The cybersecurity budget pie is getting carved up differently these days. Enterprise security leaders are moving dollars away from traditional penetration testing and vulnerability scanning tools, redirecting them toward AI-native solutions built by companies like Anthropic and OpenAI.
By 2026, roughly 70% of organizations are expected to allocate more than 10% of their security budgets specifically to AI-focused investments. Early-stage AI security deployments typically commanded just 5-7% of the budget. Mature deployments are now pushing into the 15-25% range and beyond.
Why CISOs are making the switch
Advanced AI models, including Anthropic’s Claude and OpenAI’s GPT-based offerings, can detect vulnerabilities and automate defensive strategies at a scale that legacy tools simply cannot match. Security experts have compared the magnitude of this transition to the shift toward cloud computing.
OpenAI’s billion-dollar infrastructure play
OpenAI has pledged $1 billion in AI resources targeted at small utilities and local governments. That commitment includes training and support tailored specifically for critical infrastructure operators, including water treatment plants, power grids, and municipal networks.
The pricing problem nobody wants to talk about
AI security solutions increasingly operate on consumption-based token pricing, meaning organizations pay based on how much they actually use the models. Extensive usage of AI security models can run organizations as much as $1 million, a figure that can balloon quickly during active incident response when query volumes spike.
An open letter signed by over 100 firms, including Google and Microsoft, called for increased budgets and improved AI integration across cybersecurity frameworks.