Nvidia, Arm, or Broadcom are the best AI semiconductor stocks to buy now, according to Wall Street

Artificial intelligence (AI) will eventually reshape most industries, either by eliminating the need for human workers or by making them more productive. Analysts generally agree that the transformation will happen in phases. They define and enumerate those phases differently, but infrastructure companies will always be the first to benefit.

Powerful semiconductor and networking hardware are essential for training machine learning models and running inference in AI applications. As a result, chipmaker Nvidia (NVDA -2.25%)Arm Holdings (arm -4.96%)Broadcom (AVGO -4.05%) Some companies are already seeing benefits from AI adoption. Nvidia shares are up 160% this year, Arm shares are up 80%, and Broadcom shares are up 49%. Meanwhile, the S&P 500 index is up 100%. (^GSPC -0.32%) It increased by 18%.

But Wall Street expects the leaderboard to look different over the next 12 months. Below are the median price targets (and expected upside) for the three stocks as of August 25th.

Nvidia: $144 per share (expected to rise 11%) Arm: $135 per share (expected to fall 1%) Broadcom: $196 per share (expected to rise 18%)

Wall Street is actually expecting a decline in Arm shares. While analysts expect the other two semiconductor stocks to generate positive returns, the median price target for Broadcom suggests a bigger upside.

1. NVIDIA

Nvidia’s graphics processing units (GPUs) are the industry standard for accelerating complex data center workloads like scientific computing and artificial intelligence (AI) applications. According to the Wall Street Journal, “Nvidia’s chips power every cutting-edge AI system, and the company’s market share is estimated at more than 80%.”

Nvidia has opportunities to make money beyond GPUs. The company recently unveiled its first data center server chip, called Grace, and CEO Jensen Huang has said it’s gearing up for a multi-billion-dollar product line. Nvidia also sells InfiniBand and Ethernet networking platforms; that division of the company recently achieved annual revenue rates of more than $12 billion.

Finally, Nvidia has further expanded its ability to monetize AI by diving deeper into software and services. Nvidia AI Enterprise is a subscription software suite that streamlines the development and deployment of AI applications, and DGX Cloud brings together Nvidia hardware and software to create an end-to-end AI-as-a-service offering. Overall, Nvidia’s software and services division recently achieved a revenue run rate of $1 billion.

Going forward, Wall Street expects the company’s non-GAAP earnings to grow 39% annually through fiscal 2027 (the end of January 2027). Compared to that forecast, the current valuation of 71.7 times adjusted earnings places the stock somewhere between cheap and expensive.

2. ARM Holdings

Arm designs central processing unit (CPU) architectures and development tools and licenses its intellectual property (IP) to customers. The company’s customers include the world’s five largest companies. Arm IP underpins Apple’s private cloud computing, Nvidia’s Grace CPUs, Microsoft’s Cobalt chips, Alphabet’s Axion CPUs and Amazon’s Graviton processors.

This lineup is particularly interesting because these chips are used in data centers, a market historically dominated by Intel and AMD. Specifically, Arm chips have traditionally been limited to mobile devices, as Intel and AMD’s x86 processors have offered superior compute performance. Companies on both architectural sides have tried to take market share from the other, but Arm has had more success.

While the company maintains its dominance in smartphones and other mobile devices, it has been gaining market share in data centers by boosting the processing power of its chips, especially as it relates to AI. That means Arm is well-positioned to monetize AI in two areas: in personal devices like Apple’s iPhone and Microsoft’s Copilot+ PC, and in the infrastructure and platform services offered by public cloud providers like Amazon and Alphabet.

The only problem with Arm is valuation: Wall Street expects the company to grow adjusted earnings at 25% annually through fiscal 2026 (the end of March 2026). At that consensus estimate, its current valuation of 96 times adjusted earnings looks prohibitively expensive.

3. Broadcom

Broadcom offers IT solutions across two broad categories: infrastructure software and semiconductors. Specifically, the company sells virtualization, mainframe, and cybersecurity software that helps businesses manage, optimize, and secure their IT infrastructure. Of the three software verticals, virtualization is central to Broadcom’s growth story. The company’s subsidiary, VMware, is a leader in server virtualization, a market expected to grow 18% annually through 2032.

But the biggest opportunity for Broadcom is in semiconductor solutions, because that sector will benefit from the demand for artificial intelligence infrastructure. Broadcom is the market leader in data center networking chips and application-specific integrated circuits (ASICs). As Argus analyst Jim Kelleher recently wrote, “Networking is the business most impacted by AI, and Broadcom has long-established market leadership in enterprise and cloud data centers.” But that may change in the next few years.

Broadcom is also a market leader in high-end ASICs, custom silicon made for specialized applications like artificial intelligence. The company has already designed AI accelerators for Google and MetaPlatform, and recently landed its third big customer. Broadcom announced the deal in March without disclosing the name of the customer, but Reuters reported it was TikTok’s parent company ByteDance.

Currently, custom AI accelerators account for less than 10% of AI processors, as the market is dominated by GPUs (Nvidia GPUs, to be exact). But Morgan Stanley analysts believe that sales of custom AI chips will grow faster than GPU sales over the next few years, with ASICs accounting for up to 30% of AI chips by 2027. In that scenario, Broadcom, which controls 55% to 60% of the market, would likely be the biggest winner, according to the analysts.

Wall Street expects the company’s profits to grow 24% annually through 2025, which makes its current valuation at 38 times earnings seem reasonable, making now a good time for patient investors to buy a small stake in this semiconductor stock.