investor support for Huskeys and its thesis, but it does not establish Network Edge Security Management as a recognized purchasing category. Participation by established security investors may help with integrations and distribution, yet it does not prove that enterprises have created a separate NESM budget or adopted the company’s terminology.

Evidence that a durable category is forming would include published integration coverage, deployments spanning genuinely different vendors and clouds, and independently measured improvements in policy consistency, false positives and rollback performance. Comparable outcome data would say more than aggregate traffic volume because it would show whether the layer improves enforcement rather than merely observing more requests.

The financing terms and product thesis are now public. What remains unresolved is whether Huskeys can make existing edge defenses operate as one governed system without becoming the additional console—and additional dependency—that security teams were trying to avoid.

Also read:

• UK Managed IT Generates £51 Billion, but Security Raises the Bar
• Obsidian Security Raises $85M as AI Agents Inherit SaaS Permissions
• Lumilens Raises $700M to Attack the Network Bottleneck in AI Data Centers