What does the most beautiful business model of all time look like?

First, imagine that your supply is free. Second, imagine that your customers willfully compete against each other to raise your prices. Third, imagine that your users decide which of your customers gets the privilege of paying you. All you have to do is build a bit of infrastructure to make it all happen, pay a nominal bit of depreciation on that infrastructure, and make billions of dollars on some of the greatest margins in the history of business.

I am, of course, describing Google, a company so good that Warren Buffett, the legendary investor, could never quite bring himself to invest in it. Buffett explained in the 2017 Berkshire Hathaway annual meeting:

We were their customer very early on with GEICO, for example, and we saw — these figures are way out of date — but as I remember, we were paying them $10 or $11 a click, or something like that. And any time you’re paying somebody $10 or $11 bucks every time somebody just punches a little thing where you got no cost at all, you know, that’s a good business unless somebody’s going to take it away from you. And so we were close up seeing the impact of that…But, you know, you’ve almost never seen a business like it.

One of the characteristics of an Aggregator like Google is the way in which they maximize absolute value at the expense of relative value. For supply — i.e. content on the web — Google dramatically increases the number of visitors, even as the value of any one visitor who comes from Google is worth much less than a visitor who visits directly; for an advertiser, the value of one click makes up for thousands of impressions of an ad that make no difference; for a user, Google helps them discover what they are looking for amidst the overwhelming abundance that is downstream from distribution being free. In every case the Aggregator increases quantity at the expense of relative quality, confident that the absolute amount of quality will be more in the long run.

What is interesting is that this is the exact inverse in terms of why these companies have been valued by investors. The best tech companies are “asset-light”, predicated on maximizing zero marginal costs. Yes, they spend a lot of money on R&D and on the infrastructure to make markets happen, but they don’t actually participate in those markets; simply taking a skim and keeping the vast majority of that skim is what gets Wall Street excited. In other words, it was the relative amount of money made that was generally more important to the market than the absolute amount of money.

Berkshire Hathaway and Productive Capital

Berkshire Hathaway was, before Buffett acquired it, a failing textile business; Buffett originally invested because the stock was worth less than the liquidation value, and ended up owning it outright after a dispute with management. It was a decision he regretted; from the company’s 1989 letter to shareholders:

If you buy a stock at a sufficiently low price, there will usually be some hiccup in the fortunes of the business that gives you a chance to unload at a decent profit, even though the long-term performance of the business may be terrible…Time is the friend of the wonderful business, the enemy of the mediocre…

I could give you other personal examples of “bargain-purchase” folly but I’m sure you get the picture: It’s far better to buy a wonderful company at a fair price than a fair company at a wonderful price. Charlie understood this early; I was a slow learner. But now, when buying companies or common stocks, we look for first-class businesses accompanied by first-class managements.

One of the first-class businesses Berkshire Hathaway acquired was See’s Candies in 1972. Buffett explained in the 2007 shareholder letter:

We bought See’s for $25 million when its sales were $30 million and pre-tax earnings were less than $5 million. The capital then required to conduct the business was $8 million. (Modest seasonal debt was also needed for a few months each year.) Consequently, the company was earning 60% pre-tax on invested capital…

Last year See’s sales were $383 million, and pre-tax profits were $82 million. The capital now required to run the business is $40 million. This means we have had to reinvest only $32 million since 1972 to handle the modest physical growth – and somewhat immodest financial growth – of the business. In the meantime pre-tax earnings have totaled $1.35 billion. All of that, except for the $32 million, has been sent to Berkshire (or, in the early years, to Blue Chip).

The “problem” with a See’s Candies is that there is nothing to be done with all of that profit; if it’s privately held then its owners end up with more cash than they know what to do with, and if it’s public, then the job is to figure out how to return that cash to shareholders through some combination of dividends and stock buybacks. What Berkshire Hathaway did, however, was use that cash to grow:

After paying corporate taxes on the profits, we have used the rest to buy other attractive businesses. Just as Adam and Eve kick-started an activity that led to six billion humans, See’s has given birth to multiple new streams of cash for us. (The biblical command to “be fruitful and multiply” is one we take seriously at Berkshire.)

