At the 2026 Berkshire shareholder meeting, AI will become a core topic.

This is the first time Warren Buffett has officially stepped into the background and hosted independently for the first time by his successor, Greg Abel.

However, the opening remarks of the conference were full of cyberpunk black humor: a man who called himself "95-year-old Warren" on the big screen asked Abel a question, and Abel revealed it on the spot - this was actually an AI-generated "deepfake" video. This slightly frightening opening also directly sets the tone for Berkshire's core attitude towards AI this year: extreme vigilance and pragmatism.

Greg Abel: "Never AI for AI's sake"

At a time when Silicon Valley and Wall Street are frantically chasing the AI boom, Berkshire seems particularly calm. The new CEO, Greg Abel, made it clear that the company is very pragmatic about AI and will never blindly engage in AI in order to follow the trend.

Abel continued Buffett's consistent investment philosophy, emphasizing that AI will only be used if it can actually help companies create value. Therefore, Berkshire will not be fully rolled out, but will focus on some specific business scenarios that can really make money.

When asked if he would blindly invest because technology companies are making money now, Abel also gave a negative answer. He said Berkshire will not pay money just because a company has the aura of "technology".

Investment must return to the old business, and look at three questions: Do we understand this business? Understand its risks and opportunities? Will it continue to make money in the next 5 to 10 years?

In addition, in response to a practical problem brought about by the development of AI - data centers consume extremely much power, which even leads to an increase in electricity bills for ordinary households, Abel also spoke. He called on these large data centers and electricity companies to bear all the costs of electricity themselves, and not let ordinary people pay for their energy consumption.

AI is a tool, not a decision-maker

Although AI technology is hot in all walks of life now, Berkshire still firmly believes that human judgment is the company's core competitiveness, and AI can only be used as an auxiliary tool and must not be allowed to make major decisions for people.

Ajit Jain, the head of the insurance business, was straightforward. He believes that although AI may change the world in the future, for now, it is more of a "productivity tool" that can help companies save effort and reduce costs, and is specifically used to deal with repetitive and boring daily chores.

至于保险里最核心的定价、理赔这些需要复杂权衡的环节，AI想要取代人类，起码还得再等上好几年。贾恩甚至还打了个比方：如果你指望AI来告诉你该买哪只股票、卖哪只股票，那基本是不可能发生的。

不过，这并不代表他们完全不用AI。CEO阿贝尔就举了个实际的例子：以前，经验丰富的核保人员可能最多只能盯住5个最大的风险点；但现在有了AI的帮忙，团队可以顺带把另外15个潜在风险也看得清清楚楚，大大提高了工作效率。

His core logic is very clear:

Reject concept hype: Berkshire will never blindly follow a company just because it is in the technology sector.

Back to the essentials of business: AI must bring substantial gains (create value) to the business, not to catch up with fashion.

|AI-driven BNSF railway transformation

Under the pressure of declining freight volumes, BNSF's railway business is actively introducing AI technology, hoping to stabilize profits and maintain an advantage in the fierce competition.

CEO Abel said that AI technology can be applied to the company's other businesses in the future. Berkshire will form its own team to develop the technology it needs and provide solutions to customers, but this is undoubtedly a huge challenge that cannot be achieved overnight.

BNSF executive Katie shared a very impressive set of results at the meeting: AI and other technologies have improved the efficiency of single transportation, and in the first quarter of this year, they transported more goods than in the first quarter of last year with 260 fewer locomotives.

She revealed that BNSF is vigorously recruiting data scientists and experts, focusing on exploring "digital twin" technology. To put it simply, it is to simulate a real railway network in the computer for drills, so as to more accurately predict the arrival time of goods and speed up the turnover of vehicles and goods, so as to compete with the competition of trucking.

You know, BNSF has not had a good time in recent years due to the decline in freight volume, and even carried out several rounds of layoffs in 2024 and 2025. Nowadays, technology empowerment has obviously become the key to their reversal.

Seize two types of signals and two types of opportunities

Although Buffett took a back seat, Berkshire's "soul" has not changed.

In the face of AI, a variable that can change the fate of mankind, Abel handed over a calm, restrained and very practical answer. He told us: No matter how cool technology is, if it cannot be converted into profit and efficiency, it is a castle in the air.

Signal 1: Cash is king, wait patiently

Abel reiterated that Berkshire holds up to $397 billion in cash and U.S. Treasury bonds. This is not only to deal with risks, but also to have the confidence to take action calmly when the market really has a "fat" opportunity (or AI really generates disruptive value). At present, market valuations are high, and patience is a virtue.

Signal 2: Core assets are unshakable

Abel clarified the foundations of Berkshire's equity investments: Apple, American Express, Moody's, Coca-Cola, and Japan's five major trading companies. These are still the ballast stones of the cycle through the cycle.

Opportunity 1: AI infrastructure and energy

Berkshire's logic verifies that the business of "selling shovels in the AI gold rush" is the most stable. Focus on utilities that provide power support for AI data centers, as well as traditional industry leaders (such as railways and logistics) that can effectively use AI to reduce costs and increase efficiency.

Opportunity 2: Network security track

"Deepfakes" have entered the room, and the cybersecurity threat posed by AI is systemic. Whether it is enterprises or individuals, the demand for high-level network security services and identity authentication technology will grow exponentially.