Because while Buffett hasn’t bought into the AI gold rush, the landscape is shifting beneath his feet.
Warren Buffett’s Philosophy: Why Tech Has Always Been a Hard Pass
Buffett doesn’t hate technology. He just hates not understanding it. This isn’t about resistance to change—it’s about discipline. The Oracle of Omaha built his empire on the principle of investing within his “circle of competence,” a concept he’s repeated for decades. If he can’t forecast a company’s earnings over 10 years, he walks away. And that changes everything when you’re looking at deep-learning startups burning cash for market share. Take Amazon in the early 2000s—Buffett admitted he missed it not because he thought it would fail, but because he didn’t grasp the scalability of cloud infrastructure. He’s candid about blind spots, which earns him credibility even when he’s wrong.
Yet, people don’t think about this enough: his skepticism isn’t rooted in ignorance. It’s strategic. He watched the dot-com bubble burst. He saw companies with no earnings, no business models, and absurd valuations collapse overnight. And that’s exactly where AI stands today for many investors—teetering between transformation and hype. Buffett prefers predictability. He likes brick-and-mortar businesses with pricing power, durable brands, and steady cash flow. Think: See’s Candies, Geico, BNSF Railway. These aren’t flashy. They don’t trend on X (formerly Twitter). But they print money. Year after year.
Circle of Competence: The Bedrock of Berkshire’s Strategy
This principle isn’t just jargon—it’s a filter. Buffett and his partner Charlie Munger used it to avoid entire sectors during the tech surge of the late 1990s. They sat out most of the internet revolution, focusing instead on insurance, finance, and consumer staples. That restraint made them richer in the long run, as the crash wiped out trillions in paper gains. Their returns weren’t explosive in the short term, but compound growth over decades turned $10,000 in 1965 into over $40 million by 2023. That’s not luck. It’s consistency.
Recent Shifts: Has Buffett Softened on Technology?
Not exactly. But his team has. Since Munger’s passing in 2023 and Buffett stepping back from day-to-day decisions, portfolio managers like Todd Combs and Ted Weschler have taken larger roles. They’ve made moves Buffett never would have alone—like buying Apple stock in 2016. At first glance, Apple seems like a tech stock. But Buffett redefined it in his mind: a consumer brand with a loyal ecosystem, more like Coca-Cola with screens. That mental reframing allowed the purchase. Today, Apple makes up about 47% of Berkshire’s equity portfolio, valued at over $130 billion. It’s not an AI bet, but it’s a crack in the armor.
AI in the Portfolio—But Not Where You’d Expect It
So does Buffett own AI stocks? Not directly. But AI is creeping into his holdings like moisture into old brick. Take American Express. Berkshire owns over $20 billion in the company. Amex uses machine learning algorithms to detect fraud in real time—processing 70 million transactions daily across 110 currencies. Their system flags anomalies with 95% accuracy, saving billions. Is that "AI"? Technically, yes. But it’s not generative AI creating art or chatbots. It’s operational AI—quiet, unglamorous, and effective. And it’s happening across the portfolio.
Another example: BNSF Railway. One of the largest freight railroads in North America. They use predictive maintenance models to anticipate track failures and locomotive breakdowns. Sensors collect terabytes of data daily, feeding models that optimize fuel use, reduce delays, and improve safety. A single efficiency gain of 2% across their 32,500-mile network could save $150 million annually. This is artificial intelligence working behind the scenes, not in Silicon Valley boardrooms.
Financial Services and AI: The Quiet Revolution
Bank of America, in which Berkshire holds a stake, deploys AI through its virtual assistant Erica—used by over 25 million customers. It handles balance checks, payment scheduling, and credit counseling using natural language processing. JPMorgan Chase, another indirect link via Berkshire’s financial bets, uses COiN, an AI platform that reviews legal documents in seconds—work that once took 360,000 hours annually. These aren’t headline-grabbing investments, but they’re transformative. And they’re profitable.
Insurance and Risk Modeling: Where AI Fits Buffett’s World
Geico, wholly owned by Berkshire, uses AI-driven analytics to price policies. Telematics data from smartphones and dongles feed algorithms that adjust premiums based on driving behavior. Safe drivers pay less. Risky ones pay more. It’s a feedback loop that improves underwriting margins—the lifeblood of insurance profitability. Does Buffett brag about it? No. But the results show up in quarterly earnings. Geico’s combined ratio improved from 101.7 in 2021 to 97.3 in 2023. That’s a $1.2 billion swing in underwriting profit. AI didn’t do it alone, but it helped.
Buffett vs. the AI Hype: A Philosophical Divide
When asked about AI at the 2023 Berkshire annual meeting, Buffett had a telling response: “It’s going to change a lot of things. But I don’t know how much it will change our businesses.” That’s classic Buffett—measured, open-minded, yet cautious. He sees AI as a tool, not a savior. Contrast that with Nvidia CEO Jensen Huang, who called AI “the new industrial revolution.” Two perspectives. One grounded in cash flow, the other in exponential vision. Who’s right? Honestly, it is unclear. History favors innovators, but also punishes overreach.
