Fastino, a startup bucking the trend of massive, expensive GPU clusters for AI model training, has secured $17.5 million in funding led by Khosla Ventures. Instead of relying on costly, high-end GPUs, Fastino focuses on training AI models using cheaper, readily available gaming GPUs.
This approach could democratise AI development, making it accessible to smaller companies and individual researchers who lack the resources for large-scale infrastructure. Fastino's technology optimises AI training workloads to run efficiently on consumer-grade hardware. The company's success hinges on overcoming the limitations of gaming GPUs, such as lower memory capacity and compute power compared to their professional counterparts.
With this funding, Fastino plans to expand its engineering team and further develop its platform. If Fastino's approach proves viable, it could significantly disrupt the AI training market, potentially reducing costs and increasing accessibility for AI development.