The London-based artificial intelligence technology startup Fractile has entered a significant investment phase to produce next-generation processing hardware. To meet the increasing demands of AI models, the company has secured $220 million in funding, planning to accelerate its research and development efforts with this investment.
Investors and Goals
Fractile's Series B funding round was led by prominent investors Accel, Factorial Funds, and Founders Fund. Other notable participants in the funding round include firms like Conviction, Gigascale, and Felicis. The company's strategy aims to make AI data processing more efficient by positioning memory and processing units closer together, unlike traditional processor architectures.
Fractile has developed an architecture specifically aimed at enhancing the performance of large AI models. With these systems, it is targeted to achieve up to a 25% increase in processing speed in certain scenarios compared to existing solutions, while also projecting a reduction in costs by up to 90%.
Increasing Processing Load and Requirements
Additionally, issues arising from modern AI applications consuming more resources are coming to the forefront. As the token consumption of more complex models increases, it significantly impacts latency and energy consumption in data center infrastructures. Fractile aims to meet these requirements by providing more effective solutions.
With the new funding secured, Fractile plans to commercialize its chips and expand its operations. The company continues its efforts to grow its teams in the United States, the United Kingdom, and Taiwan.