Gimlet Labs, a startup founded by Stanford adjunct professor Zain Asgar, has secured an $80 million Series A funding round to address the AI inference bottleneck challenge. According to TechCrunch, Gimlet Labs has developed a software solution called the ‘multi-silicon inference cloud’ that enables AI workloads to operate concurrently on various hardware platforms such as NVIDIA, AMD, Intel, ARM, Cerebras, and d-Matrix chips.
This approach allows AI applications to utilize both traditional CPUs and AI-optimized GPUs, as well as high-memory systems, optimizing performance and efficiency. As Asgar explained to TechCrunch, the software dynamically allocates tasks to different hardware components based on their specific capabilities, enhancing overall workload efficiency.
The lead investor, Tim Tully from Menlo Ventures, highlighted the significance of Gimlet Labs’ technology in a blog post, noting the evolving hardware landscape and the need for a versatile software layer to fully leverage the potential of diverse hardware configurations.
Given the projected exponential growth in data center spending by 2030, with current hardware utilization rates hovering between 15% to 30%, Gimlet Labs aims to improve AI workload efficiency, potentially unlocking cost savings and resource optimization.
Source: TechCrunch