Looking back at AI stock forecasts from 2025, there's one prediction that deserves a closer look -- and it's about a company that's been central to the entire AI boom.

Nvidia caught a lot of attention last year as analysts were positioning it to dominate the AI stock landscape. The reasoning was solid: while the broader AI sector had taken some lumps due to tariff concerns and economic uncertainty, Nvidia showed it could adapt. When export controls hit, they pivoted. When tariff talk emerged, they announced U.S. manufacturing investments. That kind of resourcefulness matters, especially in volatile markets.

What made the case compelling was the company's position with its biggest customers. Microsoft, Amazon, and Meta all doubled down on their AI spending plans, signaling that the infrastructure buildout was far from over. These weren't tentative commitments either -- they were concrete investments in the strongest possible AI systems. And guess where most of that hardware was flowing? Nvidia's way.

The financial picture supported the optimism too. With $53 billion in cash, Nvidia had the firepower to handle whatever came next. More importantly, the company had shifted focus toward inference -- the computational work that powers AI reasoning models. Their new Blackwell architecture was purpose-built for this exact use case. That positioning looked like it could unlock the next wave of growth.

Of course, China remained the elephant in the room. The $4.5 billion charge from restricted H20 chip sales was a real hit, and uncertainty around future export policy created genuine risk. But even accounting for that, the AI stock forecast for 2025 seemed reasonable: Nvidia had the technology, the customers, the cash, and the market tailwinds.

Valuation-wise, things looked interesting too. Trading around 32 times forward earnings -- down from 50 just months earlier -- the stock didn't look wildly expensive for a company sitting at the center of the AI infrastructure boom.

The broader AI stock market itself was expected to expand massively, from roughly $300 billion to potentially $2 trillion within the decade. Nvidia's dominance in GPU supply meant it would likely capture a meaningful slice of that growth. Whether you agreed with every detail of the analysis or not, the core logic was hard to dismiss: this was still early in the AI cycle, and companies positioned at the foundation of that infrastructure had significant runway ahead.