Microsoft chief executive Satya Nadella told analysts on July 29, 2026, that "all of us are reading this 1873 as the book to be read," according to a transcript of the company's fiscal fourth-quarter 2026 earnings call. The book is "1873: The Rothschilds, the First Great Depression, and the Making of the Modern World," a 368-page history of the world's first global financial crash by Liaquat Ahamed, published by Penguin Press on June 2, 2026. Nadella made the remark in answer to a question about whether the AI industry is building too many data centres - on a call where Microsoft disclosed $41.0 billion of capital expenditure for a single quarter.

In Short

Microsoft's boss told investors that everyone at the company is reading a history book about a giant railway building boom in the 1860s and 1870s that ended in a crash. The book matters because Microsoft and its rivals are now spending hundreds of billions of dollars on data centres for AI, and many people worry that history could repeat itself. The author of the book says Microsoft's boss took the wrong lesson from it, so the recommendation has turned into a debate about whether one careful company can avoid a collapse caused by everyone overbuilding at once.

What Nadella said, and when

The exchange came late in the July 29 call. Mark Moerdler of Bernstein Research asked how Microsoft would protect itself if the industry overbuilt data centres and chips, and how it would handle rising component prices without raising customer prices or hurting margins, according to the transcript published by roic.ai. Moerdler described sentiment around AI as "incredibly volatile," according to the same transcript.

Chief financial officer Amy Hood answered first. She said demand was running ahead of supply in what she called "a sort of relatively extreme moment," according to the transcript, and that most capital spending goes into CPUs and GPUs, short-lived assets that the company can slow down if demand changes. Land and data centre buildings, she said, are more flexible on timing.

Nadella then added his own answer. "I thought Amy captured it well," he said, according to the transcript. "All of us are reading this 1873 as the book to be read." He went on to say that Microsoft has to get its product shape and its portfolio of customers, geographies, segments and workloads right when it builds capacity. "Then you've got to run an efficient railroad," he said, according to the transcript. And on the cycle itself: "Then we know that there will be ups and downs of what is the cycle here, but the secular shift is clear," according to the same source.

Two details are worth noting about the wording. First, the phrase that circulated afterwards - "the book to be read" - is accurate to the transcript, but the full sentence is grammatically loose, as spoken remarks often are. Second, Penguin Random House now carries a slightly different version on the book's page, attributed to The Times of London: "The book to read right now [is 1873] . . . All of us are reading this." The bracketed insertion is the publisher's or the newspaper's, not Nadella's.

At least one analyst newsletter published after the call, Nikhs on Substack, attributed the "efficient railroad" line to Hood. Transcripts published by roic.ai and BigGo Finance attribute it to Nadella.

What the book is about

Ahamed's book covers the boom of the 1850s and 1860s and the crash that followed. According to the publisher's description, "over the course of the 1850s and 1860s, during the first era of globalization, the world experienced an unprecedented economic boom." The fuel, in the publisher's account, was "an explosion in the global bond market," at the hub of which stood the Rothschilds, "arguably the wealthiest banking family in history."

Railways were the central investment. "While the giant sums of capital provided through the bond market built the railroads, the century's most transformative investments, the money raised also unleashed a frenzy of speculation, massive overinvestment, and wasteful borrowing by governments," according to Penguin Random House. Then, in the early 1870s, the bubble burst. "Stock markets from Vienna to New York crashed, and dozens of railroads and many governments defaulted," according to the same description.

The book does not stop at the market crash. According to the publisher, it follows currency reforms that deepened a long deflation, the effects of the crisis on Reconstruction in the United States and on the Ottoman Empire, and the antisemitic scapegoating of the Rothschilds that followed. In a Library Journal starred review dated May 1, 2026, Thomas Davis described policy errors such as the abandonment of bimetallism and rising tariffs as "blundering, self-inflicted wounds that choked global liquidity." He added that "pervasive pessimism bred a rising sense of resentment and political instability."

The hardcover lists at $32.00, according to Penguin Random House. An audiobook runs 9 hours and 41 minutes, read by George Newbern.

Reception

Early reviews were broadly favourable. Patrick Foulis wrote in the Financial Times that Ahamed "thrillingly brings back to life a boom not unlike today's," according to the publisher's page. Trevor Jackson in The New York Times Book Review wrote that "Ahamed tells his story with an easy fluency and a high velocity," while John Cassidy of The New Yorker called the book "timely and engaging," according to the same page. Kirkus, in a starred review, described it as "an exemplary work of economic history, with many lessons for the present."

