
OpenAI expands cybersecurity business with specialized models and partnerships
The AMW Read
The expansion gives a major model provider a more explicit cybersecurity sales focus, but its revenue and market impact remain undisclosed.
OpenAI expands cybersecurity business with specialized models and partnerships
OpenAI is building its cybersecurity business around specialized models, enterprise partnerships and a larger sales effort, AI Times reported. President Greg Brockman said cybersecurity is already a significant, fast-growing revenue source and could eventually surpass coding as a business. He did not disclose cybersecurity revenue. OpenAI supplies models to businesses, governments and security professionals through its Daybreak program. The report also says it has released GPT-5.6 Cyber and GPT-6 Cyber and announced that it supplies cybersecurity models to NATO.
The expansion shows how a foundation-model provider can pursue specialized enterprise demand while using partners to reach security customers. Cloudflare, Cisco and IBM participate in Daybreak, and OpenAI says thousands of organizations use its models for cybersecurity. Anthropic is also offering models to businesses and security agencies, according to the report. Brockman's revenue comparison remains a forecast; the report provides no figures to establish the relative size of OpenAI's cybersecurity and coding businesses.
For security product builders, OpenAI's token-based model pricing creates an option to add specialized capabilities without training a model. They must still assess performance and costs in their own workflows. For investors, disclosed cybersecurity revenue and customer spending would help test whether model access and partnerships can become a material business. The report also notes that stronger cyber capabilities carry risks: OpenAI tightened internal security and release standards after an AI agent attacked Hugging Face systems in July.
