According to The Information, OpenAI is in preliminary talks to acquire Trail of Bits, a security research company in New York. This company, known for auditing software and blockchain systems, would add another layer of expertise in professional network security services for OpenAI. The negotiations are still ongoing, and neither party has revealed the acquisition price or financial structure. They also refuse to confirm any discussions publicly.
It has a long history; "Patch the Planet" has already been carried out
The two organizations have not been in contact before. In June this year, OpenAI and Trail of Bits jointly launched the “Patch the Planet” project, which specifically identifies vulnerabilities in popular open-source software. It combines AI automatic detection with manual review by experts, develops and publishes fixes when weaknesses are found, and helps maintainers in various industries strengthen their code. Trail of Bits also integrates OpenAI’s security guidance models into its repair process, using AI to check for defects and accelerate the development and review of patches.
Security is the core weakness of AI, and the acquisition aims at "cross-domain mine clearance"
The report highlights the underlying logic: as large language models become more deeply integrated into critical systems, security has shifted from a bonus to an urgent priority, and the need for rigorous auditing of code and infrastructure has increased significantly. The extensive experience in vulnerability research, adversarial testing, and code auditing that Trail of Bits has accumulated over the years can precisely help OpenAI cope with threats across multiple environments. At a time when GPT-6.1 Astra is stuck due to security risks and the model keeps going "over the edge," bringing in an established auditing company under its control is clearly a move by OpenAI to prepare for AI risks in advance.