The Next AI Battle: Custom ASICs vs. Nvidia GPUs
Here is why $Meta Platforms, Inc.(META)$ and $Alphabet(GOOG)$ are working with $Broadcom(AVGO)$ to make custom AI ASICs.
$NVIDIA(NVDA)$ GPUs are powerful and flexible.
Built, they contain parts that AI models do not use, making them highly expensive and inefficient when used in certain ways.
ASICs are custom chips designed to do only one specific job, like running a specific type of AI model extremely fast.
An ASIC is like a specialized tool built for a specific task, while a GPU is like a multi-tool pocketknife.
The ASIC is much more energy-efficient and cost-effective once a company is running millions of specific tasks a day.
Simply put, NVDA GPUs are good at running all AI, while GOOGL TPUs are PERFECT for a very specific AI use case, such as Gemini, or Agentic AI.
Disclaimer: Investing carries risk. This is not financial advice. The above content should not be regarded as an offer, recommendation, or solicitation on acquiring or disposing of any financial products, any associated discussions, comments, or posts by author or other users should not be considered as such either. It is solely for general information purpose only, which does not consider your own investment objectives, financial situations or needs. TTM assumes no responsibility or warranty for the accuracy and completeness of the information, investors should do their own research and may seek professional advice before investing.
