Toronto-based AI company Cohere Inc. has signed a letter of intent to buy Prague cybersecurity startup Aisle, a move that puts the Canadian firm squarely in the race to defend software systems against increasingly powerful AI-driven attacks.
The acquisition, announced Thursday, comes as security researchers and governments sound alarms about AI models capable of both finding vulnerabilities and exploiting them. Anthropic revealed earlier this year that its Mythos model excels at locating software flaws, while AI agents from OpenAI have broken into government and corporate systems during testing.
Financial terms of the Cohere-Aisle deal remain confidential.
Aisle, founded in 2024, developed AI systems that hunt for and help fix software vulnerabilities. The startup reports discovering more than 500 previously unknown flaws across over 100 codebases in the past year, performance it says matches or beats rival platforms.
Cohere CEO Aidan Gomez praised Aisle’s track record on LinkedIn. The startup has focused on protecting critical open source infrastructure, he wrote, outperforming many higher-profile security initiatives.
Cohere plans to fold Aisle’s autonomous security research into its own AI models. The company also intends to invest heavily in Prague, expanding Aisle’s existing operations and making the Czech capital its global hub for AI cybersecurity.
The three Aisle co-founders have backgrounds at Google DeepMind, Anthropic, and Czech security firm Avast. Ondrej Vlcek will lead Cohere Cyber, a new division. Jaya Baloo becomes Cohere’s chief information security officer. Stanislav Fort takes on the role of vice-president of AI security.
This marks Cohere’s third major deal this year. The company acquired biopharma firm Reliant AI in May and announced a merger with German AI company Aleph Alpha that still awaits regulatory approval. Cohere is also reportedly in talks to raise up to US$3-billion, which would value the company at US$20-billion, with backing from Canadian and German governments.
Meanwhile, Cohere co-founder Ivan Zhang pushed back on fears that AI hands attackers the upper hand. Defenders actually hold the advantage, he argued, since patching known vulnerabilities proves simpler than chaining exploits together for an attack.
The deal also signals strengthening Canada-Europe ties as tensions with the United States grow and American tech giants dominate the AI landscape.
