Last month, Nvidia (NASDAQ: NVDA) shares reached a record $974 on March 8. This red-hot stock has been cooling since then, so investors haven't had much to shout about.
Recently, Nvidia stock has fallen 11% from its record high. Wall Street fears the stock's bubble may have burst and its AI-powered growth may have stalled due to AI chip competition. Nvidia stock fell 2% on April 9 after Intel's (NASDAQ: INTC) announcement of a new AI processor that targeted Nvidia's top performers. Nvidia investors may be concerned about Intel's recent move to enter the fast-growing AI chip industry.
Intel challenges Nvidia's AI dominance Intel has struggled to enter the AI chip industry. Intel's AI chip revenue pipeline was "well above $2 billion and growing," according to CEO Pat Gelsinger, in 2023. That pales in comparison to Nvidia's fiscal 2024 data center revenue of $47.5 billion, showing it leads Intel in the AI processor market.
Chipzilla's latest Gaudi 3 accelerator challenges Nvidia's AI dominance. The launch press release stated that this new Intel processor is "delivering 50% on average better inference and 40% on average better power efficiency than Nvidia H100 -- at a fraction of the cost."
The press statement added that Gaudi 3 can train the Llama2 large language model (LLM) with 7 billion and 13 billion parameters and the 175 billion parameter GPT-3 model in 50% less time. Dell Technologies, Super Micro Computer, Lenovo, and Hewlett Packard Enterprise are also anticipated to provide systems based on the Gaudi accelerator and server processor, according to Intel.
Intel also believes Gaudi 3 might match Nvidia's forthcoming H200 AI GPU. Since Intel's latest AI processor uses Taiwan Semiconductor Manufacturing's 5-nanometer (nm) technology, it's not surprising. Nvidia's flagship, the H100, uses TSMC's 5nm technology. Nvidia's H200 CPU will also use TSMC's 5nm N5 technology. Investors may believe that Intel's aggressive pricing and strong performance can challenge Nvidia's AI chip market dominance.
Investors should notice that Nvidia's H100 GPU was revealed two years ago and mass-produced in 2022. The third and fourth quarters of 2024 will see Intel's Gaudi 3 chip go into mass production. By the time Intel ships these chips, Nvidia will have improved.
Dislodging Nvidia is difficult. Nvidia launched the Blackwell GPU architecture last month, claiming it will let "organizations everywhere to build and run real-time generative AI on trillion-parameter large language models at up to 25x less cost and energy consumption than its predecessor."
The Blackwell B200 GPU should outperform the H100 by four times during AI training and 30 times during inference. Due to its 4nm production node, the B200's performance is far higher than Nvidia's prior GPUs.
Better still, Nvidia will aggressively price Blackwell processors. Nvidia may be able to offset Intel's cost advantage with its performance and efficiency advances. Amazon, Alphabet's Google, Meta Platforms, Microsoft, and Oracle are expected to utilize the B200 for these reasons. Later this year, Nvidia will release Blackwell-based CPUs, which could steal Intel's thunder and retain its AI chip market dominance. Analysts predict Nvidia's data center revenue will treble to $95 billion in 2024.
Nvidia leads the AI processor market with over 80% market share, thus the market expects it to continue. Intel's AI chip announcement may not be a threat to Nvidia's next generation of more powerful and efficient chips, therefore investors should ignore it.
stay turned for development