The recent market sell-off has created some nice opportunities in the market for long-term investors, particularly among artificial intelligence (AI) stocks. This includes several AI semiconductor stocks besides Nvidia, which has also seen its stock come under pressure.

Let's look at two AI chip stocks that investors can look to buy ahead of the next AI stock bull run.

Advanced Micro Devices

Advanced Micro Devices (AMD -8.57%) shares are down nearly 46% over the past year. However, that doesn't mean the company is performing poorly. In fact, AMD has seen strong AI-related growth over the past year, with its data center revenue soaring 94% to $12.6 billion. Data center growth decelerated a bit in its most recent reported quarter but still grew a robust 69% year over year, while overall revenue climbed 24%.

While it's the distant No. 2 player in the graphic processing unit (GPU) space, the company is still riding the tide in AI infrastructure spending higher. AMD's chips continue to be used more for inference rather than training, which could be a good position to be in for the future. There should eventually be a shift toward inference as AI models mature, and there is a need for more real-time, rapid AI inference for things like cybersecurity detection, autonomous driving, and even personal shopping recommendations. Given that its chips are cheaper than those of Nvidia, it could eventually take some share in this market.

Meanwhile, one place where the company currently shines is in the central processing unit (CPU) space. While GPUs provide the processing power, CPUs act more like the brain. While the market for data center CPUs isn't nearly as large as it is for GPUs, AMD has become the leader in the space and is gaining share. Last quarter, it said among hyperscalers (companies with massive data centers) that it had well above a 50% market share with its CPUs. It noted strong CPU gains within personal computers (PCs) as well. Management said AMD has a 70% market share on popular platforms such as Amazon, Newegg, and MindFactory.

Management admits AMD's gaming segment has been weak, as there has not been a new gaming console launched in a few years. However, there is speculation a new Xbox could be released in 2027 and a new Playstation in 2028, which should lead to a rebound in this segment in a few years.

Trading at a forward price-to-earnings ratio (P/E) of 22 times 2025 analyst estimates, AMD's stock is attractively valued, given its growth and the opportunities still in front of it.

Image source: Getty Images

Broadcom

Broadcom (AVGO -4.97%) is another semiconductor company that has seen its stock take a hit, with its shares trading down roughly 32% so far in 2025. However, it has perhaps one of the biggest opportunities of all chip stocks ahead.

Broadcom is making a name for itself as the go-to company to help customers design their own custom AI chips called ASICs (application-specific integrated circuits). While there are more initial upfront costs, and it takes considerable time to design them, custom AI chips perform better at the specific tasks for which they were designed and use less power than off-the-shelf GPUs. These chips are not as flexible in their uses as GPUs, but given the upfront costs, customers are not designing them for small projects.

Alphabet was the first company to employ Broadcom in helping it design its custom Tensor Processing Units (TPUs), which are tightly integrated with Google Cloud's TensorFlow, its open-source library for machine learning. This tight integration has helped improve performance and lower power consumption costs while running Google AI workloads.

Since then, Broadcom has been gaining more new custom AI customers. Including Alphabet, the company has three custom AI chip customers that are further along in their development. It believes these three customers represent a $60 billion to $90 billion serviceable market opportunity in its fiscal year 2026 (ending October 2026) alone. While some of this opportunity will go to GPUs, this is still a huge potential growth driver for the company.

Meanwhile, it has been adding more AI chip customers recently, including Apple. It took about 15 months for Alphabet's chips to be designed and then deployed, which was considered very fast. So, this should lead to additional growth in a few years.

In addition to custom chips, Broadcom also participates in the AI infrastructure buildout through its portfolio of networking technology components. These components, such as switches and network interface cards (NICs), help manage data flow and network communication. As AI clusters grow in size, this is becoming a more important part of the AI infrastructure puzzle.

While its more commoditized chip businesses have been a drag, the company also owns an attractive software piece with VMware. It is driving growth by transitioning customers from perpetual licenses to subscriptions while also upselling them to its full VMware Cloud Foundation stack, which can virtualize entire data centers.

At a forward P/E of under 26 times, Broadcom is an attractive stock to buy, given the big catalysts in front of the stock, especially with its custom AI chips.