Broadcom (NASDAQ:AVGO) was late to the artificial intelligence game, but it is making up for lost time. Despite being a premiere chipmaker for smartphones, it was Broadcom’s transition to data center chips that cemented its position in the AI market.

Although Nvidia (NASDAQ:NVDA) continues to grab the headlines for its graphics processing unit (GPU) chips to train and build generative AI models, it will be Broadcom that soon takes the pole position as artificial intelligence readies itself for the next phase of growth.

24/7 Wall St. Insights:

• Generative artificial intelligence models are rapidly maturing and will soon transition from their training phase to the next stage of growth, called inference.
• Broadcom (AVGO) has already pivoted away from simple mobile handset chips to data center infrastructure reliant upon AI chips and is setting the groundwork to lead this new market forward.
• If you’re looking for some stocks with huge potential, make sure to grab a free copy of our brand-new “The Next NVIDIA” report. It features a software stock we’re confident has 10X potential.

The diverging path for AI’s growth

AI is not a one size-fits-all solution. There are two sides to the coin. On the one hand are the training models that Nvidia’s chips excel at. They are the first stage for an AI model where it learns from the inputs received to offer a desired output. 

Think of the ChatGPT chatbot where you ask a question and it gives you an answer. That model did not know beforehand what the answer would be, but after a long period of having data fed into it from multiple sources, it can now give a reasonable response.

The second stage of AI models is inference, where it can determine a response even though it hasn’t seen that data before. It is able to use predictive analytics to determine an outcome. Whether it is a self-driving car recognizing and reacting to data such as road signs, traffic patterns, or weather conditions; analyzing medical imaging to diagnose a condition; or optimizing inventory management to determine potential product demand peaks, these are examples of AI models inferring an outcome based on new, incoming information.

Moreover, where training is a very intensive compute process requiring substantial amounts of data, inference is fairly lean, requiring far less compute for each run and use case. It will also have significantly lower cost requirements to be economical.

This is the next phase for generative AI models that literally take off their training wheels and is where Broadcom will take the lead.

Staking out the early lead in the future of AI

ChatGPT creator OpenAI is working with Broadcom to specifically develop a new AI chip to allow models to run and respond to user requests after they have been trained, according to Bloomberg. This is the inference process, and together they are working with Taiwan Semiconductor Manufacturing (NYSE:TSM), the world’s largest third-party chipmaker to bring these chips forward.

As more businesses adopt artificial intelligence and begin to demand more complex usage for the models deployed, the need for inference chips will explode. Workloads are just going to get more complex, which will take it beyond the inference tasks that Nvidia’s GPUs are already able to handle.

According to billionaire investor Chamath Palihapitiya, there is a new capital expenditure cycle that will be pouring into the market. Earlier this year he told the All-In Podcast, “AI is really two markets, training and inference. Inference is going to be 100 times bigger than training. Nvidia is really good at training but very miscast at inference.”

Similarly, the industry analysts at McKinsey forecast the gen AI workload will transition to one that is primarily inference, “which favors specialized hardware due to lower cost, higher energy efficiency, and faster or better performance for highly specialized tasks.”

At least one estimate expects the inference chip market to grow at a 22.6% compound annual growth rate through 2030 and exceed $90 billion.

AVGO poised to become the premiere AI chip stock

This is Broadcom’s big opening, and with OpenAi and Taiwan Semiconductor, the trio are looking to introduce this new chip in 2026.

Broadcom is already seeing substantial growth in its AI chip business and it is accelerating. In its third-quarter earning conference call, president and CEO Hock Tan told analysts it now expects AI revenue to hit $12 billion in fiscal 2024, up from its previous guidance of $11.3 billion.

As the inference chip market starts to unfold, Broadcom’s early lead in the space will propel AVGO stock to new heights, much the way Nvidia staking out the early ground in AI training GPUs led to massive gains.

It makes Broadcom the chip stock to buy today for AI’s next stage of evolution.

 

Thank you for reading! Have some feedback for us?
Contact the 24/7 Wall St. editorial team.