In 1964, Warren Buffett first bought American Express when a scandal hit, but he believed the brand value of its traveller’s cheques and emerging credit card businesses was unaffected. He paid 7.6x 1963 pre-tax earnings and 2.3x book value.
Many credit his purchase of See’s Candy in 1972 at over 3x book value as his transition to seeing the value of intangible assets and earnings power, rather than buying stocks below their tangible asset value. But See’s was really the completion of Buffett’s transformation that began years earlier with American Express.
Buffett first bought Berkshire Hathaway’s stock when it was an ailing textile mill in 1965. As another tangible asset play, in 1967 Berkshire paid a dividend to extract cash from the poor business. Buffett likes to lament this as a mistake because the return on reinvesting that dividend into the conglomerate under his management would have been far greater. Berkshire has since not paid a dividend in 58 years.
He understood the value of intangibles, earnings power and reinvestment early on. The market eventually got there, but Buffett had leapfrogged it by about 30 years before the concept was widely accepted. Buffett was doing this in businesses like newspapers, ad agencies and consumer brands – not the technology companies that first adopted the capital-light reinvestment model.
For instance, computer company Digital Equipment Corporation, founded in 1957, did not pay a dividend throughout the 1960s and 1970s. It told investors that because it had a high return on equity, its earnings were better off being reinvested. Through the 1980s, Microsoft didn’t pay a dividend (until 2003) as it reinvested its earnings. Investors were satisfied.
By the 1990s, the market went a step further and investors were satisfied not only by Amazon’s lack of dividends, but its lack of profits. The market had also cottoned on to Buffett’s discovery and had rewarded non-tech companies like Walmart too. This trend accelerated in the 2010s as Tesla satisfied investors not only with no profits, but no revenues.
Today, the AI revolution is different because the companies are private – OpenAI and Anthropic – and investors appear satisfied not just by the absence of earnings and sales, but by increasing losses. The private investors are also supported by the big public tech companies – Alphabet, Amazon, Meta and Microsoft (the hyperscalers) – because they believe and will shoulder the losses with their cashflows.
Buffett’s reinvention through Berkshire
If Buffett has one trick, it has been buying high-quality businesses with great reinvestment rates at value prices. He went on to buy Coke at 10x (pre-tax earnings) in 1988 and Apple at 9x in 2016. Buffett took his mentor Ben Graham’s deep value philosophy and combined it with his own qualitative insight.
As the market was rewarding Amazon’s extreme reinvestment, Berkshire famously underperformed the market in the late 1990s but caught up through the early 2000s by holding bonds, cheap stocks and buying through the crash. Similar happened from 2007 to 2009. But the disruption of technology in the 1990s and its adoption into the 2010s did leave Buffett behind. He only caught Apple because the market missed its consumer brand qualities just as the smartphone market was consolidating.
The limits of scale
Berkshire has underperformed the S&P 500 since 2008, returning 9.8% a year compared to the market’s 10.9% (consider Berkshire’s fate without Apple). This contrasts sharply with 1985–2008, when Berkshire returned 19.7% a year against the S&P 500’s 6.8%. Even worse on a relative basis, from the 2009 market bottom Berkshire has returned 14.5% a year while the S&P 500 returned a whopping 16.1% being boosted by the Mag7’s technology-driven, capital-light businesses.
Today, American Express trades at 7.6x book and Coke 9.7x alongside Google’s 8.7x. Buffett now laments Berkshire’s size because, since 2008, he has been unable to buy capital-light businesses that are cheap enough nor large enough to move the dial.
From capital-light tech to capital-heavy reality
Since the turn of the 21st century, Berkshire has bought companies like MidAmerican Energy (electric utility), BNSF Railway (freight) and Occidental Petroleum (oil and gas). The fact these are so capital-heavy, not capital-light is worth pondering.
With the hyperscalers now pushing their cashflows into real-world assets like data centres and power generation, has Buffett been spending the last 15 years leap-frogging the market again? The capital-light tech giants have become too big to meaningfully reinvest in their digital realm. Spurred by their belief in AI, they are turning to physical assets like land, servers and gas turbines. Investors will tolerate physical investments that show as losses for OpenAI, which benefits the physical assets and operators themselves – and they are cheaper than ever.
Capital-light versus capital-heavy
As Buffett hands over to his successor, it looks like he is ahead again as the market rewards ever greater reinvestment. But this reaches a point where the balance shifts from the capital-light to the capital-heavy, where Berkshire lies in wait. Is this Buffett’s final trick?
Dan Scott Lintott is investment analyst at De Lisle Partners.The views expressed above should not be taken as investment advice.