- Published on
OpenAI puts $1 billion behind cyber defense after unveiling AI that can find zero-days
- Authors
- Name
- DeepHorus
OpenAI announced a $1 billion initiative to bolster cyber defense, offering subsidized access to its security platform over the next six months. This move comes shortly after the company revealed that its new AI model, Astra, can autonomously discover and exploit unknown software vulnerabilities, a capability previously limited to elite human hackers. The program is designed to help critical organizations build defenses against AI-powered attacks.
The financial commitment is not a direct grant but provides discounted access to OpenAI's Daybreak cybersecurity product, along with training and technical support. Daybreak operates in two tiers: Blue for standard defensive operations and Red for sensitive tasks using specialized models. OpenAI reported that approximately 2,000 organizations are already utilizing the platform.
Priority access will be granted to water and electricity infrastructure, state and local governments, community banks, nonprofits, and open-source maintainers. This latter group is particularly relevant for the cryptocurrency sector, as it includes the small, often volunteer teams who maintain vital blockchain software like Bitcoin Core and Ethereum clients, which secure billions in value.
The urgency behind this initiative stems from recent cybersecurity incidents where AI tools were used offensively. In August, a flaw in the open-source BTCPay Server exposed Lightning node credentials, leading to fund theft. More recently, Core Lightning warned operators to disconnect after AI-generated reports uncovered real vulnerabilities, and hardware wallet maker Coldcard issued new firmware following a $114 million theft.
These events culminated in an open letter from over three dozen crypto and tech companies, including Coinbase, Block, and ARK Invest. The letter urged AI labs to provide early access to their most powerful models for security research, arguing that attackers already had access to such tools, leaving defenders at a disadvantage.
OpenAI's Astra model represents a significant escalation, capable of finding and weaponizing zero-day exploits without human guidance. By providing defensive tools, the company aims to level the playing field before such capabilities become widespread. Eligible organizations can now apply through the Daybreak website.
The initiative signals a proactive stance from OpenAI on the dual-use nature of advanced AI. As AI lowers the barrier to entry for sophisticated cyberattacks, the focus on fortifying critical infrastructure and open-source projects is a crucial step. The success of this program will be measured by its ability to enhance resilience across vulnerable sectors.
