Toronto-based startup Taalas has secured $169 million in funding, bringing its total to $219 million. The company focuses on developing chips that enhance the efficiency and cost-effectiveness of artificial intelligence applications. High-profile investors, including Fidelity and venture capitalist Pierre Lamond, contributed to this funding round.
In a competitive landscape for AI technologies, Taalas's innovative chip design prints AI model segments directly onto silicon, allowing for tailored chips for specific applications. CEO Ljubisa Bajic noted that their bespoke approach enhances processing speed. Taalas constructs nearly complete chips with around 100 layers, completing final adjustments on two metal layers, which significantly reduces manufacturing time compared to traditional AI processors.
Currently, the company is capable of producing chips for simpler AI models, but it aims to develop a processor capable of handling more advanced models, such as GPT-5.2, by year-end. This trend reflects a growing interest in specialized AI inference technologies, especially following Nvidia's recent $20 billion licensing deal with Groq.