Microsoft recently revealed that its cloud computing customers could use AMD MI300X AI CPUs. Microsoft's Azure cloud computing service will provide MI300X chip clusters as an alternative to Nvidia's H100 GPUs. After this announcement, should Nvidia shares be sold for AMD??
Microsoft doesn't sell AMD GPUs to cloud clients to compete with Nvidia's GPUs, which power data center AI applications. Because Nvidia chips are so popular, companies are having trouble getting them. This issue isn't bad.
Both Nvidia and AMD are struggling to meet GPU demand. Taiwan Semiconductor Manufacturing (NYSE: TSM), the world's largest semiconductor producer, announced its advanced packaging capacity is full for this year and next. This is likely due to GPU demand from Nvidia and AMD, the company's top-10 clients.
To address AI and HPC chip demand, TSCM is quickly expanding production. After starting wafer production in Arizona, the business aims to establish a third fab. It finished its first specialty technology fab in Japan and announced a second for 2027. It will begin developing an automotive and industrial fab in Germany late this year.
TSMC also wants to perfect 2-nanometer chip technology. As semiconductor technology shrinks, more chips can fit on a wafer, increasing production capacity and lowering costs. Investors may choose to buy Nvidia, AMD, and TSMC instead of selling. The industry is struggling to meet increasing demand for high-performance computing CPUs and GPUs.
Nvidia is the frontrunner and has grown dramatically. The company's CUDA software framework, which allows GPUs to be programmed directly, made it the GPU industry standard before AI. The company can presumably sell as many chips as its foundry partners can make. Most semiconductor businesses employ contract manufacturers like TSCM instead of owning manufacturing facilities.
AMD benefits from a tight GPU market. With Q1 results, the business boosted its full-year data center GPU revenue forecast from $3.5 billion to $4.0 billion. AMD can become a number-two player in this market because Nvidia GPUs are scarce. If AMD's chips are well received, the market dynamics could offer it a longer-lasting boost since companies don't like to rely too heavily on one supplier.
As firms rush to power AI applications with GPUs and chips, TSCM benefits. TSCM will gain from the AI chip boom by increasing fabs and adopting 2-nanometer technology. It will profit as more companies join. Arm Holdings, Softbank, and Amazon have all been said to be designing AI chips. Apple executives reportedly met with TSMC to reserve 2-nanometer production to catch up in AI. All this helps TSMC.