Those lines come from the publisher's own page, which naturally selects favourable verdicts.

Who Liaquat Ahamed is

Ahamed is not an academic historian. He was born in Kenya in 1952 and studied at Trinity College, Cambridge, and at Harvard, according to Wikipedia. He worked at the World Bank in Washington, where he headed the investment division, and later at Fischer Francis Trees & Watts, a New York fixed-income firm owned by BNP Paribas, where he served as chief investment officer and chief executive.

According to Penguin Random House, he spent about twenty-five years as a professional investment manager in London and New York before turning to writing. He is a trustee of the Putnam Funds, an adviser to the Rock Creek Group, and chair of the Sun Valley Writers' Conference.

His first book, "Lords of Finance: The Bankers Who Broke the World," was published by Penguin Press in 2009. It followed the central bankers of the United States, Britain, France and Germany through the 1920s and into the Great Depression, and won the Pulitzer Prize for History in 2010. A shorter work, "Money and Tough Love: On Tour with the IMF," followed in 2014. "1873" is, in effect, a prequel: the first global depression, told by the author of the definitive popular history of the second.

Why a tech chief executive might reach for 1873

Nadella did not explain on the call which lesson he drew from the book. The framing that follows comes from Ahamed himself and from commentators who wrote about the remark; none of it is Microsoft's stated view.

Ahamed set out the comparison directly in a commentary for Fortune published on June 2, 2026, the day the book went on sale, under the headline "I won a Pulitzer for explaining the Great Depression. The AI spending boom terrifies me." His argument rests on scale and financing.

On scale, he wrote that Amazon, Microsoft, Alphabet, Meta and Oracle, along with OpenAI, Anthropic and startups, may spend $800 billion to $1 trillion a year on AI infrastructure for several years - about 2% to 3% of US GDP, according to the essay. The railway comparison is close. Between 1865 and 1872 the US rail network doubled from 35,000 to 70,000 miles, and railroad construction peaked at close to $400 million a year in the early 1870s, nearly 5% of GDP, according to Ahamed.

On financing, he pointed to bonds. "More than three-quarters of the railroad construction boom was financed by issuing bonds," he wrote in Fortune. When Jay Cooke's bank, overextended on the Northern Pacific, closed on September 18, 1873, the New York Stock Exchange suspended trading for ten days. Within two years a third of US railroad bonds had stopped paying interest, according to the essay.

The AI buildout started differently. Early spending came largely out of big tech profits, Ahamed acknowledged, but he argued that this has changed. "More and more data centers are being financed by companies issuing debt," he wrote, adding that private credit and special purpose vehicles now carry a growing share. His conclusion was blunt: "the line between visionary investment and ruinous excess is invisible until after you have crossed it."

He also offered a limit to the analogy. "Like the railroads, the AI capital investment boom - even if it suffers a temporary setback - is not going away," he wrote. The railways survived 1873, and the track that was laid stayed in use.

The author's response to Nadella

Ahamed did not take the recommendation as an endorsement of his argument. Speaking on The Ringer's "Plain English with Derek Thompson" podcast, in an episode published September 8, 2026, he said that "having the CEO of a $3.7 trillion company recommend your book is obviously a great boon." Then he added: "I have a feeling that he misread the lessons."

His reading of Nadella's takeaway, according to the episode page, was "you've just got to be very careful." What that misses, Ahamed said, is that "there's a collective action problem." Each company can try to be prudent, but when every competitor races to secure first place, the industry as a whole can still overbuild. In his Fortune essay he made the same point: competitive pressure pushes each firm to invest more, even when the sector is overbuilding.

That distinction goes to the heart of the July 29 exchange. Hood and Nadella answered Moerdler's question at the level of the firm - short-lived assets, flexible build schedules, a diversified customer base. Ahamed's argument operates at the level of the industry.

Other readings

Commentators drew their own conclusions. Shane Hoover, writing on Substack under the title "Gradually, then all at once: 1873," argued that manias share four ingredients - new innovation, regulatory shortcomings, overconfidence and excess debt - and that all four exist today. He cited S&P Global Ratings' projection that AI infrastructure investment will exceed $1.3 trillion by 2027. The Nikhs newsletter read the remark as a sign that Microsoft management was privately modelling a shakeout scenario. Both are interpretations, not statements by Microsoft.