One of the businesses Berkshire Hathaway used the See’s profits for was on the opposite end of the spectrum in terms of capital utilization: BNSF Railway. Railways require a lot of capital to operate; BNSF consumed $3.8 billion last year; they also make a lot of money: BNSF’s net income was $5.5 billion on revenue of $23.4 billion. To put that in perspective, the total amount that Berkshire Hathaway has made from See’s Candies is probably less than $3 billion (the last disclosure was “over $2 billion” in 2019), i.e. less than BNSF made last year.

So which is the better business?

Google Cloud’s Runway

In Q4 2019, the first year that Alphabet disclosed Google Cloud revenue, Google Services — the high margin beautiful business I described at the beginning — made $43.2 billion in revenue and $13.5 billion in operating profit; Google Cloud made $2.6 billion in revenue and lost $1.2 billion. Google Cloud revenue was 6% the size of Google Services.

In Q1 2023, Google Cloud made a profit for the first time. In that quarter Google Services made $62.0 billion in revenue and $21.7 billion in profit; Google Cloud made $7.5 billion in revenue and $0.2 billion in profit. Google Cloud revenue was 12% the size of Google Services, and its profit was 1% the size of Google Services.

In Q1 2026, Google Services made $89.6 billion in revenue and $40.6 billion in profit; Google Cloud made $20.0 billion in revenue and $6.6 billion in profit. Google Cloud revenue was 22% the size of Google Services, and its profit was 16% the size of Google Services.

Google Services is, needless to say, a much more scalable business than See’s Candies. The growth just over the last seven years — more than doubling revenue and tripling profits — is astounding. And yet, at the same time, Google Cloud is growing faster, and while its margins are worse — 33% last quarter as compared to 45% for Google Services — they are expanding more rapidly.

The bigger question is how big can those numbers go? Google Services’ advertising business is inherently high margin, but advertising is definitionally but a fraction of the overall economy; Google Cloud’s growth, meanwhile, is AI, which many people think/worry/hope might take over the entire economy. In other words, might we one day look back and realize that Google Services provided the cash flow to build a business with relatively worse margins but absolutely higher dollars, much like See’s helped fund BNSF?

Berkshire Hathaway and Google Equity

The context for this discussion is this news from Bloomberg:

Google parent Alphabet Inc. is raising $80 billion through a package of equity offerings, including an investment deal with Berkshire Hathaway Inc., as the company races to fund its ambitious artificial intelligence spending plans. The undertaking includes a $40 billion so-called at-the-market program to sell shares from time to time beginning in the third quarter, according to a statement Monday. The company will also offer $30 billion in underwritten offerings of shares and mandatory convertible preferred stock, as well as the $10 billion deal with Berkshire. Together, the transactions represent one of the largest equity deals of all time — and they bring an unexpected twist to a blockbuster year for initial public offerings.

First off, a decent portion of the ATM program, launching in the fall, is going towards paying tax obligations on Google equity awards (which are quite large thanks to the stock’s run-up in value).

That leaves equity being issued now, particularly the $10 billion to Berkshire Hathaway, which is fascinating for a number of reasons. The first question is why did Google issue equity instead of debt? Debt is, all things being equal, the preferred instrument for investment: the proceeds of the latter pay off the former, and existing equity holders reap all of the benefits. Equity, on the other hand, removes the risk of debt, but at the cost of giving up a share of future profits.

Google has to date funded its massive AI-related capital expenditures with free cash flow, and while the company does have around $81 billion in debt, that is more than balanced by $126 billion of cash. In other words, Google’s capacity to issue more debt — and to reap the financial benefits of doing so (because interest is tax-deductible) — is substantial.

That leads to what may be the Occam’s Razor explanation: Google is also going to start issuing a lot more debt as well, which is to say that everyone continues to underestimate the amount of demand there is for compute. Of course that’s not far off from a more bearish interpretation: Google is uncertain about the return on investment of all that capex, and would prefer to share the risk (along with the upside). If there isn’t a substantial debt issuance down the road then this might be the right answer.

