35% of Warren Buffett's $296 Billion Portfolio Is Invested in 3 Artificial Intelligence (AI) Stocks

In This Article:

Since Warren Buffett became the CEO of Berkshire Hathaway (NYSE: BRK.A)(NYSE: BRK.B) in 1965, its stock has delivered a compound annual return of 19.8%. That would have been enough to turn a $1,000 investment into $42.5 million, whereas the same investment in the S&P 500 would have grown to just $325,053 over the same period.

Buffett's long-term investing strategy is simple. He buys into companies with steady growth, robust profitability, strong management teams, and shareholder-friendly initiatives like stock buyback programs and dividend schemes, which help to compound his returns over time. You won't ever see Buffett chasing the latest stock market trend -- not even one as powerful as artificial intelligence (AI).

With that said, three of the existing stocks in Berkshire's $296 billion portfolio of publicly traded securities are integrating AI into their legacy businesses in very unique ways.

Warren Buffett looking at the camera.
Image source: The Motley Fool.

1. Amazon: 0.8% of Berkshire Hathaway's portfolio

Amazon (NASDAQ: AMZN) is the world's largest e-commerce company. It's using AI in its fulfillment centers to drive efficiency, and it even created an AI shopping assistant called Rufus, which can help customers find the right products. However, most investors are focused on Amazon's cloud computing platform called Amazon Web Services (AWS), because it's trying to dominate the three core layers of AI:

  • Infrastructure: AWS operates data centers fitted with Nvidia's industry-leading chips, but it also designed its own chips called Trainium (for AI training) and Inferentia (for AI inference). Trainium can save developers up to 50% on AI training costs compared to competing chips, and demand is so high that Amazon is producing more of them than it expected.

  • Large language models (LLMs): Developers can access leading third-party models from companies like Anthropic and Meta Platforms on the AWS Bedrock platform to accelerate their AI software projects. Bedrock is also home to a family of LLMs called Titan, which Amazon built in-house.

  • Software: Amazon's new Q virtual assistant is embedded in AWS to help businesses extract valuable insights from their internal data. It can also generate computer code on demand to help developers build software.

During the third quarter of 2024, AWS revenue came in at $27.4 billion, which was a 19% increase from the year-ago period. AI revenue within AWS, however, grew by a triple-digit percentage, and Amazon says this business is growing 3 times faster than the cloud business did at the same stage of its life cycle.