The numbers on the call

The context for the remark was a quarter of heavy spending. Microsoft's capital expenditure including finance leases rose 70% to $41.0 billion in the June quarter, with roughly two-thirds going to short-lived assets such as CPUs and GPUs. Azure and other cloud services grew 43%, and Azure passed $100 billion in revenue for the fiscal year.

For the quarter, Microsoft reported revenue of about $90 billion, up 18%, according to MarketBeat's summary of the call. Hood guided first-quarter fiscal 2027 capital spending above $50 billion and said the company "remains capacity constrained," according to the same summary. Microsoft Cloud revenue exceeded $214 billion for the year, with nearly 90% coming from customers outside the frontier model companies, according to MarketBeat. That last figure bears on Moerdler's question: it was Microsoft's way of saying its demand is not concentrated in a handful of AI labs.

The prior quarter already showed the trajectory. In the March quarter Microsoft reported $31.9 billion of capital expenditure on revenue of $82.9 billion, leaving free cash flow of $15.8 billion.

Microsoft is not alone. Alphabet raised about $85 billion in equity in June 2026, including $10 billion from Berkshire Hathaway, citing demand for AI that was "meaningfully exceeding" supply - the kind of external financing Ahamed's essay flagged. PPC Land coverage in August put Alphabet's capex guidance at $195 billion to $205 billion and its second-quarter free cash flow at negative $5.9 billion, in a piece on Ed Zitron's argument that OpenAI and Anthropic could account for over 48% of Google Cloud revenue in 2027. Meta's head of infrastructure said the company would spend "well over $100 billion" in 2026.

Not all of the constraint is financial. Local opposition blocked or delayed at least $156 billion of data centre projects in 2025, according to Data Center Watch figures that media analyst Ian Whittaker cited when arguing that 2026 advertising forecasts rest on data centre construction continuing. And capacity remains short at the moment: Sundar Pichai said Google still lacks enough compute to serve everyone after Gemini 4 Argon's launch.

Why this matters for marketers

For the advertising industry, the question raised by "1873" is not academic. The hyperscaler companies building AI capacity are also the companies that sell most of the world's digital advertising: Alphabet, Meta, Amazon and Microsoft. Their advertising businesses fund a large part of the buildout. Meta, for instance, drew about 97.6% of its second-quarter 2026 revenue from advertising.

That creates two pressures that advertisers and publishers can watch. The first is pricing. Hood said on the call that new contracts can reflect higher component costs. When capital spending runs ahead of revenue, platforms have an incentive to extract more from existing revenue lines, and advertising is the largest of them at Alphabet and Meta. The second is product direction. The capacity being built is meant to run AI products - AI search, AI shopping, agent tools - that change how ads are placed and how traffic reaches publishers. Microsoft's own search advertising revenue excluding traffic acquisition costs grew 10% in the June quarter, down from 12%, and the company guided to mid-single-digit growth for the following quarter.

A correction in AI spending, if it came, would not be confined to chipmakers. Analysts such as Whittaker have tied ad market growth forecasts to the construction cycle. The rest of Ahamed's history is a reminder that the 1873 crash reached far beyond railroads - into government finances, currencies and politics on several continents.

None of this means a crash is coming. The book is history, the analogy is contested, and Nadella's own reading - by Ahamed's account - was a call for operational discipline rather than a forecast. Microsoft reports its first-quarter fiscal 2027 results on October 28, 2026, the next occasion on which analysts can ask whether anyone is still reading.

Timeline

Summary

Who: Microsoft chief executive Satya Nadella, chief financial officer Amy Hood, Bernstein analyst Mark Moerdler, and Liaquat Ahamed, the Pulitzer-winning author of "Lords of Finance" and "1873."

What: Nadella said on Microsoft's earnings call that "all of us are reading this 1873 as the book to be read," referring to Ahamed's history of the 1873 global financial crash, which the author and others connect to today's AI infrastructure spending. Ahamed later said Nadella "misread the lessons," arguing that the risk is a collective action problem rather than one company's discipline.

When: The book was published on June 2, 2026. Nadella made the remark on July 29, 2026, during the fiscal fourth-quarter 2026 call. Ahamed responded in a podcast published September 8, 2026.

Where: Microsoft's quarterly earnings call with analysts; the book is published by Penguin Press in the United States.

Why: The remark answered a question about whether the industry is overbuilding data centres, on a call where Microsoft reported $41.0 billion of quarterly capital expenditure. For marketers, the same companies funding the buildout run the largest advertising platforms, so the health of that spending cycle bears on ad pricing, product direction and market forecasts.