- ESET NeoCortex, a state-of-the-art foundational AI model powering ESET LiveCortex, enables deeper incident understanding, reconstructs missing incident context, and provides early insight into attack progression for defenders aiming for “minus-one-day defense”.
- Via highly optimized training, cybersecurity specialization, and smart engineering, ESET NeoCortex delivers significantly more accurate, defense-oriented results at 1/200th of the training cost of general-purpose AI models.
- ESET NeoCortex, with a never-go-rogue design, employs multi-layered measures including restricted sandboxes, read-only tools, continuous monitoring, a kill switch, and human expert oversight of critical response decisions.
- Developed and self-hosted within ESET’s secure environment, the model ensures that critical workloads, sensitive security data and model operations remain under ESET’s control, reducing exposure to abrupt shifts in third-party AI providers, commercial or geopolitical spheres.
BRATISLAVA, Slovakia, Oct. 28, 2026 (GLOBE NEWSWIRE) — ESET, a global leader in cybersecurity, today introduced ESET NeoCortex, its specialized, foundational cybersecurity model powering ESET LiveCortex. Available within the ESET PROTECT Platform, ESET LiveCortex connects, enriches and prioritizes fragmented security signals to turn them into actionable incidents.
Powered by ESET NeoCortex, ESET LiveCortex processes billions of raw telemetry events daily, correlating the signals that matter into actionable incidents. Just 0.00111% of those events – approximately 1 in 90,000 – contribute to incidents presented to customers, separating meaningful activity from noise. By reducing alert fatigue and prioritizing incidents that warrant attention, ESET’s AI technology strengthens human expertise and helps defenders act before attackers reach their objectives.
“AI is accelerating both cyberattacks and cyberdefenses. With ESET NeoCortex, we are applying our AI investment to develop specialized foundational cybersecurity intelligence that remains under our control and strengthens, rather than replaces, human expertise,” said Juraj Malcho, Chief Technology Officer at ESET. “This is an important step toward a new generation of predictive, AI-powered security operations leveraging world-model architecture.”
Built for structured, time-dependent and adversarial security data, NeoCortex draws on patterns learned from real security incidents to analyze activity in context, recognize similar cases and estimate an attacker’s likely next step.
The need for this contextual analysis becomes especially clear during periods such as Black October – specifically the week leading up to Oct 25, 2026 – when critical security issues affecting PHP, Linux, GitLab and other key systems demanded rapid, prioritized responses. ESET NeoCortex is designed to help analysts make these critical decisions, while keeping human defenders in control.
Lowering the noise, improving incident investigation
Modern business environments generate vast amounts of raw security data across endpoints, identities, cloud workloads, networks, and applications. With shrinking time to detect and respond to attacks, the challenge is to process the inputs and determine which signals belong together, which point to a possible threat, and what action to take next.
Within the ESET PROTECT Platform, ESET LiveCortex acts as a nerve center for business protection. It connects signals that may appear harmless in isolation – such as sign-in anomalies, suspicious process launches, unusual connections, configuration changes and subsequent attempts to access data – into prioritized incidents. This gives defenders a focused view, reduces manual investigation and prevents repeated low-value alerts from consuming analyst time.
ESET NeoCortex – an AI model built on world-model architecture – enhances the process of reducing raw telemetry to focused incidents by learning how digital environments behave, how events relate and what may happen next, helping security teams investigate incidents more efficiently and identify potential attack progression early.
Minus-one-day defense: Cost-effective and built for security
These investigation capabilities require a model built specifically for the structure and behavior of security data. While general-purpose large language models are designed to produce plausible responses across a wide range of topics, security operations require consistent analysis of entities, timestamps, indicators, process relationships and incomplete observations across a developing incident.
ESET NeoCortex is designed for these requirements. It can reconstruct missing incident context, create compact representations of complex incidents and identify anomalous events within a wider attack sequence. Its world-model approach also estimates how an attack may progress, giving defenders an opportunity to intervene before the next stage unfolds – a “minus-one-day defense.”
It does not operate as a standalone decision maker. It forms part of a layered security architecture combining specialized models, deterministic controls, correlation logic, validation, auditability and expert human oversight.
ESET’s internal research shows that a specialized model such as ESET NeoCortex can be trained at a 200 times lower cost than a large general reasoning model, while delivering more accurate, defense-focused results. These efficiencies come from ESET- optimized training, focus on cybersecurity vocabulary, purpose-built security engineering, lower computing requirements and privately owned hardware.
For defenders, this translates into more frequent updates to learned incident patterns, lower infrastructure overhead and more predictable costs, with computing resources focused on security analysis rather than broad conversational capabilities.
Securing AI tools and sovereignty
As AI becomes integral to cybersecurity infrastructure and a target itself, ESET has designed ESET NeoCortex to operate under strict safeguards to ensure that the AI never escapes and never “goes rogue” on the open internet. These include restricted sandboxes, read-only tools, continuous monitoring, defined usage and token limits, a kill switch and human expert oversight of critical response decisions.
Developed and operated on ESET’s own infrastructure, ESET NeoCortex keeps sensitive security data and model operations under ESET’s control. This self-hosted approach reduces reliance on third-party AI providers and supports stronger data governance, consistent protection and operational continuity reducing customer exposure to shifts in geopolitical, commercial, and AI provider spheres.
About ESET®
ESET protects organizations, critical infrastructure, and individuals, helping them build digital resilience and confidence in an increasingly AI-driven world. With more than 35 years of cybersecurity expertise, over 25 years of AI innovation, and 11 global R&D centers, ESET develops proprietary technologies designed to protect against evolving cyber threats and secure the broader AI ecosystem. Grounded in scientific rigor and taking a preemptive approach to security, ESET delivers effortless protection that remains under expert human oversight. Headquartered in the European Union and privately owned, ESET protects millions of users and more than 500,000 organizations across 178 countries. Committed to responsible AI and customer trust, ESET protects the progress that technology enables. For more information, visit www.eset.com or follow ESET’s social media, podcasts, and blogs.
Media contact: Jessica Beffa [email protected] 720-413-4938
Disclaimer: The above press release comes to you under an arrangement with GlobeNewswire. AddOn News takes no editorial responsibility for the same.