Microsoft has unveiled two custom-designed AI chips, the Azure Maia 100 AI Accelerator and the Azure Cobalt 100 CPU, to address the shortage of GPUs used in AI model development. The scarcity of GPUs, including Nvidia's AI cards, has created challenges for companies in the AI space. Maia 100, a 5-nanometer chip with 105 billion transistors, is designed to power large AI workloads in Microsoft Azure, while Cobalt 100 is an energy-efficient, 128-core chip optimized for cloud-native offerings. These chips will be rolled out to Azure data centers, initially supporting Microsoft AI services and OpenAI models, with second-generation hardware already in development.

‍

Microsoft's move into designing its own AI chips aims to reduce its dependence on GPU vendors and remain cost-competitive in the AI market. GPU shortages and high costs have hindered progress for both Microsoft and OpenAI, making in-house hardware development a strategic move. However, the success of Microsoft's AI chips remains uncertain, as seen with challenges faced by other tech giants like Meta and Google in their custom AI chip efforts.