Artificial intelligence (AI) investing is still the market's most prevalent theme. As a result, investors need to pay attention to what's going on and see what stocks are the best buys during any given month. On the flip side, they also need to be aware of stocks that may have gotten too hot and should be trimmed or avoided.

For February, I have three stocks on my list that look great to buy and one that doesn't quite make the cut.

AI tailwinds are still strong for Nvidia, Meta Platforms, and Alphabet

When it comes to the companies to buy, Nvidia (NVDA -1.02%), Meta Platforms (META -1.39%), and Alphabet (GOOG -1.34%) (GOOGL -1.37%) top the list.

Nvidia is still a buy because the language from some of its largest customers indicates that 2025 will be an amazing year for it. The company makes GPUs (graphics processing units), which are the muscle behind most of the developed AI models. They are also used for inference, which occurs when someone uses a pre-trained AI model.

Alphabet, Meta Platforms, Microsoft, and Amazon have all told investors to expect massive capital expenditures this year to meet the growing AI demand. Each company also has some in-house computing chips that compete with Nvidia, but they still heavily use GPUs, making this stock one that investors want to own.

Meta Platforms is one of the biggest spenders in the AI race, but it has the cash flows to do it. It derives most of its revenue from its social media business, which includes properties like Facebook and Instagram. In the fourth quarter alone, this division produced $46.8 billion in revenue and converted an incredible 61% of that into operating profit.

This allows Meta to use those cash flows to build out its AI computing capacity, and it sounds close to a game-changing breakthrough. CEO Mark Zuckerberg thinks 2025 will be the year when it becomes possible to build an engineering AI agency with the same problem-solving skills as a good mid-level engineer. That breakthrough would be massive and further amplify the company's engineering resources.

Meta is an AI powerhouse, and its stock is one of the best on the market.

Alphabet is similar to Meta in that it is primarily an advertising business with AI on the side. It integrated its AI module, Gemini, into many of its advertising tools, which has helped the company stay relevant even if its generative AI model isn't viewed as best-in-class.

However, plenty of people are still using it through its Google Cloud platform, which allows clients to rent computing power. Because many companies can't afford to build their own servers, this has become an attractive option, leading to strong growth for Alphabet, with cloud revenue rising 30% in the fourth quarter.

I'm not buying Broadcom

Broadcom (AVGO -0.62%) is one AI stock I'm avoiding in February, because it just doesn't belong in the same class as these other three.

It has significant AI tailwinds with its custom AI accelerator and connectivity product lines, but they don't contribute enough revenue to offset some of the weakness in its legacy business. Furthermore, its massive VMware acquisition is still struggling to find its footing, as there are reports of it increasing its prices by more than 1,000%.

As a result, I'm not as high on Broadcom as I am on the other three. And it trades at a premium to each of these stocks despite not having growth that's as strong.

On a forward price-to-earnings (P/E) basis, Broadcom is the most expensive stock of the four.

NVDA PE Ratio (Forward) Chart

NVDA PE ratio (forward), data by YCharts.

This could be justified if the company were growing at warp speed or had another catalyst coming up. But that's not the case.

Each of these four companies operates on a different fiscal calendar, so direct comparisons are a bit difficult. But to get an idea of the growth each company will experience in the coming quarter, we'll use the revenue projections for the next quarter.

NVDA Revenue Growth Estimate for Current Quarter Chart

NVDA revenue growth estimate for current quarter, data by YCharts.

While Broadcom is growing faster than Meta or Alphabet, it's not so much faster that it deserves a massive premium over these two stocks. Furthermore, Nvidia is growing far faster than Broadcom yet trades at a discount.

Broadcom may be a good AI stock in the future, but it's not right now. As a result, I'm looking at Nvidia, Meta Platforms, and Alphabet in February, since they represent better values.