Consider the 2024 run-up in AI stocks. Nvidia’s market cap surged from $500 billion in early 2023 to over $2.2 trillion by June 2024. Microsoft and Amazon gained over 40% in the same period. But how many of these gains are based on real earnings versus future potential? Nvidia’s trailing P/E ratio hit 75—astronomical compared to Berkshire’s core holdings, which average around 15. Buffett would call that speculative. And he’s not wrong. Yet, missing the trend altogether carries opportunity cost. Even he admitted, “The one thing I do know is that AI is going to change the world.”
Apple: The One Tech Bet That Rewrote the Rules
How does Apple fit into Buffett’s world? It’s the exception that proves the rule. He didn’t buy Apple for its AI research. He bought it for its brand loyalty, massive cash reserves, and ecosystem lock-in. But now, Apple is investing heavily in on-device AI. The A17 chip in the iPhone 15 supports real-time language translation, photo enhancement, and voice processing—all without sending data to the cloud. Their 2024 “Apple Intelligence” rollout integrates generative AI directly into iOS 18. Siri is getting a major upgrade. And developers can build AI features using Apple’s frameworks.
That said, Apple’s AI strategy is conservative. No race to release a ChatGPT rival. No data harvesting. Privacy-first design limits data availability, which constrains model training. But it might also be sustainable. While Meta and Google face regulatory scrutiny over data use, Apple’s approach could win long-term trust. And Berkshire’s $130 billion stake rides that wave—whether Buffett intended to bet on AI or not.
AI Alternatives: Where Buffett Might Look Next
If Buffett won’t buy AI startups, where might he go? One possibility: infrastructure. Not semiconductors—too cyclical. But data centers? Utilities? Fiber networks? These are the “picks and shovels” of the AI era. Berkshire owns MidAmerican Energy, which runs data centers for tech clients. They’re expanding renewable-powered facilities in Iowa and Utah. A 100-megawatt solar farm powers part of their Des Moines campus. Could this scale? Possibly. Because AI demands power—lots of it. One ChatGPT query uses 10 times more energy than a Google search. Training GPT-3 consumed 1,287 megawatt-hours—enough to power 120 homes for a year. Energy demand is becoming a bottleneck. And that might be Buffett’s kind of opportunity: boring, essential, and cash-flow positive.
Utilities vs. Semiconductors: A Tale of Two Investments
Compare Nvidia’s volatility to NextEra Energy’s steady growth. Nvidia stock swung over 20% in single days during 2023. NextEra, a utility Buffett has praised, rarely moves more than 2%. One is a rocket ship. The other, a freight train. Which would Buffett prefer? The answer seems obvious. But what if AI power needs surge 500% by 2030, as some forecasts suggest? Then the freight train becomes a high-speed rail. Berkshire’s existing energy assets could quietly benefit—without making a splash.
Industrial Automation: The Unseen AI Play
Another area: manufacturing. Precision Castparts Corp, owned by Berkshire, supplies aerospace parts. They use AI to optimize CNC machining, reduce waste, and predict tool wear. A 5% efficiency gain here saves millions. Is it AI? Yes. Is it headline-worthy? No. But it fits Buffett’s model: improve margins in a capital-intensive business. It’s a bit like upgrading a factory’s thermostat to a smart system—one that learns. To give a sense of scale, PCC’s annual revenue exceeds $10 billion. Even small gains compound.
Frequently Asked Questions
Does Berkshire Hathaway own stock in Nvidia or Microsoft?
No. As of Q1 2024, Berkshire Hathaway does not hold shares in Nvidia, Microsoft, or Meta—despite their AI leadership. The company has never disclosed a position in these firms. Its largest tech holding remains Apple. Some analysts speculate that the size of these stocks makes entry difficult without moving markets—another factor in Buffett’s caution.
Has Warren Buffett ever invested in any artificial intelligence companies?
Not directly. Buffett avoids companies where the future is uncertain or the technology too complex to evaluate. However, subsidiaries like Geico and BNSF use AI in operations. So while he doesn’t own “AI stocks,” he owns companies using AI to boost profits. The distinction matters.
Could Buffett change his mind about AI in the future?
We’re far from it being impossible. Buffett has evolved before—just look at Apple. If an AI company emerges with durable advantages, strong margins, and predictable earnings, he might reconsider. But it would need to pass the 10-year test. And that’s a high bar.
The Bottom Line: No Direct Bets, But Exposure Through the Back Door
Warren Buffett doesn’t own AI stocks in the way retail investors do—no calls on Nvidia, no options on Palantir. But pretending he’s untouched by AI is naive. His companies use it. His largest holding, Apple, is integrating it deeply. And his energy arm stands to gain from the massive power demands AI will require. The irony is palpable: the man who dodged the tech wave is now surfing it on a raft he didn’t build. I find this overrated as a contradiction. It’s just adaptation.
My take? Buffett won’t lead the AI charge. But he won’t be left behind. He’s playing a different game—one focused on enduring value, not quarterly hype. If you’re looking for a sign to mimic his portfolio, don’t chase the flash. Look for quiet efficiency. Find the companies where AI isn’t the product, but the engine. Because in the end, profits matter more than press releases. And that’s exactly where Berkshire excels.