NeuralTrust, a pioneer in autonomous AI security and governance infrastructure, has closed a $20 million (€17.2 million) seed funding round. This milestone represents the largest cybersecurity seed financing raised by a European Union enterprise to date. The capital injection arrives at a critical turning point in corporate tech adoption, as multinational organizations rapidly shift autonomous, agentic AI architectures out of experimental sandbox phases and into high-volume live production environments.
NeuralTrust’s oversubscribed financing was led by Munich-based Alstin Capital, along with strong backing from VentureFriends, Seaya, Kibo Ventures, Banc Sabadell, EA Ventures Plug & Play Fund, and Finaves (venture capital wing of IESE Business School). Moreover, NeuralTrust is funded through highly competitive grants for innovation from the European Innovation Council (EIC) and the Spanish State Research Agency (AEI), thus making the company one of the most trustworthy and sovereign security solutions providers for regulated industries and procurement systems.
Resolving the Governance Blind Spot Created by Autonomous Agents
While corporations are aggressively deploying AI agents to automate data manipulation and customer-facing interactions, agentic adoption is pacing far ahead of traditional IT security capabilities. Traditional security tools—such as legacy firewalls or security information and event management (SIEM) systems—fail to detect or mitigate agent-specific vulnerabilities. They cannot identify when a digital agent is being systematically manipulated via natural language to leak internal corporate secrets or bypass programmatic system controls.
Also Read: AppViewX Introduces Agent Identity Security to Protect Enterprise Infrastructure in the AI and Post-Quantum Era
Unchecked development leads to severe “agent sprawl,” where disconnected teams build or acquire specialized agents on separate foundational models. Each independent agent connects directly to core relational databases, private APIs, and external communication channels. Most enterprise security teams remain unable to audit how many agents are currently executing code within their networks, what data permissions they hold, or if they have been compromised by prompt-injection vectors.
“AI agents are entering enterprise infrastructure faster than security teams can adapt, so the window to establish control is now. They’ve built a genuine control layer, not another point solution, and enterprises can operationalise it today,” said Alexander Meyer-Scharenberg, partner at Alstin Capital.
The True Scale of the Agent Threat Vector:
According to active telemetry analyzed across NeuralTrust’s enterprise deployment base, the platform inspects millions of autonomous interactions daily. Strikingly, approximately 1.2% of all active enterprise AI agent interactions require direct real-time security intervention due to malicious prompt-injection techniques, tool manipulation, or unauthorized data extraction attempts.
A Unified Architecture for Real-Time Agent Monitoring
NeuralTrust delivers a centralized control layer that enables chief information security officers (CISOs) to discover, govern, and defend their entire AI ecosystem through three integrated software products:
- TrustGate: A secure, high-performance gateway that brokers every large language model (LLM), model context protocol (MCP), and API tool call, serving as the single, low-latency enforcement point through which all corporate agent traffic must flow.
- TrustGuard: A reactive, real-time threat detection and runtime guardrail engine designed to inspect agent inputs and outputs, intercepting multi-step attacks regardless of how or where the underlying agent was built.
- TrustLens: A comprehensive posture management layer that inventories, maps, and continuously tracks agent behaviors across both internal corporate servers and external security perimeters.
“This round allows us to keep building the infrastructure layer that makes AI adoption measurable, governable, and safe. Our mission has not changed since day one: turn AI security into a strategic advantage for the enterprises that will define the next decade,” said Joan Vendrell, co-founder and CEO of NeuralTrust.
Driving European Sovereignty and Leading Threat Research
The platform’s growth indicates an unusual commercial trajectory for early-stage software companies. NeuralTrust’s active enterprise client portfolio includes major global institutions such as Iberia, AirEuropa, Abanca, and Banc Sabadell—with 92% of its customer base generating more than $1 billion in annual revenue. Moving into the first half of 2026, NeuralTrust successfully doubled its full-year 2025 annual recurring revenue (ARR) within the first quarter alone.
Beyond commercial product delivery, NeuralTrust functions as a core contributor to international open-source AI security frameworks. The company’s dedicated threat research lab recently discovered and documented two novel vulnerability techniques: “Echo Chamber” (a sophisticated multi-turn conversational jailbreak method) and “Semantic Chaining” (a complex multimodal attack vector). Both definitions have been officially integrated into the OWASP AI Security Project threat taxonomy to help global organizations defend their digital assets.
The newly acquired funding will be utilized to accelerate global engineering recruitment, deepen the native interoperability across NeuralTrust’s three software modules, and expand market operations deeper into the European, UK, and North American corporate sectors.