Artificial intelligence has a crucial role in silicon demand; this procurement amplifies AMD’s Software power, paving the way for effortless AI model deployment.
Advanced Micro Devices (AMD) acquires French firm Mipsology to bolster its artificial intelligence (AI) software skills, intensifying its competition with AI chip giant Nvidia. This move aligns with AMD CEO Lisa Su’s strategy to counter Nvidia with top-tier graphics processing units(GPUs), central processing units (CPUs), and adaptive AI solutions for inference and training. This partnership with AMD is for a significant duration. By integrating Mipsology into its operations, the company anticipates that it will significantly boost the pace of its customer interactions while simultaneously broadening its capacity and expertise in AI software development.
Mipsology specialises in “plug-and-play” software, Zebra, that boosts AI inference speeds without needing modifications to existing neural network models. It enhances field programmable gate arrays(FPGAs) performance, which AMD obtained from the Xilinx acquisition and now terms adaptive computing chips. These chips cater to varying workloads, from edge devices to data centres. The software aligns with major industry frameworks like TensorFlow, PyTorch, and ONNX Runtime. This will further hasten AI solutions, offering AI inference, optimisation, and tools tailored to the company’s devices. It complements the company’s Unified AI (UAI) Software Stack, ensuring consistent AI operations from edge to cloud. The acquisition will fortify complete AI software offerings, enhancing the open ecosystem and facilitating efficient AI model deployment on the company’s hardware.
“AI is our primary focus, expected to drive silicon demand in the next decade. Incorporating the expertise of the Mipsology team enhances our software strength, further enabling global customers to harness AI’s immense potential. The team will bolster our comprehensive AI software suite, enhancing our open array of tools, libraries, and models for effortless AI model deployment on AMD hardware,” stated Vamsi Boppana, senior Vice President of AMD’s AI division.