The Nasdaq entered correction territory earlier this week, and a number of leading artificial intelligence (AI) semiconductor stocks have been swept up in the market downturn. However, spending on AI infrastructure has not suddenly dried up, and in fact it is still on the rise.

The three big cloud computing companies, for example, have budgeted spending a combined $250 billion in capital expenditures (capex) this year largely related to AI infrastructure. Meanwhile, a group of companies led by OpenAI and Softbank have pledged spending $500 billion over the next few years to go toward building AI data centers through Project Stargate. At the same time, AI start-ups and other leading tech companies are also building out AI infrastructure, including Meta Platforms, which plans to spend up to $65 billion on AI infrastructure this year. That's a lot of spending that will benefit AI chip companies this year and beyond.

Let's look a three AI chip companies set to benefit that are worth buying during the current Nasdaq correction.

Nvidia

Nvidia (NVDA 1.66%) has established itself as the leading AI chipmaker through its market-leading graphics processing units (GPUs). It's been able to establish an approximately 90% market share with GPUs in large part thanks to its CUDA software platform, which was developed to allow its chips to be programmed beyond their original task of speeding up graphics rendering in video games.

Nvidia's revenue growth exploded when AI started to become mainstream due to the fast processing times of its chips, which were used to help train AI models and run inference. As AI models advanced, more and more GPUs were required to provide the needed computing power. Meanwhile, the company continued to further increase its software lead by creating a collection of libraries, microservices, and tools designed specifically for AI and high-performance computing. Today, its chips are the backbone of AI infrastructure.

Following the recent sell-off, the stock is inexpensive, trading at a forward price-to-earnings (P/E) ratio of under 24 times 2025 analysts' estimates and a price/earnings-to-growth (PEG) of below 0.5, with PEGs below 1 typically considered undervalued.

Image source: Getty Images.

Broadcom

While Nvidia is the leader in mass-merchant AI chips, Broadcom (AVGO 3.06%) has been carving out a strong niche in custom AI chips. It helps customers design application-specific integrated circuits, or ASICs. These custom chips are used for very specific tasks and as such have better performance at these tasks while using less power. However, they lack the flexibility of GPUs.

After helping Alphabet develop its custom tensor-processing unit (TPU) called Trillium, Broadcom has been gaining increasing interest from new customers. It now has three established customers, who it says represent a $60 billion to $90 billion serviceable addressable market in its fiscal 2026. While Nvidia will likely get its fair share of this AI chip spending, this represents a huge opportunity for Broadcom.

Meanwhile, the company also now has four newer custom AI chip customers, including Apple. It took about 15 months for Alphabet's custom chip to go from development to being deployed, which was considered fast. As such, it likely will take a year and half to two years before these new customers can provide some meaningful revenue. Nonetheless, Broadcom is seeing a lot of AI chip momentum.

Trading around 28.5 times fiscal 2025 analyst estimates, the stock is attractively priced given the big opportunity in front of it.

Advanced Micro Devices

While Advanced Micro Devices (AMD 0.14%) is the No. 2 player in the GPU market behind Nvidia with about 10% market share, what the company has done a good job at is gaining share in the central processing unit (CPU) market within the data center. While GPUs are known for providing the power, CPUs act more as the brain of the operations. The CPU market for data centers isn't as big as the market for GPUs, but it is still expanding nicely as the AI infrastructure buildout continues.

Last quarter, AMD said market share for its EPYC CPUs was well above 50% among hyperscalers, which are companies with massive data centers. The company's CPUs have also been gaining share in the personal computer (PC) retail space, as well. It said it had an over -0% market share on platforms such as Amazon, Newegg, and MindFactory.

Meanwhile, the company is still seeing growth in the GPU market. It noted that Microsoft and Meta Platforms are both using its MI300X GPUs, while it has seen strong interest in its next-generation MI350 series GPUs. Meanwhile, it plans to launch its MI400 GPUs in 2026. Currently, there is a pretty big gap between Nvidia's and AMD's software, which will likely keep it a distant second in the GPU market; however, its chips are finding a nice niche on the inference side of the market, which is nicely growing.

With a forward P/E of only 15, AMD's stock is inexpensive. Meanwhile, it's doing well in the data center with its CPU chips, while the overall tide in AI spending should help its GPU revenue as well, making the company a solid option to consider buying at these levels.