Aikido Launches Local AI Model for Cybersecurity
Belgium-based cybersecurity company Aikido has released an open-weight AI model designed to help organisations detect and defend against software vulnerabilities without sending sensitive code to external AI providers.
Belgian cybersecurity company Aikido has introduced Altar, an open-weight AI model built specifically for cybersecurity. The model is designed for local deployment, allowing organisations to use AI-powered security capabilities while keeping sensitive source code and security data inside their own infrastructure.
Artificial intelligence is becoming a new battleground in cybersecurity.
The same technology that can help security teams discover vulnerabilities can also help attackers find weaknesses more quickly. As AI-powered cyber threats become more sophisticated, companies are looking for defensive systems that can analyse software without exposing sensitive information to outside services.
Belgium-based cybersecurity company Aikido is responding with a new approach: an AI model that organisations can run within their own infrastructure.
Aikido Introduces Altar
Aikido has released Altar, an open-weight artificial intelligence model designed specifically for cybersecurity applications.
The company says Altar is a compressed and customised version of Z.AI's open model GLM-5.3. It is intended to power Aikido's security products while allowing customers to deploy the technology locally.
This local deployment model is particularly important for companies that cannot or do not want to send source code, internal infrastructure information or other sensitive security data to an external AI service.
The central idea behind Altar is simple: powerful cybersecurity AI should be able to operate where the sensitive data already lives.
Why Local AI Matters for Security
Many businesses are cautious about sending proprietary software and security information to third-party AI providers.
Source code can contain intellectual property, authentication details, business logic and information about how critical systems operate. For financial institutions, governments and other organisations handling highly sensitive information, keeping that data inside controlled infrastructure can be particularly important.
An open-weight model that can operate locally gives security teams another option: using advanced AI capabilities without necessarily transferring their most sensitive information to an external inference provider.
Built for Defensive Cybersecurity
Altar is being developed as part of Aikido's wider effort to use AI for automated cybersecurity.
The company has already developed AI-powered penetration-testing systems capable of searching for vulnerabilities, validating whether weaknesses can actually be exploited and helping organisations address them.
With Altar, Aikido is taking that approach further by putting a dedicated security model directly into its technology stack.
Demand Is Growing as AI Changes Cyberattacks
The launch comes as cybersecurity companies face a changing threat environment.
Criminal groups can increasingly use AI to analyse software, identify potential weaknesses, generate malicious code and accelerate parts of an attack. That creates pressure on defenders to adopt AI themselves.
But the use of AI in security also creates a new question: how can organisations use increasingly capable models without giving those models unnecessary access to sensitive information?
Belfius Among the Customers
Aikido plans to integrate Altar into security solutions used by customers including Belgian bank Belfius.
Financial institutions are particularly significant users of cybersecurity technology because they manage valuable financial information and operate systems that are attractive targets for cybercriminals.
Local AI deployment could therefore become increasingly relevant for organisations that need advanced security analysis while maintaining strict controls over where their data is processed.
A European Push for More Control
Aikido's launch also fits into a broader European technology debate about digital sovereignty.
European companies and governments have increasingly focused on maintaining control over sensitive data, infrastructure and critical technology. AI has added another dimension to that discussion because powerful models often require substantial computing resources and are frequently accessed through external cloud services.
Open-weight models that can run within an organisation's own infrastructure can potentially give businesses greater control over how AI is deployed, although local deployment also requires appropriate hardware, expertise, security controls and ongoing maintenance.
The Bigger AI Security Shift
The importance of Altar goes beyond one cybersecurity product.
The launch reflects a broader shift toward specialised AI models designed for specific industries rather than general-purpose chatbots alone. Cybersecurity is one of the areas where specialised models can potentially provide value because they can be integrated into highly technical workflows and trained or adapted for security-related tasks.
If more companies adopt this approach, cybersecurity AI could increasingly become an internal layer of digital infrastructure rather than simply another cloud-based software service.
As AI becomes part of both the attack and defence sides of cybersecurity, the next important question may not simply be how powerful the model is — but who controls where that intelligence operates and what information it can see.
