
The demand for AI computing has significantly impacted the supply and price of advanced semiconductor chips from NVIDIA and AMD. As the surge in AI computing has outpaced the supply of AI-capable chips, NVIDIA and AMD have experienced delays in meeting the high demand. This has resulted in their chips becoming very expensive and difficult to secure.
The situation has prompted major players like Microsoft, Google, and AWS to ramp up their production of custom silicon chips to reduce their dependence on NVIDIA and AMD. These companies are investing in developing their own AI chips to deliver AI compute services to customers globally.
The increasing demand for AI-based applications and services is expected to continue driving the growth of the AI chip market. As a result, buyers are recommended to delve deeply into the application areas of AI technology and expand their product lines to meet market demands and achieve business growth5. Establishing close cooperation with suppliers is also suggested to ensure timely access to high-quality, stable semiconductor products.

The need for high-performance custom silicon chips is being driven by compute-intensive tasks such as training AI large language models, which require massive amounts of computing power. These tasks are essential in various applications, including natural language processing, image and speech recognition, and predictive analytics. The dominant hyperscale cloud vendors, including Google, Microsoft, and AWS, are accelerating the production of custom silicon chips to meet the increasing demand for AI computing and to reduce dependence on dominant suppliers like NVIDIA and AMD.

The main reasons global giants like Microsoft, Google, and AWS are increasing their production of custom silicon chips are:
Reducing dependence on dominant suppliers: By producing their own custom silicon chips, these companies can reduce their reliance on the major suppliers of GPUs, such as NVIDIA and AMD.
Meeting the demand for AI computing: The surge in AI computing has led to a shortage of AI-capable chips. By ramping up custom silicon production, these companies can meet the increasing demand for AI compute services.
Enhancing AI capabilities: Custom silicon chips can be optimized for specific tasks, such as AI and generative AI tasks. This allows companies like Microsoft and Google to enhance their AI capabilities and offer better services to their customers.
Cost optimization: Custom silicon options could help organizations meet sustainability goals and reduce overall cloud costs by providing more hardware options and allowing customers to more discerningly choose appropriate compute.
Improved performance and latency results: The competition provided by cloud providers, in tandem with chip suppliers, could drive advances in chip performance, whether that is in the high-performance computing category for AI model training, or innovation for inferencing, where latency is a big consideration.