Nvidia (NVDA +1.23%) and Microsoft (MSFT +2.05%) have been two of the market's hottest artificial intelligence (AI) stocks. Nvidia, the world's leading producer of data center GPUs, sells the best picks and shovels for training AI algorithms. Microsoft, the world's largest software company, owns a major stake in ChatGPT's creator OpenAI. It's also been integrating its Copilot generative AI services into most of its cloud-based services and applications.
From the beginning of 2020 to the end of 2025, Nvidia's stock skyrocketed 3,070% as Microsoft's stock rallied 207%. But this year, Nvidia's stock has risen only 8%, while Microsoft's has declined nearly 60%. Both stocks lost their luster amid concerns about inflation, the Fed's reluctance to cut interest rates, and the escalating Middle East conflict. Those headwinds drove investors toward more conservative investments and chilled the AI market's multi-year rally.
Image source: Getty Images.
However, I believe that pullback represents a good buying opportunity for long-term investors who can tune out the near-term noise. Here's why I'm still willing to buy more AI stocks.
The AI market will continue to evolve
Over the past few years, most companies have focused on training AI algorithms to crunch massive amounts of data across large GPU clusters in data centers. The leaders in the training market include OpenAI's GPT, Anthropic's Claude, Meta Platforms' Llama, and Alphabet's Google Gemini.
Those tech giants are investing billions of dollars in their data centers to train those large language models (LLMs), but that spending is lumpy and cyclical. For long-term stability, more investors are pivoting toward the inference market -- which focuses on using software applications to access that trained data. Simply put, "inference" is the act of asking OpenAI's ChatGPT or Google Gemini a question, or using AI software to create a new image or video.
NASDAQ: AVGO
Key Data Points
Today, many AI-oriented companies spend more on inference than on training. That's great news for Broadcom (AVGO +5.06%), which develops custom application-specific integrated circuits (ASICs) to accelerate AI inference tasks for hyperscalers. They're also buying more data center CPUs from Intel (INTC 1.51%) -- which was left behind when companies focused on buying more GPUs from Nvidia for their training tasks -- to support those inference tasks.
As those hyperscalers expand their data center infrastructure, they'll ramp up spending on optical hardware from companies like Lumentum and Corning. They'll also lease more rack space from data center real estate investment trusts (REITs) like Digital Realty and strike big renewable energy deals with companies like Brookfield Renewable. They'll also outsource some of their training and inference tasks to cloud-based AI infrastructure provides like CoreWeave and Nebius.
In other words, there are dozens of ways to invest in the AI market's secular expansion. While some conversations might focus on Magnificent Seven companies like Nvidia and Microsoft, all of these companies still have plenty of room to expand without trampling each other.
The AI market will continue to expand
The global AI market could still expand at a 30.6% CAGR from 2026 to 2033, according to Grand View Research, driven by the rapid enterprise adoption of generative AI and agentic AI solutions. A recession could actually drive the market's acceleration as more companies use AI tools to streamline their businesses and replace more of their human workers with AI agents.
If you expect that to happen, it doesn't make sense to shun AI stocks today simply because oil prices are high, the macro headwinds are messy, and the Fed isn't cutting its benchmark rates. Instead, it makes more sense to nibble on the top AI stocks even as they cool off this year. Their near-term gains might be limited, but the AI sector still has plenty of room to run and will reward its most patient investors.