The second question is why is Berkshire Hathaway suddenly, after all these years, interested in Google, and at only a slight discount to its all-time high price? Does it really just come down to the fact that Buffett is no longer making investment decisions, and Greg Abel, his successor as CEO, is?

In fact, you can make the case that Abel is actually just replaying Buffett’s strategy, only this time Berkshire Hathaway is See’s Candies, and Google is BNSF. At the end of last quarter Berkshire Hathaway had $373 billion in cash, and $25 billion in free cash flow in 2025. How many companies could actually employ that cash in a way that generated a high rate of return?

It’s hard to imagine a better option than Google. The company is not only investing in AI, but has optionality in terms of outcomes: its Services business benefits from the investment, it is in contention at the model layer with Gemini, and it can sell capacity to the frontier labs. Moreover, that capacity has a sustainable cost advantage because of TPUs, which means that in a world where compute becomes a commodity — as hard as that is to imagine right now — Google is the hyperscaler that is poised to make the most profit.

It is worth noting that $10 billion is a relatively small amount of money to both companies. To that end, perhaps the primary utility is as a signaling mechanism. On Google’s side, the signal is that the expected demand is actually far greater than anyone thinks, and that the company is ready and willing to fund supply using all means at its disposal, including equity; for them Berkshire Hathaway’s investment is an endorsement of this view and a validation of the wisdom of the investment. And, on the flip side, if the signal is correct, then Berkshire Hathaway is getting a deal and putting its cash flow machines to work building the future.

Cash the Ultimate Commodity

A couple of months ago, when Anthropic was clearly ascendant, OpenAI backers tried to make the case that actually OpenAI was in the driver’s seat because the frontier lab had secured more compute; I made the case in Mythos, Muse, and the Opportunity Cost of Compute that this was not at all dispositive:

OpenAI is betting that this compute constraint — and the deals they have made to overcome it — will matter more than Anthropic’s current momentum with end users…I’m less certain that this will be dispositive. When it comes to AI, distribution and transaction costs are still free — the two preconditions for Aggregators — which means that the winners should be those with the most compelling products. Those products will win the most users, providing the money necessary to source the compute to serve them; consider Anthropic’s deal to secure a meaningful portion of TPU supply, which, given the capacity constraints at TSMC, is ultimately an example of taking supply from Google. I suspect that Anthropic can take more, including already built hyperscaler and neocloud capacity. Yes, that compute will be more expensive, but if demand is high enough the necessary cash flow will be there.

That thesis was proven correct just weeks later when SpaceX ponied up the supply Anthropic needed (and yes, it was expensive):

This deal is a perfect example of what really is basic economics. First, if demand exceeds supply, then prices should increase. At the same time, prices are elastic: if they are lower there is more demand, and if they are higher there is less demand. In this case, while there is broad demand for computing, Anthropic has arguably the most demand; furthermore, Anthropic has the most willingness to pay, not just because they are making meaningful revenue, but also because they have the capacity to raise money in the pursuit of winning in AI.

Implicit in this analysis was that there was enough compute capacity in the world to be bought; what happens, however, when and if there isn’t? What if the ultimate battle — the one that determines who gets compute — becomes a matter of who can bring the most cash to bear? And what if that advantage compounds, such that the company with the most cash capacity ends up with the most compute capacity (which we already know they will sell, in addition to using themselves) driving the ability to generate more cash? In that world, what company would be your best bet?

We now know which one Berkshire Hathaway is betting on.¹

1. As an aside, it’s notable that Alphabet has another business — Waymo — where the company has so far rejected an asset-light model of licensing their software to OEMs, and has instead to date pursued a much more capital intensive approach of owning and operating their own cars; that’s a choice that has always felt at odds with Google Services, but is perhaps more compelling and aligned with Google Cloud and the Google Capital Company. ↩︎