After watching the excellence of Google's business model firsthand, Berkshire Hathaway chose to wait for a full twenty years. Now, its successor is deploying a $10 billion investment, replicating Warren Buffett's most classic capital allocation logic—except this time, Berkshire Hathaway itself is playing the role of the "See's Candies," while Google Cloud may be growing into the BNSF Railway of the AI era.

According to reports, Alphabet, Google's parent company, is raising $80 billion through a package of equity financing. This includes a $10 billion private placement deal with Berkshire Hathaway, a $30 billion underwritten offering, and a $40 billion at-the-market equity offering plan. This is one of the largest equity transactions in history, and Berkshire's entry has surprised the market.

Ben Thompson, founder of technology strategy analysis firm Stratechery, believes the core logic of this investment is that Google is not only one of the best business models today, but its cloud business, fueled by exploding AI demand, could grow into a new engine whose absolute scale far surpasses its advertising business.

The Timing of a New Captain's Move

The timing of this move by Greg Abel, Buffett's successor and the new CEO of Berkshire Hathaway, is significant.

Berkshire currently holds $373 billion in cash, with free cash flow projected to reach $25 billion in 2025. Against the backdrop of an intensifying AI infrastructure race and computing power demand consistently exceeding market expectations, Google—with its cost advantages from its self-developed TPU chips and its multiple bets across the model, service, and cloud capacity sales layers—is seen as one of the few options capable of effectively absorbing Berkshire's massive capital and generating high returns.

Buffett's Regret: The Business He "Hardly Ever Saw"

Google is a business that Buffett has publicly admitted to missing.

At the 2017 Berkshire Hathaway Annual Shareholders Meeting, he confessed that Berkshire's GEICO insurance company was an extremely early Google advertising client, paying $10 to $11 per click at the time. "Anytime you're paying somebody 10 or 11 dollars for a click on something that costs them nothing, you know it's a good business—unless somebody takes it away from you," Buffett said. "You hardly ever see a business like that."

However, it was precisely this "light-asset, high-margin" business model that made it difficult for Buffett to invest.

Ben Thompson's analysis in Stratechery points out that Google's Aggregator nature means it maximizes absolute value by sacrificing relative value—content creators on the supply side, advertisers and users on the demand side all contribute value within this mechanism, but the relative return per unit is diluted. This does not naturally fit Buffett's preferred framework of "buying wonderful companies at fair prices." He is more familiar with heavy-asset companies that rely on tangible asset accumulation and have stable, predictable cash flows.

See's Candies and BNSF: Two Faces of Capital Compounding

Understanding this investment requires first understanding Berkshire's capital operation logic.

In 1972, Berkshire bought See's Candies for $25 million. At the time, See's had sales of $30 million, pre-tax profit under $5 million, and required operating capital of just $8 million. By 2007, See's annual sales had reached $383 million with pre-tax profit of $82 million, while cumulative additional investment was only $32 million. Cumulative pre-tax profit, however, reached $1.35 billion. This is what Buffett calls "time is the friend of the wonderful business."

Berkshire deployed the cash continuously generated by See's Candies into a completely different business—BNSF Railway.

Railroads are classic heavy-asset industries; BNSF's capital expenditure last year was $3.8 billion. But its absolute returns are also substantial: revenue reached $23.4 billion with net profit of $5.5 billion. Ben Thompson notes that the cumulative profit See's Candies has contributed to Berkshire (last disclosed as "over $2 billion" in 2019) might not even match BNSF's annual net profit for a single year. The two businesses are, in essence, the same logic: using the output of a cash flow machine to leverage a heavy-asset business of larger absolute scale.

Google Cloud: Lower Margins, Greater Potential

Data cited by Ben Thompson shows this history is replaying inside Google.

In Q4 2019, Google Services (the core ad business) revenue was $43.2 billion with operating profit of $13.5 billion; Google Cloud revenue was just $2.6 billion with a loss of $1.2 billion. In Q1 2026, Google Services revenue grew to $89.6 billion with operating profit of $40.6 billion; Google Cloud revenue reached $20 billion with operating profit of $6.6 billion, a margin of 33%, compared to Google Services' 45% margin.

Over seven years, Google Services revenue more than doubled and profit nearly tripled. Meanwhile, Google Cloud revenue as a percentage of Google Services rose from 6% to 22%, and its profit share increased from almost zero to 16%, growing faster and expanding its margin more significantly.

The more critical question is: the ceiling for the advertising business is constrained by the share of advertising spend in the overall economy, while the potential market for cloud and AI could, in some scenarios, encompass the entire economy. As Ben Thompson writes, the role of Google Services today may precisely be to provide the ammunition for a future business that is larger in scale, with slightly lower margins but higher absolute profit.

Why Berkshire is Entering Now and Why Google is Issuing Equity

This transaction also raises two noteworthy questions. First, why did Google choose to issue equity rather than debt?

From a financial logic perspective, debt financing is more economical—interest is tax-deductible and it does not dilute existing shareholder equity. Google currently has about $81 billion in debt on its books but holds a massive $126 billion in cash, theoretically leaving significant room for more borrowing.

Ben Thompson's assessment is that Google is highly likely to issue large-scale debt thereafter. The signal of this equity financing is that the scale of computing power demand is still systematically underestimated by the market, and Google is willing to use all available financing tools to meet supply.

Of course, another interpretation is that Google is uncertain about the returns on massive capital expenditure and wishes to share the risk. If large-scale debt financing does not follow, the credibility of this interpretation would rise.

Second, what is Berkshire's logic?

Ben Thompson believes Greg Abel's move essentially replicates Buffett's playbook: with Berkshire playing See's Candies and Google playing BNSF.

Facing a $373 billion cash pile and $25 billion in annual free cash flow, targets that can absorb such a volume of capital while maintaining high returns are extremely scarce. Google's advantage lies in having multiple options in play: its advertising services business directly benefits from AI investment, Gemini maintains a competitive position at the model layer, and with the cost advantage of TPUs, Google Cloud can maintain excess profits even in a scenario of commoditized compute, while selling compute capacity to external customers.

Cash is King: The Foundational Logic of the AI Race

Ben Thompson previously wrote in Stratechery that the key to AI competition is not merely who has more computing power first, but who has sufficient cash flow to continuously purchase computing power.

He used Anthropic as an example: despite OpenAI once claiming superiority in compute resources, Anthropic, with its sustained commercial revenue and fundraising ability, eventually secured a large compute procurement deal with SpaceX, proving that entities with ample cash flow could continuously acquire needed computing resources in the secondary market.

This logic extends to the ultimate competitive level—when compute supply tightens, the company with the strongest cash-generating ability will gain the most sustainable compute advantage, forming a positive flywheel. In Ben Thompson's view, at this moment, the company that best fits this description is Google. And Berkshire Hathaway has given its answer with $10 billion.