Many investors see Nvidia (NVDA 2.63%) as the quintessential artificial intelligence (AI) stock because its chips provide the computing horsepower needed to train the most advanced AI systems, such as OpenAI's ChatGPT and Tesla's full self-driving software.

However, some hedge fund billionaires sold Nvidia stock during the first quarter, while purchasing shares of Palantir Technologies (PLTR 1.49%) and/or Super Micro Computer (SMCI -1.80%), two supercharged AI stocks with year-to-date returns of 59% and 198%, respectively.

  • Louis Bacon of Moore Capital Management sold 2,006 shares of Nvidia in the first quarter, reducing his stake by 19%. Meanwhile, Bacon started a small position in Super Micro Computer.
  • Israel Englander of Millennium Management sold 720,000 shares of Nvidia in the first quarter, reducing his stake by 35%. Meanwhile, Englander increased his positions in Palantir and Super Micro Computer by 4% and 235%, respectively.
  • Philippe Laffont of Coatue Management sold 2.9 million shares of Nvidia in the first quarter, reducing his stake by 68%. Meanwhile, Laffont increased his position in Palantir by 40%.

Trades made by Israel Englander are especially noteworthy because Millennium Management handily beat the S&P 500 over the last three years , and it ranks as the second-best performing hedge fund of all-time as measured by net gains since inception. Here's what investors should know about Palantir and Supermicro.

1. Palantir Technologies

Palantir specializes in data analytics. Its software lets government and commercial clients integrate data, develop artificial intelligence (AI) and machine learning (ML) models, and build applications that lean on those data sets and models to improve decision-making. Palantir recently introduced its Artificial Intelligence Platform (AIP), a product that brings support for large language models and generative AI to its existing software.

Certain industry analysts have praised the company for its technological prowess. Forrester Research ranked Palantir Foundry as the best AI/ML platform in a report published in July 2022. And Dresner Advisory Services ranked Palantir as a leader in the AI/ML and data science market in a report published in August 2023.

However, other analysts are skeptical. Rishi Jaluria of RBC Capital says conversations with industry observers and company employees have led to the conclusion that Palantir does not offer "anything truly differentiated when it comes to generative AI."

Palantir reported reasonably good financial results in the first quarter, beating estimates on the top line and meeting expectations on the bottom line. Its customer count increased 42% to 554, and the average existing customer spent 11% more. In turn, revenue increased 21% to $634 million, the third consecutive sequential acceleration, and non-GAAP earnings increased 60% to $0.08 per diluted share.

CFO Dave Glazer said the commercial segment benefited from "unprecedented demand driven by momentum from AIP." However, the stock still slipped 7% following the first-quarter report because management guided for full-year revenue growth of 20%, implying a slight deceleration in the coming quarters. Analysts anticipated full-year revenue growth of 22%.

Going forward, Wall Street expects Palantir to grow adjusted earnings per share at 22% annually through 2026. That consensus estimate makes its current valuation of 97 times earnings look very expensive. Investors should be cautious with this stock. Personally, I plan to avoid Palantir until earnings growth accelerates or the valuation improves.

2. Super Micro Computer

Super Micro Computer designs high-performance computing platforms for enterprise and cloud data centers. Its portfolio includes servers and storage systems, ranging from individual devices to full rack solutions. Its products can be optimized for use cases like artificial intelligence and 5G infrastructure, and they feature chips like Nvidia graphics processing units (GPUs) and Intel central processing units (CPUs).

Importantly, Supermicro is the market leader in AI servers due to manufacturing prowess and its building-block approach to product development. To elaborate, nearly half of its workforce are engineers, and it handles most research and development internally. "Our engineering aptitude, coupled with our internal manufacturing capability, enables rapid prototyping and product roll-out."

Additionally, Supermicro's modular product design reduces time to market, and affords clients flexibility in designing custom solutions. It can "quickly assemble a broad portfolio of solutions by leveraging common building blocks across product lines." In other words, Supermicro can rapidly integrate the latest CPUs, GPUs, and memory into preassembled server chassis, such that it often beats competitors to market by two to six months.

Indeed, Supermicro "anticipates being first-to-market in deploying full rack clusters featuring Nvidia Blackwell GPUs." That is advantageous because businesses are eager to purchase AI hardware, so they are turning to the server maker that brings computing products to market at the fastest clip. As a result, Supermicro's AI server market share is expected to reach 23% by the end of 2024, up from 10% at the beginning of the year.

Going forward, Wall Street expects Supermicro to grow earnings per share at 48% annually over the next three years. That consensus estimate makes its current valuation of 47 times earnings look very reasonable. Indeed, it gives a PEG ratio -- the price-to-earnings ratio divided by the forecasted earnings growth -- of roughly 1. For context, using the same methodology, Palantir currently has a PEG ratio of 4.4.